INTERNATIONAL ORGANISATION FOR
STANDARDISATION
ORGANISATION INTERNATIONALE DE NORMALISATION
ISO/IEC JTC 1/SC 29/WG 11
CODING OF MOVING PICTURES AND AUDIO
N8434
ISO/IEC JTC 1/SC 29/WG 11
Hangzhou, CN – October 2006
Source: Leonardo Chiariglione
Title:
Report of 78th meeting
Status
Report of 78th meeting
WG11 report
Annex A – Attendance list
Annex B – Agenda
Annex C – Input contributions
Annex D – Output documents
Annex E – Requirements report
Annex F – Systems report
Annex G – MDS report
Annex H – Video report
Annex I – Audio report
Annex J – 3DG report
Annex K – Test report
Annex L – ISG report
Annex M – Liaisons report
WG11 report
1
Opening
The 78th MPEG Meeting was held on 2006/10/23T09:00-27T20:50 at Vaton Yunqi Resort
Hotel, Hangzhou, China at the kind invitation of the Chinese National Body and hosted by
Zhejiang University.
2
Roll call of participants
Annex 1 provides the list of participants.
3
Approval of agenda
This is given in annex 2.
1
4
Allocation of contributions
The list of input documents is given in annex 3.
5
Communications from Convenor
There were no specific communications made.
6
Report of previous meeting
This was approved
7
Processing of NB Position Papers
National Body Position Papers were presented, discussed and, where required, responses were
prepared and agreed.
8
Work plan
8.1 Media coding
8.1.1 Fixed point implementation of DCT/IDCT
The following documents were approved
8479 ISO/IEC CD 23002-2 Fixed point IDCT and DCT
8480 Description of Core Experiments on Fixed-Point DCT/IDCT
8481 Software Testbed for fixed-point DCT/IDCT V 5.0
8.1.2 Colour spaces
The following documents were approved
8445 Disposition of Comments on ISO/IEC 13818-2:2000/FPDAM 2
8446 Text of ISO/IEC 13818-2:2000/FDAM 2 Support for Colour Spaces
8447 Disposition of Comments on ISO/IEC 14496-2:2004/FPDAM3
8448 Text of ISO/IEC 14496-2:2004/FDAM 3 Support for Colour Spaces
8450 Disposition of Comments on ISO/IEC 14496-10:2005/FPDAM1
8451 Text of ISO/IEC 14496-10:2005/FDAM 1 Support for Colour Spaces and Aspect
Ratios
8.1.3 Multiview Video Coding
The following documents were approved
8458 Working Draft 1 of ISO/IEC 14496-10:2005/Amd.4 Multiview Video Coding
8459 Joint Multiview Video Model (JMVM) 2
8460 JMVM 2 Software
8.1.4 Symbolic Music Representation
The following documents were approved
8613 DoC on ISO/IEC 14496-3:2006/PDAM 6, Symbolic Music Representation
8631 Request for Subdivision, Symbolic Music Representation
8632 ISO/IEC 14496-23:200x/FCD, Symbolic Music Representation
2
8.1.5 Reconfigurable Video Coding
The following documents were approved
8483 Request for Subdivision: ISO/IEC 23002-4 Video Tool Library
8484 WD 2 of ISO/IEC 23002-4 Video Tool Library
8485 White Paper on Configurable Video Coding (RVC)
8486 Description of Core Experiments in RVC
8487 RVC Simulation Model (RSM) V2.0
8488 RVC Work Plan
8.1.6 Scalable audio and speech coding
The following document was approved
8640 Workplan for Exploration of Speech and Audio Coding
8.2 Description Coding
8.2.1 Technologies for digital photo management using MPEG-7 visual tools
The following documents were approved
8469 Disposition of Comments on ISO/IEC TR 15938-8:2002/DAM3
8470 Text of ISO/IEC TR 15938-8:2002/FDAM3 (Technologies for digital photo
management using MPEG-7 visual tools)
8.2.2 MPEG-7 Query Format
The following documents were approved
8509 MPEG-7 Query Format Requirements
8510 Final Call for Proposals on MPEG-7 Query Format
8.3 Systems support
8.3.1 Fragments Request Unit
The following document was approved
8683 DoC on ISO/IEC 23001-2/CD (Fragment Request Unit)
8.4 IPMP
8.4.1 MPEG-21 IPMP Component Base Profile
8563 DoC for ISO/IEC 21000-4/PDAM 1: MPEG-21 IPMP Components Base Profile
8564 ISO/IEC 21000-4/FPDAM 1: IPMP Components Base Profile
3
8.4.2 REL Profiles
The following documents were approved
8565 Request for Amendment 3 of ISO/IEC 21000-5 ORC (Open Release Content) Profile
8566 ISO/IEC 21000-5/PDAM 3 ORC (Open Release Content ) Profile
8.5 Digital Item
8.5.1 Dynamic and Distributed Adaptations
The following documents were approved
8569 Disposition of Comments on ISO/IEC 21000-7/FPDAM 2
8570 Text of ISO/IEC 21000-7/ FDAM 2 Dynamic and Distributed Adaptation
8.6 Transport and File Format
8.6.1 Transport of MPEG Surround data in AAC
The following documents were approved
8610 DoC on ISO/IEC 13818-7:2006/PDAM 1
8611 ISO/IEC 13818-7:2006/FPDAM 1, Transport of MPEG Surround data in AAC
8.6.2 File Format extensions for Description of Timed Metadata
The following documents were approved
8658 DoC on ISO/IEC 14496-12/FPDAM1 (Description of Timed Metadata)
8659 Text of ISO/IEC 14496-12/FDAM1 (Description of Timed Metadata)
8.6.3 Flute Hint Track
The following documents were approved
8660 DoC on ISO/IEC 14496-12/PDAM2 (Flute Hint Track)
8661 Text of ISO/IEC 14496-12/FPDAM2 (Flute Hint Track)
8.6.4 Digital Item Streaming
The following documents were approved
8575 DoC of ISO/IEC FCD 21000-18 Digital Item Streaming
8576 Text of ISO/IEC FDIS 21000-18 Digital Item Streaming
Request for Amendment 1 of ISO/IEC 21000-18 Digital Item Streaming: Simple
8579
Fragmentation Rule
8580 ISO/IEC 21000-18/PDAM/1 Digital Item Streaming
4
8.7 Multimedia architecture
8.7.1 M3W Component Download
The following documents was approved
8608
Text of ISO/IEC 23004-5/FCD Component Download
8.7.2 M3W Fault Management
The following documents was approved
8700
Text of ISO/IEC 23004-6/FCD Fault Management
8.7.3 M3W System Integrity Management
The following documents was approved
8701
Text of ISO/IEC 23004-7/FCD System Integrity Management
8.8 Application formats
8.8.1 Protected Music Player MAF
The following documents were approved
8581 DoC on ISO/IEC CD 23000-2 MPEG-A Music Player 2nd edition
8582 ISO/IEC FCD 23000-2 MPEG-A Music Player 2nd edition
8.8.2 Photo Player MAF
The following documents were approved
8471 Disposition of comments on ISO/IEC FCD 23000-3
8472 Text of ISO/IEC FDIS 23000-3
8473 Request for ISO/IEC 23000-3/Amd.1: Reference Software for Photo Player MAF
8.8.3 Musical Slide Show MAF
The following documents were approved
8673 DoC on ISO/IEC 23000-4/CD (Musical Slide Show MAF)
8674 Text of ISO/IEC 23000-4/FCD (Musical Slide Show MAF)
8.8.4 Media Streaming MAF
The following document was approved
8584 ISO/IEC CD 23000-5 Media Streaming Player
9
Liaison matters
The following liaison documents were issued
5
8526
8527
8528
8529
8530
8531
8532
8533
8534
8535
8536
8537
8538
8539
8540
8541
8542
8543
8544
8545
8546
8702
Liaison Statement to UHAPI concerning M3W
Liaison Statement to ITU-T FG/IPTV WG 6 concerning M3W
Liaison Statement to 3GPP2
Liaison Statement to ITU-R SG6 WP 6J concerning colour space amendments
Liaison Statement to ITU-R SG6 WP 6Q on Call for Proposals
Liaison Statement to SMPTE
Liaison Statement to SMPTE on 4:2:2 and 4:2:0 Intra-only profiles of AVC
Liaison Statement to SMPTE on 4:4:4 Intra-only profile of AVC
Liaison Statement to FLO Forum
Liaison Statement to IEC TC100
Liaison Statement to W3C MMSem-XG
Liaison Statement to ITU-T SG9 concerning FTV ad MVC
Liaison Statement to OMA BAC MAE
Liaison Statement to DVB
Liaison Statement to ISO TC184 SC4
Liaison statement to AES
Liaison Statement to SCTE
Liaison Statement to WG1 (JPEG)
Liaison Statement to Khronos
Liaison Statement to ITU-T FG/IPTV WG 6 concerning work on IPTV
Liaison Statement to ITU-T SG16 Q23
Liaison Statement to 3GPP
10 Organisation of this meeting
10.1.1 Tasks for subgroups
The following tasks were assigned
S
P
A
4
10
10 3
21
4
Requirements
7
21
A
C
W
X
Y
Z
5
X
2
AVC Profiles
SVC Profiles
Laser use case and requirements
IPMP Component Profile requirements
GPS, colour, calibration etc. reqs
REL Profiling
MAFs under consideration
IDCT
Free Viewpoint TV
MPEG-7 Query Format
Terminology
Dual track licensing approach
Systems
2
4
1 2
3
4 17
22
Transport of Auxiliary Data
JPEG2000 support
ATG conformance
Audio BIFS conformance
6
5
11
12
7
21
15
20
7
23
24
25
26
12
14
1
1
2
1
1
2
8
14
16
Synthetised texture conformance
File format conformance
Laser conformance
Open Font Format Conformance
File Format Reference software
Open Font Format Reference Software
SMR
Description of timed metadata
FLUTE hint track
SVC File Format
Lightweight Scene Representation
Fast Access Extension Conformance
File Format Reference Software
File Format Conformance
Binary DI conformance
A
B
4
6
1 1
2
2
E
xxx
1
2
3
4
5
6
7
8
1
Musical Slide Show MAF
Audio Archival MAF
Reference software and conformance
Extension on encoding of wild cards
Fragment Request Unit
MPEG Multimedia Middleware
Reference Software
MS MAF Protocols
MDS
21
4
4
6
7
8
14
1
2
2
1
IPMP Components Amendment 1
IPMP Components Amendment 2
RDD Implementation Issues
DIA Dynamic and Distributed Adaptation
Reference software
IPMP Components
DIA
DIP
ER
FID
DIS
Conformance
IPMP Components
DIA
DIP
7
ER
FID
Digital Item Streaming
Schemas
Protected Music Player MAF
Photo player MAF
Media Streaming MAF
Audio Archival MAF
MAFs under consideration
18
A
2
3
5
6
X
2
4
7
A
B
C
2
2
8 3
3
4
2
3
4
10 1
Video
JVT
4
2
3
3
4
Colour spaces
Colour spaces
New Visual Extensions
Photo Player
Reconfigurable Video Coding
Fixed-point 8x8 IDCT and DCT
Auxiliary video data
Reconfigurable Video Coding
Colour spaces
AVC profiles
4:4:4 profiles
Scalable Video Coding
SVC Profiles
MV Video Coding
Audio
4
7
A
D
3 1
3 6
3 5
3 7
4 14
15
16
17
18
5
6
2 1
2
6
1 1
1
1
Low delay AAC profile
Symbolic Music Representation
BSAC and SBR
Audio/systems interaction
BSAC conformance
1 bit lossless conformance
MPEG-1 and -2 on MPEG-4 conformance
ALS conformance
SLS conformance
New Audio Extensions RS
New Audio Extensions conformance
Music Player MAF Conformance and reference software
Protected Music Player MAF
Audio Archival MAF
MPEG Surround extensions
MPEG Surround Reference Software
MPEG Surround Conformance
Scalable Audio and Speech Coding
4 12
16
21
5 9
Conformance of Morphing and Textures
Conformance MPEG-J GFX
Conformance of Geometry and shadow
Reference software of Morphing and Textures
X
3DG
4
8
11 Reference software MPEG-J GFX
13 Reference Software of Geometry and shadow
16 2 AFX (Geometry and shadow)
Test
4
X
10 3
4
9 2
3
6
SVC verification tests
Dual track codecs
ISG
7
Reference Hardware Description
Reference Hardware Description
Reference software
Liaison
1
2
4
7
21
A
B
X
10.1.2 Joint meetings
The following joint meetings were held
Groups
Req, mds
Req, mds, sys, vid
Sys, mds, req
Mds, vid
Req, sys
Mds, aud
Req, vid, isg
Req, vid, jvt
Mds, sys
Vid, sys, mds, isg
Req, vid
Sys, mds
Req, aud
What
Mp7, mp21 profile, maf
New maf proposals
File formats
BSDL in SVC
Laser reqs
Audio archival, MP MAF
IDCT
AVC & SVC profiles
Mp21 in laser
RVC
Free View Point TV
Conversion between metadata systems
Speech and audio coding
Day
Tue
Tue
Tue
Wed
Wed
Wed
Wed
Wed
Wed
Thu
Thu
Thu
Thu
Where
req
req
req
mds
req
aud
isg
jvt
sys
vid
Req
Sys
Aud
Time
09:00-12:00
14:00-17:30
17:30-18:00
11:00-11:30
11:30-12:00
14:00-15:00
14:00-15:00
16:00-18:00
16:30-17:30
09:00-10:00
10:00-10:30
10:00-11:00
14:00-14:30
10.1.3 Development of MPEG standards
The following document was approved
8513 Dual Track Straw Man for IP Free and IP bearing but Royalty Free Standards Making
9
11 Administrative matters
11.1.1 Schedule of future MPEG meetings
#
78
79
80
81
82
83
84
85
86
City
Hangzhou
Marrakech
San José
Lausanne
Shenzhen
Sun City?
Geneva?
Hannover
?
Country
CN
MA
US
CH
CN
ZA?
CH?
DE
?
yy
06
07
07
07
07
08
08
08
08
mm
10
01
04
07
10
01
04
07
10
dd-dd
23-27
15-19
23-27
02-06
22-26
14-18
21-25
21-25
13-17
11.1.2 Promotional activities
The following documents were approved
8512
8600
8601
8602
8603
8604
8641
8642
8687
8688
8689
8690
8691
8692
8693
First MAF Awareness Event
MPEG-21 Session Mobility One Pager
MPEG-21 Digital Item Processing Amendment 1 One Pager
MPEG-21 Conformance to Digital Item Processing One Pager
MPEG-21 Conformance to Digital Item Processing Amendment 1 One Pager
MPEG-21 Reference Software One Pager
Audio Bifs version 3
Audio Conformance and Reference Software Assets
M3W White Paper: Multimedia Middleware Architecture
M3W White Paper: Multimedia API
M3W White Paper: Component Model
M3W White Paper: Resource and Quality Management
M3W White Paper: Component Download
M3W White Paper: Fault Management
M3W White Paper: System Integrity Management
12 Planning of future activities
The following ad hoc groups were established
8699 Ad Hoc Group on MAF Under Development in Systems
8698 Ad Hoc Group on MPEG File Formats
8697 Ad Hoc Group on Scene Representation
8507 AHG on 3DGC documents, experiments and software maintenance
8643 AHG on Audio Standards Maintenance
8644 AHG on Exploration of Speech and Audio Coding
8515 AHG on IPTV Requirements
8517 AHG on MAFs Awareness Event
10
8441 AHG on Maintenance of MPEG-4 Visual related Documents, Reference Software and
Conformance
8443 AHG on Maintenance of MPEG-7 Visual related Documents, Reference Software and
Conformance
8606 AHG on MDS MAFs Under Development
8645 AHG on MPEG Surround Verification Test and SAOC CfP
8605 AHG on MPEG-21 DIS
8520 AHG on MPEG-4 Part 9: Reference Hardware Description Phase 1 and 2.
8516 AHG on MPEG-7 Query Formats
8444 AHG on MPEG-7 Visual and Photo Player MAF
8442 AHG on Reconfigurable Video Coding
8589 AHG on SVC Verification Test
8514 AHG on the development of MPEG standards
8440 AHG on Video IDCT Specification
13 Resolutions of this meeting
These were approved
14 A.O.B
There was no other business
15 Closing
The meeting closed at 2006/10/27T20:50
11
Annex A – Attendance list
#
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Given Name
Ian
Gerrard
Michael
Christian
Dan
Jan
Davy
Saar
Jan
Bart
Silviu
Winger
Demin
Hua
Zhibo
Quqing
Chi-Cheng
Dandan
Zhengzhong
Yongying
Wen
Yun
Huan
Tiejun
Junyan
Xiangyang
Xin
Sixin
Jianguo
Jian
Hongfei
Zhibo
Honggang
Fang
Jianye
Lifeng
Li
Jing
Ting-hong
Ning
Jiangtao
Wei
Lianhuan
Xiaozhong
Family Name
Burnett
Drury
Ransburg
Timmerer
Cernea
De Cock
De Schrijver
De Zutter
Lievens
Masschelein
Simbotelecan
Lowell
Wang
Cai
Chen
Chen
Chu
Ding
Du
Gao
Gao
He
Hou
Huang
Huo
Ji
Jin
Lin
Liu
Lu
Ma
Ni
Qi
Qin
Rong
Song
Song
Wang
Wang
Wang
Wen
Xiao
Xiong
Xu
12
Country
Australia
Australia
Austria
Austria
Belgium
Belgium
Belgium
Belgium
Belgium
Belgium
Belgium
Canada
Canada
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
Jizheng
Lijing
Zhijie
Ping
Haitao
Lu
Pengxin
Yung-Chang
Shuhua
Jun
Xiaozhen
Lihua
Gang
Chris
Tanya
Miroslaw
Leszek
Catherine
Mike
Robert
Niels
Ping
Ying
Justin
Kemal
Ye-Kui
Olivier
Abdellatif
Vincent
Arnaud
Sébastien
Nathalie
Philippe
Sylvain
Jean-Claude
Marc
Patrick
Marc
Jean
Francois-xavier
Patrice
Stephane
Pierrick
Marius
Francoise
Jerome
David
Titus
Peter
Xu
Xu
Yang
Yang
Yang
Yu
Zeng
Zhang
Zhang
Zhang
Zheng
Zhu
Zhu
Barlas
Beech
Bober
Cieplinski
Grant
Nilsson
O'Callaghan
Rump
Wu
Chen
Ridge
Ugur
Wang
Avaro
Benjelloun Touimi
Bottreau
Bourge
Brangoulo
Cammas
De Cuetos
Devillers
Dufourd
Emerit
Gioia
Guez Vucher
Le Feuvre
Nuttall
Onno
Pateux
Philippe
Preda
Preteux
Vieron
Virette
Zaharia
Amon
13
China
China
China
China
China
China
China
China
China
China
China
China
China
England
England
England
England
England
England
England
England
England
Finland
Finland
Finland
Finland
France
France
France
France
France
France
France
France
France
France
France
France
France
France
France
France
France
France
France
France
France
France
Germany
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
Gero
JOhennaes
Bernhard
Harald
Bernhard
Oliver
Juergen
Steffen
Stefan
Tilman
Peter
Detlev
Matthias
Oliver
Tobias
Jens-Rainer
Joern
Thomas
Andreas
Florian
Heiko
Ralph
Chun Hui
Herbert
Thomas
Thomas
Mathias
Steffen
Ingo
Alex
W.H.A. (Fons)
Jean H.A.
Johan
Werner
Jan
Pierfrancesco
Maurizio
Filippo
Leonardo
Giovanni
Diego
Livio
Massimo
Kohtaro
Yukihiro
Takeshi
Toshiaki
Satoshi
Takashi
Base
Boehm
Feiten
Fuchs
Grill
Hellmuth
Herre
Kamp
Kraegeloh
Liebchen
List
Marpe
Narroschke
Niemeyer
Oelbaum
Ohm
Ostermann
Rathgen
Schneider
Schreiner
Schwarz
Sperschneider
Suen
Thoma
Wedi
Wiegand
Wien
Wittmann
Wolf
Eleftheriadis
Bruls
Gelissen
Muskens
Oomen
Vander Meer
Bellini
Campanai
Chiariglione
Chiariglione
Cordara
Gibellino
Lima
Mancin
Asai
Bandoh
Chujoh
Fujii
Hasuo
Ito
14
Germany
Germany
Germany
Germany
Germany
Germany
Germany
Germany
Germany
Germany
Germany
Germany
Germany
Germany
Germany
Germany
Germany
Germany
Germany
Germany
Germany
Germany
Germany
Germany
Germany
Germany
Germany
Germany
Germany
Greece
Holland
Holland
Holland
Holland
Holland
Italy
Italy
Italy
Italy
Italy
Italy
Italy
Italy
Japan
Japan
Japan
Japan
Japan
Japan
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
Satoshi
Itaru
Hideaki
Takahiro
Takuyo
Mayumi
Takehiro
Tomokazu
Tokumichi
Joji
Sei
Hiroya
Harada
Toshiyuki
Takeshi
Ryoma
Yukiko
Hideki
Satoru
Shun-ichi
Takanori
Masato
Kazuhiro
Yoshinori
Toshiyasu
Teruhiko
Seishi
Thiow Keng
Masayuki
akiyuki
Zhixiong
John
Akio
Yoshihisa
Tomoyuki
Yoshiyuki
Hyunsoo
Chang Beom
Jeong-Hwan
Seong Seon
Seungkwon
Hyouk Jean
Eun-Young
Maeng Sub
A-Young
Young-Hoon
Dae-Sung
Haechul
Byeongho
Ito
Kaneko
Kimata
Kimoto
Kogure
Koike
Moriya
Murakami
Murakami
Naito
Naito
Nakamura
Noboru
Nomura
Norimatsu
Oami
Ogura
Ohtaka
Sakazume
Sekiguchi
Senoh
Shima
Shimauchi
Sugihara
Sugio
Suzuki
Takamura
Tan
Tanimoto
Tanizawa
Wu
Wus
Yamada
Yamada
Yamamoto
Yashima
Ahn
Ahn
Ahn
Baek
Beack
Cha
Chang
Cho
Cho
Cho
Cho
Choi
Choi
15
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Japan
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
Kwonyul
Won Jun
Sang Bae
Hyon-Gon
Sung Moon
KangJae
Minsoo
Jong-Ki
Mahnjin
Yo-Sung
Park
Min Cheol
Jae-Ho
Chi Jung
Euee.S.
Dalwon
Byeong-Moon
Yongjoon
Dong Seok
Seyoon
Jie
Sung Ho
Jae Bum
Jaewoo
Yang-Won
Jung Won
SangWon
Mun Churl
Hae Kwang
Hyungyu
Hyun Mun
Jae-Gon
Dae Yeon
Dong kyun
DaeYeon
Hui Yong
Jong Lak
Tae Hyun
Kibeom
JungHoe
Kwangki
Kyuheon
Minsoo
Sangmi
SungMin
Tae Hyeon
Je Woo
MiJung
Sang-Kyun
Choi
Choi
Chon
Choo
Chun
Chung
Hahn
Han
Han
Ho
Hochong
Hong
Hur
Hwang
Jang
Jang
Jeon
Jeon
Jeong
Jeong
Jia
Jin
Jun
Jung
Jung
Kang
Kang
Kim
Kim
Kim
Kim
Kim
Kim
Kim
Kim
Kim
Kim
Kim
Kim
Kim
Kim
Kim
Kim
Kim
Kim
Kim
Kim
Kim
Kim
16
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
Yong-Hwan
Han-Suh
Young Seok
Kyung Jun
Yong Hun
Bae-Keun
Bumshik
Sunil
SangHeon
Sinwook
Yung Lyul
Alex Chungku
Hyunkook
James
Sangyoun
Sang Rae
Sunyoung
Young-Kwon
Jeongyeon
SungChang
Jaehyun
Taebeom
Sangil
Jung-hak
Kwan-Jung
Weon Geun
Eunmi
Henney
Jae Yul
Seoung-Jun
Hee-Suk
SeungWook
Seanae
Min Woo
Gwang-Hoon
SooJun
SungHo
DongHwan
Hee-Cheol
Jungdong
Chan-Won
Jeongil
Woo-Sung
Donggyu
Giseok
Youngjoo
Won Seon
Doug Young
Hendry
Kim
Koo
Koo
Lee
Lee
Lee
Lee
Lee
Lee
Lee
Lee
Lee
Lee
Lee
Lee
Lee
Lee
Lim
Lim
Lim
Lim
Lim
Na
Nam
Oh
Oh
Oh
Oh
Oh
Oh
Pang
Park
Park
Park
Park
Park
Park
Park
Seo
Seo
Seo
Seo
Shim
Sim
Son
Song
Song
Suh
Tan
17
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
Truong Cong
Jungyoup
Jeong-Hyu
Sehoon
Joung
Chang Dong
Won-Young
Kyoungro
Gisle
Peder
Arild
Thomas
Pereira
Lekha
Kok Seng
Haibin
Zhengguo
Te
Chong Soon
Mike
Zhongkang
Susanto
Qibin
Yih Han
Jo Yew
Wei
JunLi
Marc
Francisco
Kenneth
Kristofer
Frojdh
Heiko
Jonas
Anisse
Touradj
Marco
Vetro
Yiliang
Lazar
Marina
Lulin
Yi-Jen
Harinath
Onur
Munsi
Barry
Zhongli
Arianne
Thang
Yang
Yang
Yea
YeSun
Yoo
Yoo
Yoon
Bjontegaard
Drege
Fuldseth
Skjoelberg
Fernando
Chaisorn
Chong
Huang
Li
Li
Lim
Loh
Lu
Rahardja
Sun
Tan
Tham
Yao
Yuan
Gauvin
Morán Burgos
Andersson
Kjoerling
Per
Purnhagen
Roeden
Taleb
Ebrahimi
Mattavelli
Anthony
Bao
Bivolarski
Bosi
Chen
Chiu
Garudadri
Guleryuz
Haque
Haskell
He
Hinds
18
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Korea
Norvegia
Norvegia
Norvegia
Norvegia
Portugal
Singapore
Singapore
Singapore
Singapore
Singapore
Singapore
Singapore
Singapore
Singapore
Singapore
Singapore
Singapore
Singapore
Singapore
Spain
Spain
Sweden
Sweden
Sweden
Sweden
Sweden
Sweden
Switzerland
Switzerland
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
Michael
Shih-Ta
Faisal
Michael
Marta
Arkady
Gwo Giun
Shawmin
Vladimir
He-Yuan
James
Ning
Sam
Wen-Hsiao
Yingyong
Schuyler
Majid
Shankar
Yuriy
Jesus
Andrew
Rane
David
Yeping
Gary
Shijun
Huifang
Peter
Andrew
Pankaj
Alexandros
Chun-Jen
Yi-Shin
Xianglin
Chang
Wang
Yong
Yan
Peng
Haoping
Weimin
Minhua
Horowitz
Hsiang
Ishtiaq
Isnardi
Karczewicz
Kopansky
Lee
Lei
Levantovsky
Lin
Liu
Lu
Narasimhan
Peng
Qi
Quackenbush
Rabbani
Regunathan
Reznik
Sampedro
Segall
Shantanu
Singer
Su
Sullivan
Sun
Sun
Symes
Tescher
Topiwala
Tourapis
Tsai
Tung
Wang
Wo
Xin
Yan
Ye
Yin
Yu
Zeng
Zhou
19
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
United States
Annex B – Agenda
1.
2
3
4
5
6
7
8
1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
2
1
2
3
1
2
3
4
4
1
2
5
6
7
1
2
3
1
2
Agenda item
Opening
Roll call of participants
Approval of agenda
Allocation of contributions
Communications from Convenor
Report of previous meeting
Processing of NB Position Papers
Work plan
Media coding
Fixed point implementation of DCT/IDCT
Colour spaces
Colour spaces
Colour spaces
Advanced 4:4:4 Profile
Scalable Video Coding
Multiview Video Coding
Auxiliary Video Data Representation
BSAC Extensions
Symbolic Music Representation
Geometry and Shadow
XML Binarisation
Reconfigurable Video Coding
Increased video compression
Scalable audio and speech coding
Composition coding
Lightweight Scene Representation
Symbolic Music Representation
Description Coding
Schema definition
Visual Descriptor Extensions
Technologies for digital photo mgmt using MPEG-7 visual tools
MPEG-7 Query Format
Systems support
Fragments Request Unit
JPEG2000 support in MPEG-4 Systems
IPMP
MPEG-21 IPMP Component Base Profile
REL Profiles - the DAC profile
Rights Data Dictionary
Digital Item
Dynamic and Distributed Adaptations
Transport and File Format
20
1
2
3
4
5
6
7
8
8
1
2
3
4
5
6
7
9
1
2
3
4
5
6
10
1
2
3
4
5
6
7
8
9
10
11
11
1
2
3
4
5
6
7
8
9
10
11
Transport of Auxiliary Video Data
Transport of MPEG Surround data in AAC
File Format extensions for Description of Timed Metadata
Flute Hint Track
AVC File Format extensions for FRExt
AVC File Format extensions for SVC
File Format Issues for Support of Audio Media
Digital Item Streaming
Multimedia architecture
M3W Architecture
M3W Multimedia API
M3W Component Model
M3W Resource and Quality Management
M3W Component Download
M3W Fault Management
M3W System Integrity Management
Application formats
Protected Music Player MAF
Photo Player MAF
Musical Slide Show MAF
Media Streaming MAF
Audio Archival MAF
MAFs under consideration
Reference implementation
MPEG Surround Reference Software
File Format Reference Software
Morphing & Textures Reference Software
MPEG-J GFX Reference Software
Reference Hardware Description
MPEG-7 Systems Reference Software
Perceptual 3D Shape Reference Software
MPEG-21 REL Reference Software
MPEG-21 DIA Reference Software
Binary MPEG format for XML Reference Software
M3W Reference Software
Conformance
Audio BIFS v3 Conformance
MPEG-1 and -2 Audio in MPEG-4 Conformance
BSAC conformance
1-bit Oversampled Audio Conformance
Audio Lossless Conformance
Audio Scalable to Lossless conformance
MPEG Surround conformance
File Format conformance
Morphing & Textures Conformance
Advanced Text and Graphics Conformance
Synthesized Texture Conformance
21
12
13
14
15
16
17
18
12
1
2
3
4
5
6
9
10
11
12
13
14
15
MPEG-J GFX Conformance
Open Font Format conformance
Perceptual 3D Shape Conformance
IPMP Components Conformance
Event Reporting Conformance
Fragment Identification of MPEG Resources Conformance
Music Player Application Format Conformance
Maintenance
Systems coding standards
Video coding standards
Audio coding standards
Visual description coding standards
Audio description coding standards
MDS standards
Liaison matters
Organisation of this meeting
Tasks for subgroups
Joint meetings
Development of MPEG standards
Administrative matters
Schedule of future MPEG meetings
Promotional activities
Planning of future activities
Resolutions of this meeting
A.O.B
Closing
22
Annex C – Input contributions
No.
Authors
Title
13744 Wo Chang
Document Register for SC29/WG11 Meeting Hangzhou,
China
Marius Preda
Jeong-Hwan Ahn
13745
Francisco Morán
Vishy Swaminathan
AHG on 3DGC documents, experiments and software
maintenance
Marco Mattavelli
G. Sullivan
13746 A. Hinds
Y. Reznik
P. Topiwala
AHG on Video IDCT Specification
13747
Yi-Shin Tung
Chung-Neng Wang
AHG on Maintenance of MPEG-4 Visual related
Documents, Reference Software and Conformance
13748
Euee S. Jang
Yoshihisa Yamada
AHG on Reconfigurable Video Coding
Sang-Kyun Kim
13749 Robert O'Callaghan
Akio Yamada
AHG on Maintenance of MPEG-7 Visual related
Documents, Reference Software and Conformance
Miroslaw Bober
Sang-Kyun Kim
13750
Akio Yamada
Wo Chang
AHG on MPEG-7 Visual and Photo Player MAF
13751
Robert Turney
Marco Mattavelli
AHG on MPEG-4 Part 9 Reference Hardware Description
Phase 2 and 3
13752 R. Sperschneider
AHG on Audio Standards Maintenance
13753 S. Quackenbush
AHG on Exploration of Audio Spatialization and Speech
and Audio Coding
13754
Gerrard Drury
Peder Drege
Stefan Kraegeloh
13755 Filippo Chiariglione
Noboru Harada
AHG on MPEG-21 DIS
AHG on MAFs Under Development
13756
Young-Kwon Lim
Cyril Concolato
AHG on Scene Representation
13757
David Singer
Visharam Mohammed
AHG on MPEG File Formats
Chris Barlas
13758 Takuyo Kogure
Andy Tescher
AHG on the Development of MPEG standards
James A.G. Annesley
James Orwell
13759
Jim Aldridge
Kate Grant
AHG on Surveillance MAF
13760
H. Jean Cha
Herbert Thoma
AHG on Portable Video Player MAF
23
13761 SC 29 Secretariat
Summary of Voting on ISO/IEC 14496-20:2006/DCOR 1
[SC 29 N 7729]
13762 SC 29 Secretariat
Summary of Voting on ISO/IEC 13818-2:2000/FPDAM 2
[SC 29 N 7736]
13763 SC 29 Secretariat
Summary of Voting on ISO/IEC 14496-2:2004/FPDAM 3
[SC 29 N 7737]
13764 SC 29 Secretariat
Summary of Voting on ISO/IEC 14496-4:2004/FPDAM
12 [SC 29 N 7738]
13765 SC 29 Secretariat
Summary of Voting on ISO/IEC 14496-5:2001/FPDAM 9
[SC 29 N 7739]
13766 SC 29 Secretariat
Summary of Voting on ISO/IEC 14496-10:2005/FPDAM
1 [SC 29 N 7740]
13767 SC 29 Secretariat
Summary of Voting on ISO/IEC 15938-7:2003/FPDAM 2
[SC 29 N 7741]
13768 SC 37 via SC 29 Secretariat
Liaison Statement from SC 37/WG 3 [SC 29 N 7742]
13769 SC 29 Secretariat
Summary of Voting on ISO/IEC 14496-3:2005/PDAM 6
[SC 29 N 7748]
13770 SC 29 Secretariat
Summary of Voting on ISO/IEC 14496-12:2005/FPDAM
1 and ISO/IEC 15444-12:2005/FPDAM 1 [SC 29 N 7749]
13771 ITTF via SC 29 Secretariat
Table of Replies on ISO/IEC 15938-4:2002/FDAM 2 [SC
29 N 7751]
13772 SC 29 Secretariat
Summary of Voting on ISO/IEC 14496-11:2005/FPDAM
5: [SC 29 N 7752]
13773 Jean-Claude Dufourd
Draft study of DCOR1, on going
13774 Jean-Claude Dufourd
Draft study of FPDAM1, on going
13775 Jean-Claude Dufourd
Time encoding issues for discussion at Sophia AHG
13776 Jean-Claude Dufourd
Stream management issues for discussion at Sophia AHG
13777 Jean-Claude Dufourd
Events and animation issues for discussion at Sophia AHG
13778 Jean-Claude Dufourd
Update extensibility issues for discussion at Sophia AHG
13779 Jean-Claude Dufourd
mini2 improvements for discussion at Sophia AHG
13780
ITU-R SG 6/WP 6J via SC 29
Secretariat
Liaison Statement from ITU-R SG 6/WP 6J [SC 29 N
7764]
13781 Jean-Claude Dufourd
On SMIL MediaClipping in LASeR for discussion at
Sophia AHG
13782 Jean-Claude Dufourd
On fullscreen video for discussion at Sophia AHG
13783 Jean-Claude Dufourd
On the usage of the a element in LASeR for discussion at
Sophia AHG
13784 Yuriy Reznik
updated IDCT algorithm for CD selection
13785 Jean Le Feuvre
On LASeR Conditional Execution
13786 Jean Le Feuvre
On SAF Configuration
13787 Jean Le Feuvre
Comments on LASeR and SAF DCOR
13788 Jean Le Feuvre
On LASeR animateScroll
13789 Cyril Concolato
On SAF global streams
24
Jean Le Feuvre
13790
Cyril Concolato
Jean Le Feuvre
On LASeR Waiting Tree
13791 Lazar Bivolarski
Updated Connex Proposal of Low Complexity IDCT for
CD Selection
13792 IEC TC 100 via SC 29 Secretariat
IEC CDV 62455 [SC 29 N 7769]
13793 SC 29 Secretariat
Summary of Voting on ISO/IEC CD 21000-14.2 [SC 29 N
7777]
13794 Young-Kwon Lim
Report of LASeR AHG meeting in Sophia
13795
Choudhury A. Rahman
Wael Badawy
A HW Block for H.264/AVC Context Adaptive Variable
Length Coding (CAVLC)
13796 SC 29 Secretariat
Summary of Voting on ISO/IEC 14496-20:2006/FPDAM
1 [SC 29 N 7781]
Honggang Qi
Wen Gao
13797
Debin Zhao
Siwei Ma
AAN IDCT Design for CD Selection
13798 W3C via SC 29 Secretariat
Liaison Statement from W3C [SC 29 N 7782]
13799 Trac D. Tran
FastVDO IDCT proposal for CD
Navarro
13800 Reznik
Silva
Improved IDCT
13801 Jean-Claude Dufourd
Draft LASeR 2nd edition (DCOR1 + FPDAM1)
13802 SC 29 Secretariat
Summary of Voting on ISO/IEC 21000-7:2004/FPDAM 2
[SC 29 N 7784]
Navarro
13803 Reznik
Silva
Improved IDCT- Replacing M13800
13804 Arianne T. Hinds
Updated MPEG-4 testbed
13805 Zhibo Ni
Updated MPEG-2 testbed
13806 Yuriy Reznik
Updated H.263+ testbed
13807 Tanya Beech
Proposal for improvements to Geographic Position in
Mpeg7 Part 5
13808
ITU-R SG 6/WP 6Q via SC 29
Secretariat
Liaison Statement from ITU-R SG 6/WP 6Q [SC 29 N
7794]
13809 3GPP2 via SC 29 Secretariat
Liaison Statement from 3GPP2 [SC 29 N 7795]
13810 SC 29 Secretariat
Summary of Voting on ISO/IEC FCD 21000-18 [SC 29 N
7802]
13811 ITTF via SC 29 Secretariat
Table of Replies on ISO/IEC FDIS 23002-1 [SC 29 N
7819]
13812 SC 29 Secretariat
Summary of Voting on ISO/IEC FCD 23000-3
13813 SC 29 Secretariat
Summary of Voting on ISO/IEC 14496-1:2004/DCOR 2
13814 SC 29 Secretariat
Summary of Voting on ISO/IEC 138184:2004/Amd.2:2005/DCOR 1
25
13815 SC 29 Secretariat
Summary of Voting on ISO/IEC 13818-7:2006/PDAM 1
13816 SC 29 Secretariat
Summary of Voting on ISO/IEC 144964:2004/Amd.11:2006/DCOR 2
13817 SC 29 Secretariat
Summary of Voting on ISO/IEC 14496-4:2004/PDAM 14
13818 SC 29 Secretariat
Summary of Voting on ISO/IEC 14496-4:2004/PDAM 18
13819 SC 29 Secretariat
Summary of Voting on ISO/IEC 14496-4:2004/PDAM 19
13820 SC 29 Secretariat
Summary of Voting on ISO/IEC 14496-4:2004/PDAM 20
13821 SC 29 Secretariat
Summary of Voting on ISO/IEC 14496-15:2004/PDAM 2
13822 SC 29 Secretariat
Summary of Voting on ISO/IEC 14496-18:2004/DCOR 1
13823 SC 29 Secretariat
Summary of Voting on ISO/IEC 159383:2002/Amd.1:2004/DCOR 2
13824 SC 29 Secretariat
Summary of Voting on ISO/IEC 15938-6:2003/PDAM 2
13825 SC 29 Secretariat
Summary of Voting on ISO/IEC 15938-7:2003/PDAM 3
13826 SC 29 Secretariat
Summary of Voting on ISO/IEC TR 15938-8:2002/PDAM
3
13827 SC 29 Secretariat
Summary of Voting on ISO/IEC 15938-10:2005/DCOR 1
13828 SC 29 Secretariat
Summary of Voting on ISO/IEC 14496-12/PDAM 2 &
15444-12/PDAM 2
13829 SC 29 Secretariat
Summary of Voting on ISO/IEC 21000-4:2006/PDAM 1
13830 SC 29 Secretariat
Summary of Voting on ISO/IEC CD 23000-2 [2nd
Edition]
13831 SC 29 Secretariat
Summary of Voting on ISO/IEC CD 23000-4
13832 SC 29 Secretariat
Summary of Voting on ISO/IEC 23001-1:2006/DCOR 1
13833 SC 29 Secretariat
Summary of Voting on ISO/IEC CD 23004-5
13834 SC 29 Secretariat
Summary of Voting on ISO/IEC CD 23004-6
13835 SC 29 Secretariat
Summary of Voting on ISO/IEC CD 23004-7
Jaime Delgado
13836 Eva Rodriguez
Marc Gauvin
Final Comments on the Ontological Analysis of the Study
of DCOR/2 of ISO/IEC 21000-6
13837 Francisco Morán
ESNB position paper: Problems with the inconsistency of
the MPEG-21 Rights Data Dictionary
13838
Haque
Sanjay
Use of UPnP Content Directory Service (CDS) in MP7QF
13839
Shinjun Lee
Jeong-Hwan Ahn
Results of Exploration Experiments (EE1: Static and
Animated 3D Object Compression)
13840 SMPTE
SMPTE Liaison to JTC1/SC29 - SMPTE 421M ISO Base
Media File Format
13841 SMPTE
SMPTE Liaison to JTC1/SC29 Constraints on high profile
13842 SMPTE
SMPTE Liaison to JTC1/SC29 New profile for production
13843 Arnaud Bourge
Proposed WD for ISO/IEC 23002-3 Conformance
13844
Trac D. Tran
Lijie Liu
High-Performance Low-Complexity Dyadic Re-Sampling
Filters for SVC
26
Pankaj Topiwala
Trac D. Tran
13845 Lijie Liu
Pankaj Topiwala
Core Experiments for Re-Sampling Filters in SVC
Trac D. Tran
13846 Lijie Liu
Pankaj Topiwala
FastVDO 16-bit IDCT Proposal for CD: Performance and
Comparison
Trac D. Tran
13847 Lijie Liu
Pankaj Topiwala
Core Experiments for IDCT
13848 David Singer
Updated ISO Base Media File Format Reference Software
13849 David Singer
Updated MP4 Conformance Files from Apple
13850 David Singer
An improved rate-share algorithm for the ISO Base file
format
13851 David Singer
Comments and suggestions on the SVC File Format draft
13852 David Singer
SMPTE KLV meta-data in ISO Base Media File format
files
Pierfrancesco Bellini
13853 Maurizio Campanai
Paolo Nesi
Editors' Study of ISO/IEC 14496-3/PDAM6
Pierfrancesco Bellini
13854 Mauzio Campanai
Paolo Nesi
Editors' Study of ISO/IEC 14496-11/FPDAM5
Per Fröjdh
Thorsten Lohmar
13855
Miska Hannuksela
Imed Bouazizi
Study on 14496-12:2005/PDAM2 ALC/FLUTE server file
format
Hae Kwang Kim (Sejong University)
13856 B.S Manjunath(UCSB)
Weon Geun Oh (ETRI)
Image and video signature techniques
13857 Korean National Body
Request on MAF standardization for DMB
13858 Korean National Body
A T-DMB White Paper and a Introductory Movie (6
minutes)
Munchurl Kim
Jeongyeon Lim
Hui Yong Kim
13859 Hyon-Gon Choo
Yong Han Kim
Jinhan Kim
Sung Ho Jin
Requirements for DMB MAF
Munchurl Kim
Jeongyeon Lim
Hui Yong Kim
13860 Hyon-Gon Choo
Yong Han Kim
Jinhan Kim
Sung Ho Jin
Proposal for DMB Multimedia Application Format
13861
Weon-Geun Oh
Eun-Ku Jung
A Test Image Management System for MPEG-7 core
experiments
27
Hae-Kwang Kim
Mayumi Koike
13862 Takuyo Kogure
Hiroshi Yasuda
Adaptation to MPEG MAF of Digital Video/Cinema file
format
Mayumi Koike
13863 Takuyo Kogure
Hiroshi Yasuda
Requirement of Color Management Information to
MPEG-7 for Digital Video/Cinema
Marcos Avilés
13864 Francisco Morán
Marius Preda
Implementation of JPEG 2000 elementary stream support
in MPEG-4 reference software
13865 Andy Tescher for USNB
USNB Contribution: SMPTE ARO Position (Concurrently
submitted to WG 1)
13866 Andy Tescher for USNB
USNB Contribution: Speech and Audio Coding
Exploration Support
Sangki Kim
Sangyoun Lee
13867
Myung Gil Jang
Jeong Hur
CE Report for VCE-5
Weon-Geun Oh
Hyeong-yong Jeon
13868 Jung-Sub Shin
Chi-Jung Hwang
Maeng-Sub Cho
An Image Identifier Based on Singular Value
Decomposition and Feature Point
Ik-Hwan Cho
Seok-Kyoo Shin
13869
Weon Geun Oh
Dong-Seok Jeong
The Category and Complexity based Test Image
Extraction Method on MPEG-7 VCE-6
Soo-Jun Park
13870 Sung Min Kim
Chee Sun Won
Proposal for a new MPEG-7 input query format: Queryby-Layout
13871
Soo-Jun Park
Seon Hee Park
Report of Core Experiment: VCE-3 - Person-Identitybased clustering, indexing and retrieval of images
13872
Soo-Jun Park
Seon Hee Park
Dataset for VCE-3 by ETRI, Version3
13873 Ryoma Oami
A proposal for a referencing mechanism of person
information for MPEG-A Photo Player
13874 Tobias Oelbaum
The SVT High Definition Multi Format Test Set
Kyoungro Yoon
Hee-Cheol Seo
13875
Hyunki Kim
Myung-Gil Jang
Comparison of MP7QF Requirements and TV-Anytime
Technology
Hee-Cheol Seo
Hyunki Kim
13876
Myung-Gil Jang
Kyoungro Yoon
Comparison of XQuery and MPEG-7 Query Format
13877 Jean Le Feuvre
On AAC SBR storage in ISO Media File
13878
Jean Le Feuvre
Jean-Claude Dufourd
Discussion on SAF global streams
13879
Jean Le Feuvre
Cyril Concolato
On LASeR Fraction events
28
Noboru Harada
13880 Takehiro Moriya
Yutaka Kamamoto
Proposed text to MPEG-4 audio extensions for 64-bit
address space file format support
Noboru Harada
13881 Takehiro Moriya
Yutaka Kamamoto
Proposed text for WD of Audio Archival MAF
Weon-Geun Oh
Chi-Jung Hwang
13882
Dong-Seok Jeong
Hae-Kwang Kim
Request of Amendment in VCE-6 Specifications
Adrian Munteanu none Maryse Stoufs
Error-resilient profile for MeshGrid: robust encoding of
13883 none Alin Alecu none Jan Cornelis none
the reference-grid
Peter Schelkens
Thomas Skjølberg
Peder Drege
13884
Gerrard Drury
Joseph Thomas-Kerr
Report of CE on DIS TuC
13885 FLO Folum via SC 29 Secretariat
Liaison Statement from FLO Folum [SC 29 N 7821]
13886
Stephen Davis
Gerrard Drury
Editors input on 23001-2 FRU
13887
Donggyu Sim isha1012@kw.ac.kr
SueKyung Park
Face detection
Khaled Mamou
Titus Zaharia
13888
Marius Preda
Françoise Prêteux
Results of evaluation experiment EE1 on static and
animated 3D mesh coding : skinning-based dynamic mesh
compression
Hui Yong Kim
Jeong Hyun Yoon
Hee Kyung Lee
13889 Han Kyu Lee
Sung Ho Jin
Jae-Seok Jang
Yong Man Ro
Requirements for DTV MAF
Sang-Kyun Kim
13890 Yong-Ju Jung
Yong Man Ro
CE Report on Person-Identity based photo clustering and
indexing (VCE-3)
13891
Sang-Kyun Kim
Ryong Lee
Request for adding Query Requirements related to data
manipulation against MPEG-7 DB on current MPEG-7
Query Format Requirement
Kisong Yoon(ETRI)
Taehyun Kim(DRM inside)
13892 Eva Rodriguez(DMAG-UPC)
Jaime Delgado(DMAG-UPC)
Hogab Kang(DRM inside)
Proposed MPEG-21 REL Open Release Profile
Weon-Geun Oh
A-Young Cho
13893 Ik-Hwan Cho
Jun-Woo Lee
Dong-Seok Jeong
VCE-6 Results for Non-geometric Modification
Masayuki Tanimoto
13894 Toshiaki Fujii
Shigeyuki Sakazawa
Requirements on Free Viewpoint Television (FTV) v.0
29
Hideaki Kimata
13895 Jean Le Feuvre
On SAF streams redefinition
YeSun Joung
Young-kwon Lim
13896 Won-sik Cheong
Jihun Cha
KyungAe Moon
Implementation of LASeR uDOM Interface in LASeR
Player
YeSun Joung
Young-kwon Lim
13897 Won-Sik Cheong
Jihun Cha
KyungAe Moon
An exploration on MPEG-21 and LASeR
Youngjoo Song
13898 Young-Kwon Lim
Jechang Jeong
Improved text for GroupingDescriptor
Seungkwon Beack
Jeongil Seo
13899 Taejin Lee
Inseon Jang
Dae-young Jang
Further information of a new application for SAOC
JungHoe Kim
Eunmi Oh
Proposed updates on SLS reference software with ER
BSAC
JungHoe Kim
13901 KangEun Lee
Eunmi Oh
Proposed study on 14496-4:2004/FPDAM 14, BSAC
Conformance
13900
13902
JungHoe Kim
Eunmi Oh
Thomas Ragthen
13903 Peter Amon
Andreas Hutter
13904
JungHoe Kim
Eunmi Oh
Thomas Rathgen
13905 Peter Amon
Andreas Hutter
13906
Eva Rodríguez
Jaime Delgado
Proposed changes for BSAC Extensions combined with
MPEG Surround
Comments on the AVC file format PDAM2 document
Proposed residual coding with ER BSAC for MPEG
Surround
Improvements of SVC file format meta data statements
IPMP and the Surveillance MAF
Euee S. Jang
Sunyoung Lee
Alex Chungku Yie
13907
Eunkyung Kwak
James S.G. Yoo
Rana Lee
Reshaping Digital Media Business Models by
Reconfigurable Video Coding
Sunyoung Lee
Hyungyu Kim
Hyunsoo Ahn
Sinwook Lee
13908
Jaebum Jun
Giseok Son
Chungku Yie
Euee S. Jang
Proposed Updates of RVC Working Draft 1.0
30
Hyungyu Kim
Sunyoung Lee
Hyunsoo Ahn
Sinwook Lee
13909
Jaebum Jun
Giseok Son
Chungku Yie
Euee S. Jang
RVC CE1 : RVC based Inter Coding Implementaion
Hyunsoo Ahn
Sunyoung Lee
Hyungyu Kim
Sinwook Lee
13910
Jaebum Jun
Giseok Son
Chungku Yie
Euee S. Jang
Modified Decoder Description for Scheduling over RVC
Framework
Eun-Young Chang
JaeBum Jun
Sinwook Lee
13911 Namho Hur
Jinwoong Kim
Soo In Lee
Euee S. Jang
Comments on the inclusion of 3DMC-Extension in Part 11
Scene description and application engine
13912
Zhibo Ni
Lu Yu
Drift Problem of Fixed-Point IDCT on News Sequence
13913 Harald Fuchs
Study Text on ISO/IEC CD 23000-2 MPEG-A Music
Player 2nd edition
Dandan Ding
13914 Zhibo Ni
Lu Yu
Analysis of Hardware Implementation Cost of Fixed-Point
IDCT
Chun Hui Suen
13915 Florian Schreiner
Klaus Diepold
File Format and Event Reporting for Open Release MAF
Zhibo Ni
13916 Cixun Zhang
Lu Yu
Test Results for Technical Selection of Committee Draft
of ISO/IEC 23002-2 Fixed-Point IDCT
13917
Ralph Sperschneider
Michael Matejko
Conformance issues regarding AAC utilizing the LTP tool
13918 Giovanni Cordara
Late comment on ISO/IEC 14496-3 PDAM.6 and
ISO/IEC 14496-11 FPDAM.5
Sung-Wen Wang
Chung-Yi Weng
13919
Yi-Shin Tung
Wei-Kai Steve Su
RVC CE2: Extensibility of FUs and Interfaces between
CAL and C++
13920 Jean Le Feuvre
On LASeR Events
13921 S. Quackenbush
77th MPEG Audio Report
13922 S. Quackenbush
78th MPEG Audio Tasks
Kristofer Kjörling
Jonas Rödén
13923
Heiko Purnhagen
Werner Oomen
Further revision of the verification test proposal for
MPEG Surround
31
Johannes Hilpert
13924 Heiko Purnhagen
13925
Heiko Purnhagen
Andreas Schneider
Update on reference software for MPEG Surround
Update on conformance testing for MPEG Surround
13926 Heiko Purnhagen
Update on transport of MPEG Surround
13927 Lu Yu
Anti-IDCT for IDCT Drift Test
13928
Xin Wang
Chris Barlas
Rights Enforceability in the Open Release MAF
13929 Xin Wang
Proposal for Working on an IPTV MAF
13930 CNNB
CNNB comments on the work of fixed-point 8x8 IDCT
transform
13931
Miska M. Hannuksela
Ye-Kui Wang
Track relationship in file format
13932
Ye-Kui Wang
Miska M. Hannuksela
Generic adaptation path in file format
13933
Ye-Kui Wang
Miska M. Hannuksela
Comments on SVC file format
13934 Honggang Qi
Crosscheck for proposal m13927
Juergen Herre
13935 Werner Oomen
Kristofer Kjoerling
Thoughts on an SAOC Architecture
13936
Paul Brasnett
Miroslaw Bober
Experimental results on an image identifier (VCE-6)
13937
Paul Brasnett
Miroslaw Bober
Experimental dataset for VCE-6
Kwan-Jung Oh
13938 Yo-Sung Ho
Byeongho Choi
View Interpolation for Multi-view Video Coding
Kwan-Jung Oh
Cheon Lee
13939
Pil-Kyu Park
Byeongho Choi
Global Disparity Compensation for Multi-view Video
Coding
Kwan-Jung Oh
13940 Cheon Lee
Pil-Kyu Park
Reconstruction of Reference Frames for Multi-view Video
Coding
13941 Honggang Qi
Test Results for Selection of Committee Draft of ISO/IEC
23002-2 Fixed-Point IDCT
Marco Mattavelli
13942 Jorn Janneck
Dave Parlour
Report on results of RVC CE 1.2 Formalize XML-based
description of configuration of FUs.
13943 AHG on MAFs Under Development
Proposal of Updated Working Draft of ISO/IEC 23000-5
Media Streaming Player
Marco Mattavelli
Joseph Thomas-Kerr
13944
Jorn Janneck
Dave Parlour
Report on results of RVC CE 1.1 Implement flexible FUs
according to the processing mechanism in RVC WD using
CAL.
13945 on MAFs Under Development
Proposal of Updated Working Draft of IPMP Extensions
32
XML Messages
13946 on MAFs Under Development
Proposal of Updated Working Draft of Media Streaming
MAF Technologies
Marco Mattavelli
Andrew Kinane
13947 Christophe Lucarz
Jorn Janneck
Dave Parlour
Report on results of RVC CE 2.1 Reshape the current
MPEG-4 SP CAL decoder according to the current FU
interface in RVC WM.
Marco Mattavelli
13948 Christophe Lucarz
Andrew Kinane
Report on results of RVC CE 2.2 Explore the extensibility
of FUs
Hyon-Gon Choo
13949 Filippo Chiariglione
Bum-Suk Choi
Proposed Working Draft of ISO/IEC 21000-4/Amd 2
Media Streaming Profile
Robert O'Callaghan
Miroslaw Bober
13950
Akio Yamada
Wo Chang
Editors' input: FDIS 23000-3 (Photo-Player MAF)
Robert O'Callaghan
Miroslaw Bober
13951
Sang-Kyun Kim
Akio Yamada
Editors' input: TR 15938-8 DAM3 (Technologies for
digital photo management)
13952 Robert O'Callaghan
Defect Report: ISO/IEC 15938-3 Amd.2 (Perceptual 3D
Shape Descriptor)
13953
Robert O'Callaghan
(on behalf of the UKNB)
UKNB comments on the text of ISO/IEC 15938-7
PDAM3 & 15938-6 PDAM2
13954
Robert O'Callaghan
(on behalf of the UKNB)
UKNB comments on the text of ISO/IEC TR 15938-8
PDAM3
Hendry
Munchurl Kim
13955 Sangjin Hahm
Keunsik Lee
Keunsoo Park
Proposed Extension to SVC File Format for Efficient and
Effective Protection
13956
Hendry
Takafumi Ueno
Editor’s Study of ISO/IEC 21000-4/PDAM 1: IPMP Base
Profile
13957
Hendry
Munchurl Kim
Contribution to ISO/IEC 21000-4/PDAM 1: IPMP Base
Profile Reference Software
Markus Schnell
Ralph Sperschneider
Markus Schmidt
13958 Juergen Herre
Ralf Geiger
Gerald Schuller
Manfred Lutzky
13959
Michael Ransburg
Hermann Hellwagner
13960 Patrick Gioia
13961
Patrick Gioia
Romain Cavagna
13962 Marius Preda
Proposal for an Enhanced Low Delay Coding Mode
Contribution to ISO Base Media File Format Reference
Software
Proposal for Large 3D Environments Profile
Proposal for Geometry Related Space Partitioning Streams
www.3DoD.org: an MPEG-4 3D Database
33
Son Tran
Duc Tran
Ivica Arsov
Francoise Preteux
Davy De Schrijver
Wesley De Neve
13963 Davy Van Deursen
Saar De Zutter
Rik Van de Walle
An MPEG-21 BS Schema for the scalable extension of
H.264/MPEG-4 AVC version 6 (Joint Scalable Video
Model 6)
Thomas Wedi
Hideki Ohtaka
13964
John Wus
Shun-ichi Sekiguchi
Intra-only H.264/AVC profiles for professional
applications
Chris Poppe
13965 Saar De Zutter
Rik Van de Walle
Contribution to Utility Software for ISO/IEC 21000-10
DIP/AMD 1
13966 Wo Chang
Proposed Medical Imaging MAF (MI MAF) for
Preserving Medical Imaging Records
13967 Wo Chang
MAF to Industry
Saar De Zutter
Frederik De Keukelaere
13968 Gerrard Drury
Christian Timmerer
Xin Wang
Editor’s input to ISO/IEC 21000-8 Reference Software
(Second Edition)
Jean-Claude Dufourd
Nicolas Pierre
13969 Elouan Le Coq
Cyril Concolato
Jean Lefeuvre
Final word on the encoding of times in LASeR
13970
Jean-Claude Dufourd
Nicolas Pierre
Elements for the clarification of the waiting tree concept
in LASeR
13971 Jean-Claude Dufourd
LASeR reference software release and status
Saar De Zutter
Sylvain Devillers
13972
Thomas DeMartini
Andrew Tokmakoff
Editor's input to ISO/IEC 21000-14 Conformance Testing
Jean-Claude Dufourd
Elouan Le Coq
On communication channels management with LASeR
13974 Jean-Claude Dufourd
Request for promotion of some TuC to LASeR AMD1
13973
13975
Jean-Claude Dufourd
Nicolas Pierre
Update of the proposed LASeR mini2 profile
13976
Jean-Claude Dufourd
Elouan Le Coq
On a new caching instruction for SAF
13977 Jean-Claude Dufourd
On a few missing fixes to LASeR docs
Saar De Zutter
13978 Davy De Schrijver
Rik Van de Walle
Update to Reference Software for Conformance to
ISO/IEC 21000-10
Saar De Zutter
13979 Chris Poppe
Davy De Schrijver
Update to Reference Software for Conformance to
ISO/IEC 21000-10/Amd 1
34
Rik Van de Walle
13980
Saar De Zutter
Rik Van de Walle
Contribution to summary and 1-pager of Enhanced
Interoperability for MPEG-21 Session Mobility using DIP
Saar De Zutter
13981 Davy De Schrijver
Rik Van de Walle
Update to summary of Digital Item Technologies: Digital
Item Processing
Saar De Zutter
Chris Poppe
13982
Davy De Schrijver
Rik Van de Walle
Contribution to summary and 1-pager of Digital Item
Technologies: Digital Item Processing Amd 1
Saar De Zutter
13983 Davy De Schrijver
Rik Van de Walle
Contribution to summary and 1-pager of Conformance:
MPEG-21 Digital Item Processing
Saar De Zutter
Chris Poppe
13984
Davy De Schrijver
Rik Van de Walle
Contribution to summary and 1-pager of Conformance:
MPEG-21 Digital Item Processing Amd 1
Saar De Zutter
13985 Davy De Schrijver
Rik Van de Walle
Update to summary and 1-pager of Reference Software:
MPEG-21
13986 Justin Ridge
Request new levels for MPEG-4 Simple Profile
Christian Timmerer
13987 Michael Ransburg
on behalf of the ANB
Austrian NB comments on ISO/IEC 21000-7 FPDAM
Michael Eberhard
13988 Michael Sablatschan
Christian Timmerer
gBSDtoBin (MPEG-21 DIA) reference software update
13989
Antonio Navarro
Marco Santos
Hardware implementation of full search H.264 motion
estimation
13990
Antonio Navarro
Antonio Silva
Performance in MPEG-4 of five submitted integer IDCTs
for CD
13991
Thomas Skjølberg
Peder Drege
Delivery of dynamic resources in Digital Item Streaming
13992 Antonio Navarro
Crosschecking an integer 16 bit IDCT (M13791)
13993 Lazar Bivolarski
On implementation of IDCTs on existing 16-bit
architectures
Truong Cong Thang
Tae Meon Bae
13994 Yong Man Ro
Jung Won Kang
Jae-Gon Kim
Mechanism of AR-FGS in Conditions of FGS Motion
Refinement
13995
H. Jean Cha
Tae Hyeon Kim
Refined requirements and technologies for Portable Video
Player MAF
13996
Honggang Qi
Arianne T. Hinds
On the Usage of High Precision IDCTs in Existing MPEG
Products
13997
Joanna J. Eastment
Arianne T. Hinds
On the Cost and Performance of IDCT Implementations in
Hardware
13998
H. Jean Cha
Tae Hyeon Kim
Proposed working draft of Portable Video Player MAF
35
13999 Arianne T. Hinds
Updated T.83 testbed for IDCT testing
14000 Lazar Bivolarski
On the Complexity Analysis of IDCT Algorithms for CD
Selection
14001
Tae Hyeon Kim
H. Jean Cha
Proposed timed text formt for Musical Slide Show MAF
14002
Tae Hyeon Kim
H. Jean Cha
Usage of the transition element for Musical Slide Show
MAF
14003 Antonio Navarro
Cross check of proposed additional (CE-stage) IDCT
designs
14004 Yuriy Reznik
On clipping and dynamic range of variables in IDCT
designs
Yuriy A. Reznik
Arianne T. Hinds
Cixun Zhang
Lu Yu
14005
Zhibo Ni
Lazar Bivolarski
Honggang Qi
Siwei Ma
Additional information on IDCT CD candidates and
proposed core experiments
14006 Yuriy Reznik
Examples of existing fixed-point IDCTs
14007 Wo Chang
Testing
14008 Marc Emerit
A survey of audio middleware parameters for Audio
Scene Control reusing MPEG Surround
14009
Pierrick Philippe
David Virette
Report on the pre-selection process for MPEG Surround
verification tests
Philippe de Cuetos
14010 Gregoire Pau
Cedric Thienot
Editor's Study of 23001-1 PDAM2
14011 Philippe de Cuetos
Fixes on LASeR Amd1
14012
Sylvain Devillers
Renaud Cazoulat
Use case and requirement for LASeR
14013
Sylvain Devillers
Renaud Cazoulat
New feature for LASeR
Weon Geun Oh
14014 Eun Ku Jung
Hae Kwang Kim
An Image Data Management System for MPEG-7 VCE-6
Jeff Z. Pan
14015 Raphaël Troncy
Yannis Avrithis
Liaison Statement from W3C MMSem-XG on Exploring
Opportunities for Cooperation
14016 ITU-T SG 9 via SC 29 Secretariat
Liaison Statement from ITU-T SG 9 [SC 29 N 7830]
14017 H. Jean Cha
Proposed Work Plan for Portable Video Player MAF
Jianguo Liu
Guoyou Wang
Shengkui Dai
14018
Pingping Zhu
Xinjian Meng
Jianhua Zheng
DSP implementations of 24-bit AAN algorithms
14019 Jianguo Liu
16-bit high precision scaled AAN for fixed-point IDCT
36
Guoyou Wang
Shengkui Dai
Pingping Zhu
Xinjian Meng
Jianhua Zheng
14020 Swiss National Body
14021
Marco Mattavelli
Jorn Janneck
Request of completing the editing of conformance
subclauses for the DTR 14496-9 2nd Edition
Proposition for update of the RVC WD
37
Annex D – Output documents
No.
Source
Title
8431 Convener
List of Documents from the Hangzhou, CN Meeting
8432 Convener
Resolutions of the Hangzhou, CN
8433 Convener
List of AHGs Established at the 78th Meeting in Hangzhou, CN
8434 Convener
Report of the 78th Meeting in Hangzhou, CN
8435 Convener
Guidelines for Electronic Distribution of MPEG and WG 11 Documents
8436 Convener
Press Release of the 78th Meeting in Hangzhou, CN
8437 Convener
Meeting Notice of the 79th Meeting in Marrakech, MA
8438 HoD
Guide for WG 11 Meeting Hosts
8439 HoD
MPEG 101
8440 Convener
AHG on Video IDCT Specification
8441 Convener
AHG on Maintenance of MPEG-4 Visual related Documents, Reference
Software and Conformance
8442 Convener
AHG on Reconfigurable Video Coding
8443 Convener
AHG on Maintenance of MPEG-7 Visual related Documents, Reference
Software and Conformance
8444 Convener
AHG on MPEG-7 Visual and Photo Player MAF
8445 Video
Disposition of Comments on ISO/IEC 13818-2:2000/FPDAM 2
8446 Video
Text of ISO/IEC 13818-2:2000/FDAM 2 Support for Colour Spaces
8447 Video
Disposition of Comments on ISO/IEC 14496-2:2004/FPDAM3
8448 Video
Text of ISO/IEC 14496-2:2004/FDAM 3 Support for Colour Spaces
8449 Video
Defect Report on ISO/IEC 14496-10:2005 (Version 2)
8450 Video
Disposition of Comments on ISO/IEC 14496-10:2005/FPDAM1
8451 Video
Text of ISO/IEC 14496-10:2006/FDAM 1 Support for Colour Spaces
and Aspect Ratios
8452 Video
Study Text of ISO/IEC 14496-10:2005/FPDAM2 Advanced 4:4:4
Profiles
8453 Video
Joint 4:4:4 Video Model (JFVM) 5
8454 Video
JFVM 5 Software
8455 Video
Study Text of ISO/IEC 14496-10:2005/FPDAM3 Scalable Video Coding
8456 Video
Joint Scalable Video Model (JSVM) 8
8457 Video
JSVM 8 Software
8458 Video
Working Draft 1 of ISO/IEC 14496-10:2005/Amd.4 Multiview Video
Coding
38
8459 Video
Joint Multiview Video Model (JMVM) 2
8460 Video
JMVM 2 Software
8461 Video
Disposition of Comments on ISO/IEC 15938-3:2002/Amd.1/DCOR2
8462 Video
Text of ISO/IEC 15938-3:2002/Amd.1/COR2
8463 Video
Defect Report on ISO/IEC 15938-3:2002/Amd.2
8464 Video
Description of Core Experiments for MPEG-7 New Visual Extensions
8465 Video
Disposition of Comments on ISO/IEC 15938-6:2003/PDAM2
8466 Video
Text of ISO/IEC 15938-6:2003/FPDAM2 (Perceptual 3D Shape)
8467 Video
Disposition of Comments on ISO/IEC 15938-7:2003/PDAM3
8468 Video
Text of ISO/IEC 15938-7:2003/FPDAM3 (Perceptual 3D Shape)
8469 Video
Disposition of Comments on ISO/IEC TR 15938-8:2002/DAM3
8470 Video
Text of ISO/IEC TR 15938-8:2002/FDAM3 (Technologies for digital
photo management using MPEG-7 visual tools)
8471 Video
Disposition of comments on ISO/IEC FCD 23000-3
8472 Video
Text of ISO/IEC FDIS 23000-3
8473 Video
Request for ISO/IEC 23000-3/Amd.1: Reference Software for Photo
Player MAF
8474 Video
Working Draft 1 of ISO/IEC 23000-3/Amd.1
8475 Video
Request for Subdivision: ISO/IEC 23001-4 Codec Description
Representation
8476 Video
WD 2 of ISO/IEC 23001-4
8477 Video
Request for ISO/IEC 23002-1/Amd.1 Software for Integer IDCT
Accuracy Testing
8478 Video
Text of ISO/IEC 23002-1/PDAM1
8479 Video
ISO/IEC CD 23002-2 Fixed point IDCT and DCT
8480 Video
Description of Core Experiments on Fixed-Point DCT/IDCT
8481 Video
Software Testbed for fixed-point DCT/IDCT V 5.0
8482 Video
Study Text of ISO/IEC FCD 23002-3 Representation of Auxiliary Video
and Supplemental Information
8483 Video
Request for Subdivision: ISO/IEC 23002-4 Video Tool Library
8484 Video
WD 2 of ISO/IEC 23002-4
8485 Video
White Paper on Reconfigurable Video Coding (RVC)
8486 Video
Description of Core Experiments in RVC
8487 Video
RVC Simulation Model (RSM) V2.0
8488 Video
RVC Work Plan
8489 3DGC
DoC on ISO/IEC 14496-4:2004/ FPDAM12 (Morphing & Textures)
8490 3DGC
Text of ISO/IEC 14496-4:2004/ FDAM12 (Morphing & Textures)
39
8491 3DGC
DoC on ISO/IEC 14496-4:2004/ PDAM16 (MPEG-J GFX)
8492 3DGC
Text of ISO/IEC 14496-4:2004/ FPDAM16 (MPEG-J GFX)
8493 3DGC
Request for ISO/IEC 14496-4:2004/ AMD21 (Geometry & Shadow)
8494 3DGC
Text of ISO/IEC 14496-4:2004/ PDAM21 (Geometry & Shadow)
8495 3DGC
DoC on ISO/IEC 14496-5:2001/ FPDAM9 (Morphing & Textures)
8496 3DGC
Text of ISO/IEC 14496-5:2001/ FDAM9 (Morphing & Textures)
8497 3DGC
Request for ISO/IEC 14496-5:2001/AMD13 (Geometry & Shadow)
8498 3DGC
Text of ISO/IEC 14496-5:2001/ PDAM13 (Geometry & Shadow)
8499 3DGC
3D Graphics Core Experiments Description
8500 Convener
Terms of Reference
8501 Convener
MPEG Standards
8502 Convener
Table of unpublished standards at FDIS level
8503 Convener
Work plan and time line
8504 Convener
Work item assignment
8505 Convener
MPEG Standard Editors
8506 3DGC
3D Graphics Compression FAQ 16.0
8507 Convener
AHG on 3DGC documents, experiments and software maintenance
8508 Requirements MPEG-7 Requirements
8509 Requirements MPEG-7 Query Formats Requirements
8510 Requirements Final Call on MPEG-7 Query Formats
8511 Requirements MAFs Overview
8512 Requirements MAFs Awareness Event
8513 Requirements
Dual-track Straw Man for IP Free and IP bearing but Royalty Free
Standards Making
8514 Convener
AHG on the development of MPEG standards
8515 Convener
AHG on IPTV Requirements
8516 Convener
AHG on MPEG-7 Query Formats
8517 Convener
AHG on MAFs Awareness Event
8518 ISG
Status of HDL submissions and commitments for MPEG-4 Part-9
8519 ISG
Study of “ISO/IEC PDTR 14496-9 3rd Edition Reference Hardware
Description”
8520 Convener
AHG on MPEG-4 Part 9: Reference Hardware Description Phase 1 and
2.
8521 Convener
Software assets
8522 Convener
Conformance assets
8523 Convener
Content assets
40
8524 Convener
URI assets
8525 Convener
Standards under development for which a call for patent statements is
issued
8526 Liaison
Liaison Statement to UHAPI concerning M3W
8527 Liaison
Liaison Statement to ITU-T FG/IPTV WG 6 concerning M3W
8528 Liaison
Liaison Statement to 3GPP2
8529 Liaison
Liaison Statement to ITU-R SG6 WP 6J concerning colour space
amendments
8530 Liaison
Liaison Statement to ITU-R SG6 WP 6Q on Call for Proposals
8531 Liaison
Liaison Statement to SMPTE
8532 Liaison
Liaison Statement to SMPTE on 4:2:2 and 4:2:0 Intra-only profiles of
AVC
8533 Liaison
Liaison Statement to SMPTE on 4:4:4 Intra-only profile of AVC
8534 Liaison
Liaison Statement to FLO Forum
8535 Liaison
Liaison Statement to IEC TC100
8536 Liaison
Liaison Statement to W3C MMSem-XG
8537 Liaison
Liaison Statement to ITU-T SG9 concerning FTV ad MVC
8538 Liaison
Liaison Statement to OMA BAC MAE
8539 Liaison
Liaison Statement to DVB
8540 Liaison
Liaison Statement to TC184 SC4
8541 Liaison
Liaison Statement to ITU-T SG16 Q10 comments on G722.2EV
8542 Liaison
Liaison Statement to SCTE
8543 Liaison
Liaison Statement to WG1 (JPEG)
8544 Liaison
Liaison Statement to Khronos
8545 Liaison
Liaison Statement to ITU-T FG/IPTV WG 6 concerning work on IPTV
8546 Liaison
Liaison Statement to ITU-T SG16 Q23
8547 DELETED
DELETED
8548 Liaison
Request for establishment of Category A liaison with 3GPP2
8549 Liaison
Request for establishment of Category B liaison with AES
8550 Liaison
Request for establishment of Category C liaison with Khronos
8551 Liaison
Response to National Bodies
8552 Liaison
List of Organisations with which MPEG entertains liaisons (as of
October 2006)
8553 Testing
Draft SVC Verification Test Plan
8554 Testing
Request for Video Test Sequences
8555 MDS
Metadata Conversion – Problem and High Level Solution Statement
8556 MDS
Request for Amendment 3 of ISO/IEC 15938-5 Improvements to
41
Geographic Position Descriptor
8557 MDS
ISO/IEC 15938-5/PDAM 4 Improvements to Geographic Position
Descriptor
8558 MDS
Request for Amendment 4 of ISO/IEC 15938-7 New Geographic
Position Descriptor Conformance
8559 MDS
ISO/IEC 15938-7/PDAM 4 New Geographic Position Descriptor
Conformance
8560 MDS
DoC on ISO/IEC 15938-10:2005/DCOR 1 Multimedia content
description interface — Part 10: Schema definition
8561 MDS
ISO/IEC 15938-10:2005/COR 1 Multimedia content description
interface — Part 10: Schema definition
8562 MDS
Schema Files for MPEG-21 standards v.5
8563 MDS
DoC for ISO/IEC 21000-4/PDAM 1: MPEG-21 IPMP Components Base
Profile
8564 MDS
ISO/IEC 21000-4/FPDAM 1: IPMP Components Base Profile
8565 MDS
Request for Amendment 3 of ISO/IEC 21000-5 ORC (Open Release
Content) Profile
8566 MDS
ISO/IEC 21000-5/PDAM 3 ORC (Open Release Content) Profile
8567 MDS
DoC on ISO/IEC 21000-6/DCOR 2 Rights Data Dictionary
8568 MDS
Text of ISO/IEC 21000-6/COR 2 Rights Data Dictionary
8569 MDS
Disposition of Comments on ISO/IEC 21000-7/FPDAM 2
8570 MDS
Text of ISO/IEC 21000-7/ FDAM 2 Dynamic and Distributed Adaptation
8571 MDS
MPEG-21 DIA Reference Software and Status Work Plan
8572 MDS
Study of ISO/IEC CD 21000-8: Reference Software Second Edition
8573 MDS
DoC on ISO/IEC CD 21000-14: Conformance Testing
8574 MDS
ISO/IEC FCD 21000-14: Conformance Testing
8575 MDS
DoC of ISO/IEC FCD 21000-18 Digital Item Streaming
8576 MDS
Text of ISO/IEC 21000-18 Digital Item Streaming
8577 MDS
TuC v5.0 for ISO/IEC 21000-18 Digital Item Streaming
8578 MDS
Workplan for Core Experiment on DI Streaming Technologies under
Consideration
8579 MDS
Request for Amendment 1 of ISO/IEC 21000-18 Digital Item Streaming:
Simple Fragmentation Rule
8580 MDS
ISO/IEC 21000-18/PDAM/1 Digital Item Streaming
8581 MDS
DoC on ISO/IEC CD 23000-2 MPEG-A Music Player 2nd edition
8582 MDS
ISO/IEC FCD 23000-2 MPEG-A Music Player 2nd edition
8583 MDS
Reference Software Workplan for MPEG-A Music Player 2nd edition
42
8584 MDS
ISO/IEC CD 23000-5 Media Streaming Player
8585 MDS
TuC for Media Streaming Player IPMP Technologies
8586 MDS
Reference Software Workplan for ISO/IEC CD 23000-5 Media
Streaming Player
8587 MDS
Request for Name Change of subdivision 23000-6 to Professional
Archival MAF
8588 MDS
Professional Archival MAF Under Development Workplan
8589 Convener
AHG on SVC Verification Test
8590 Empty
Empty
8591 Empty
Empty
8592 Empty
Empty
8593 Empty
Empty
8594 Empty
Empty
8595 Empty
Empty
8596 Empty
Empty
8597 Empty
Empty
8598 Empty
Empty
8599 MDS
WD of 23000-6 Professional Archival MAF - Audio
8600 MDS
MPEG-21 Session Mobility One Pager
8601 MDS
MPEG-21 Digital Item Processing Amendment 1 One Pager
8602 MDS
MPEG-21 Conformance to Digital Item Processing One Pager
8603 MDS
MPEG-21 Conformance to Digital Item Processing Amendment 1 One
Pager
8604 MDS
MPEG-21 Reference Software One Pager
8605 Convener
AHG on MPEG-21 DIS
8606 Convener
AHG on MDS MAFs Under Development
8607 Audio
ISO/IEC 11172-5:199x/DCOR 1
8608 Systems
Text of ISO/IEC 23004-5/FCD Component Download
8609 Audio
ISO/IEC 13818-4:2004/AMD 2:2005/Cor. 1
8610 Audio
DoC on ISO/IEC 13818-7:2006/PDAM 1
8611 Audio
ISO/IEC 13818-7:2006/FPDAM 1, Transport of MPEG Surround data in
AAC
8612 Audio
Study on ISO/IEC 14496-3:2005/PDAM 5, BSAC Extensions
8613 Audio
DoC on ISO/IEC 14496-3:2006/PDAM 6, Symbolic Music
Representation
8614 Audio
WD on Support for 64-bit address space in ancillary data
8615 Audio
Request for Amendment, AAC-ELD
43
8616 Audio
ISO/IEC 14496-3:2005/PDAM 9, AAC-ELD
8617 Audio
ISO/IEC 14496-4:2004/AMD11/Cor. 2 Parametric Stereo Conformance
8618 Audio
DoC on ISO/IEC 14496-4:2004/PDAM 14, BSAC Extension
Conformance
8619 Audio
ISO/IEC 14496-4:2004/FPDAM 14, BSAC Extension Conformance
8620 Audio
DoC on ISO/IEC 14496-4:2004/PDAM 18, MPEG-1 and -2 on MPEG-4
Conformance
8621 Audio
ISO/IEC 14496-4:2004/FPDAM 18, MPEG-1 and -2 on MPEG-4
Conformance
8622 Audio
DoC on ISO/IEC 14496-4:2004/PDAM 19, ALS Conformance
8623 Audio
ISO/IEC 14496-4:2004/FPDAM 19, ALS Conformance
8624 Audio
DoC on ISO/IEC 14496-4:2004/PDAM 20, SLS Conformance
8625 Audio
ISO/IEC 14496-4:2004/FPDAM 20, SLS Conformance
8626 Audio
Status of BSAC Extension conformance
8627 Audio
Status of ALS Conformance
8628 Audio
Status of SLS Conformance
8629 Audio
Status of MPEG-4 Audio Conformance
8630 Audio
Workplan for updates on SLS reference software
8631 Audio
Request for Subdivision, Symbolic Music Representation
8632 Audio
ISO/IEC 14496-23:200x/FCD, Symbolic Music Representation
8633 Audio
Request for Amendment, MPEG Surround conformance testing
8634 Audio
ISO/IEC 23003-1:2006/PDAM 1, MPEG Surround conformance testing
8635 Audio
Request for Amendment MPEG Surround reference software
8636 Audio
ISO/IEC 23003-1:2006/PDAM 2, MPEG Surround reference software
8637 Audio
Workplan for MPEG Surround verification test
8638 Audio
SAOC use cases, draft requirements and architecture
8639 Audio
Draft Call for Proposals on Spatial Audio Object Coding
8640 Audio
Workplan for Exploration of Speech and Audio Coding
8641 Audio
Audio Bifs version 3
8642 Audio
Audio Conformance and Reference Software Assets
8643 Convener
AHG on Audio Standards Maintenance
8644 Convener
AHG on Exploration of Speech and Audio Coding
8645 Convener
AHG on MPEG Surround Verification Test and SAOC CfP
8646 Systems
Text of ISO/IEC 14496-1:2004/COR2 OD Dependencies
8647 Systems
Request of ISO/IEC 14496-4/Amd.24 File Format Conformance
8648 Systems
Text of ISO/IEC 14496-4/PDAM.24 File Format Conformance
44
8649 Systems
Request of ISO/IEC 14496-4/Amd.25 LASeR Conformance
8650 Systems
Text of ISO/IEC 14496-4/PDAM.25 LASeR Conformance
8651 Systems
WD1.0 of ISO/IEC 14496-20/Amd.27 LASeR Conformance
8652 Systems
DoC of ISO/IEC 14496-5/PDAM12 File Format Reference Software
8653 Systems
Text of ISO/IEC 14496-5/FPDAM12 File Format Reference Software
8654 Systems
WD1.0 of ISO/IEC 14496-4/Amd.15 LASeR Reference Software
8655 Systems
WD1.0 of ISO/IEC 14496-4/Amd.16 Symbolic Music Representation
Ref. Soft.
8656 Systems
DoC on ISO/IEC 14496-11:2005/FPDAM5 Symbolic Music
Representation
8657 Systems
Text of ISO/IEC 14496-11:2005/FDAM5 Symbolic Music
Representation
8658 Systems
DoC on ISO/IEC 14496-12/FPDAM1 (Description of Timed Metadata)
8659 Systems
Text of ISO/IEC 14496-12/FDAM1 (Description of Timed Metadata)
8660 Systems
DoC on ISO/IEC 14496-12/PDAM2 (Flute Hint Track)
8661 Systems
Text of ISO/IEC 14496-12/FPDAM2 (Flute Hint Track)
8662 Systems
Draft DoC on ISO/IEC 14496-15/PDAM2 (SVC File Format)
8663 Systems
Study Text of ISO/IEC 14496-15/PDAM2 (SVC File Format)
8664 Systems
Text of ISO/IEC 14496-18/COR1
8665 Systems
DoC on ISO/IEC 14496-20/DCOR1
8666 Systems
Text of ISO/IEC 14496-20/DCOR1
8667 Systems
Study Text of ISO/IEC 14496-20/FPDAM1 (SVGT1.2 Support)
8668 Systems
TuC for ISO/IEC 14496-20/Amd1
8669 Systems
WD1.0 of ISO/IEC 14496-20 2nd Edition (1st Ed. + Cor + Amd.)
8670 Systems
First ideas on MPEG-21 and LASeR
8671 Systems
DoC on ISO/IEC 15938-7/FPDAM2 Fast Access Extension
Conformance
8672 Systems
Text of ISO/IEC 15938-7/FDAM2 Fast Access Extension Conformance
8673 Systems
DoC on ISO/IEC 23000-4/CD (Musical Slide Show MAF)
8674 Systems
Text of ISO/IEC 23000-4/FCD (Musical Slide Show MAF)
8675 Systems
TuC for ISO/IEC 23000-4 (Musical Slide Show MAF)
8676 Systems
Request of Subdivision of ISO/IEC 23000
8677 Systems
WD1.0 of ISO/IEC 23000-8 (Portable Video Player MAF)
8678 Systems
Request of Subdivision of ISO/IEC 23000
8679 Systems
WD1.0 of ISO/IEC 23000-9 (Digital Multimedia Broadcasting MAF)
8680 Systems
Text on ISO/IEC 23001-1/COR1 (Editorial and technical clarifications)
8681 Systems
Study of Text of ISO/IEC 23001-1/PDAM2 (Prefixes and of wild cards
45
extensions)
8682 Systems
MPEG-B Part 1 Reference software workplan
8683 Systems
DoC on ISO/IEC 23001-2/PDAM2 (Fragment Request Unit)
8684 Systems
Text of ISO/IEC 23001-2/FPDAM2 (Fragment Request Unit)
8685 Systems
Request of subdivision of ISO/IEC 23001
8686 Systems
Text of ISO/IEC 23001-3/CD (Binary to XML Mapping of IPMP-X)
8687 Systems
M3W White Paper : Multimedia Middleware Architecture
8688 Systems
M3W White Paper : Multimedia API
8689 Systems
M3W White Paper : Component Model
8690 Systems
M3W White Paper : Resource and Quality Management
8691 Systems
M3W White Paper : Component Download
8692 Systems
M3W White Paper : Fault Management
8693 Systems
M3W White Paper : System Integrity Management
8694 Systems
M3W Reference Software Plan
8695 Systems
Request for New Project on Supplemental Media Technology
8696 Systems
Text of ISO/IEC XXXXX-1/CD Media Streaming MAF Protocol
8697 Convener
Ad Hoc Group on Scene Representation
8698 Convener
Ad Hoc Group on MPEG File Formats
8699 Convener
Ad Hoc Group on MAF Under Development in Systems
8700 Systems
Text of ISO/IEC 23004-6/FCD Fault Management
8701 Systems
Text of ISO/IEC 23004-7/FCD System Integrity Management
8702 Liaison
Liaison Statement to 3GPP
46
Annex E – Requirements report
Source: Fernando Pereira (Instituto Superior Técnico, Lisboa-Portugal
Note: Requirements agenda for the Hangzhou MPEG meeting is annexed at the end of this
report.
1. Requirements documents approved at this meeting
N8508 MPEG-7 Requirements
N8509 MPEG-7 Query Format Requirements
N8510 Final Call for Proposals on MPEG-7 Query Format
N8511 MAFs Overview
N8512 First MAF Awareness Event
N8513 Dual Track Straw Man for IP Free and IP bearing but Royalty Free Standards
Making
N8514 AHG on the development of MPEG standards
N8515 AHG on IPTV Requirements
N8516 AHG on MPEG-7 Query Formats
N8517 AHG on MAFs Awareness Event
2. MPEG-4
a. AVC Profiling (joint with JVT)
13964, Thomas Wedi, Hideki Ohtaka, John Wus, Shun-ichi Sekiguchi, Intra-only H.264/AVC
profiles for professional applications
This contribution made requests for AVC profiles which were largely fulfilled with the
definition of 4 new AVC ‘Professional’ Profiles: 4:4:4 Intra, 14b; 4:2:2 Intra, 10b; 4:2:0 Intra,
10b and 4:4:4 Predictive, 14b (see more details in the JVT report).
b. SVC Profiling (joint with JVT)
During this joint meeting, the draft definition of 4 SVC Profiles (so called A, B, B Intra and
C) was discussed. These profiles’ definitions will be further refined as well as the levels
definitions. The most convincing application scenarios behind these profiles will be used for
SVC verification testing (see more details in the JVT report).
c. Laser (joint with Systems)
14012, Sylvain Devillers, Renaud Cazoulat, Use case and requirement for LASeR
This contribution was discussed and the conclusion was that the problem addressed has
already a solution in the context of MPEG standards and no more requirements are needed.
3. MPEG-7
47
a. MPEG-7 Requirements (joint with MDS)
13807, Tanya Beech, Proposal for improvements to Geographic Position in Mpeg7 Part 5
This contribution regards the refinement of the MPEG-7 requirement on Geographic Position
(MDS), notably regarding the ability for users to define a geographical position with one or
more Points, and the ability for users to define the type of GeographicPosition Point details.
Examples of possible GeographicPosition Point details might include Area, Route or Point.
This requirement was accepted and forward to the MDS subgroup for technical development.
To include this requirement, a new version of the MPEG-7 Requirements document has been
issued (N8508).
b. MPEG-7 Query Formats (M7QF) (joint with MDS)
The MPEG-7 Query Format effort will standardize the format of the request sent to the server
and the format of the response sent from the server with additional tools for query
management capability (see figure below). The MPEG-7 Query Format standard will not
specify the behavior of the server because the specific behavior of the server will differ from
implementation to implementation.
Client
Application
Input Query Format
MPEG-7
Database
Output Query Format
Query Management Tools
This meeting confirmed the schedule adopted at last meeting, notably:
 Preliminary Call for Proposals – April 2006
 Final Call for Proposals – July 2006
 Evaluation – January 2007
 CD – July 2007
 FCD – October 2007
 FDIS – January 2008
3870, Soo-Jun Park, Sung Min Kim, Chee Sun Won, Proposal for a new MPEG-7 input query
format: Query-by-Layout
13875, Kyoungro Yoon, Hee-Cheol Seo, Hyunki Kim, Myung-Gil Jang,
Comparison of MP7QF Requirements and TV-Anytime Technology
13876, Hee-Cheol Seo, Hyunki Kim, Myung-Gil Jang, Kyoungro Yoon, Comparison of
XQuery and MPEG-7 Query Format
13891, Sang-Kyun Kim, Ryong Lee, Request for adding Query Requirements related to data
manipulation against MPEG-7 DB on current MPEG-7 Query Format Requirement
13838, Munsi Haque and Addicam Sanjay, Use of UPnP Content Directory Service (CDS) in
MP7QF
Following these contributions, the M7QF Requirements has been improved with 2 additional
requirements (fitting well in the current architecture and vision) and some more examples. At
the end of this meeting, a revised version of the M7QF Requirements has been issued
(N8509) as well as the Final Call for Proposals (N8510). An AHG with the mandates to i)
48
distribute the “Final Call for Proposals on MPEG-7 Query Format (MP7QF)”, ii) organize
logistics of the evaluation, and iii) perform a preliminary evaluation of the proposals at the
AHG meeting, immediately preceding the 79th meeting, was created (N8516).
4. MPEG-21
a. MPEG-21 REL Profiling
13892, Kisong Yoon(ETRI), Taehyun Kim(DRM inside), Eva Rodriguez(DMAG-UPC), Jaime
Delgado(DMAG-UPC), Hogab Kang(DRM inside), Proposed MPEG-21 REL Open Release
Profile
This contribution proposed a REL Open Release Content profile in close connection with the
Open Release MAF. Among other features, this profile should be able to express the
intentions of CC licenses. The technical work will be carried out by MDS.
5. MPEG-A
a. MAFs Awareness Event
13967, Wo Chang, MAF to Industry
Recognizing the importance to advertise the MAFs achievements to the industry, it was
decided to organize a MAF Awareness Event - Connecting Multimedia Applications and
Services – at the Doubletree Hotel, on the April 28 (Saturday after MPEG meeting), 2007, in
San Jose, USA. This awareness event will include both technology presentations and demos
(N8512). An AHG was created to i) to plan meeting logistics (location, registration fee, etc.);
ii) to compile MAF topic and description from identified speakers; and iii) to establish MAF
website to industry (N8517).
b. Media Streaming MAF
c. A joint Requirements-MDS meeting was held to clarify the requirements
regarding this MAF. The proposed requirements have been approved and
included in the revised MAFs Overview document (N8511). It was also
agreed that the requirements must be as much as possible fulfilled using
a combination of MPEG technologies, independently of the standard or
parts used. It was also agreed that MAFs may complement existing
MPEG technologies in the same way that industry consortia used to do.
d. Open Release MAF
13915, Chun Hui Suen, Florian Schreiner, Klaus Diepold, File Format and Event Reporting
for Open Release MAF
13928, Xin Wang, Chris Barlas, Rights Enforceability in the Open Release MAF
e. Following these contributions, the requirements for this MAF have been
revised in the MAFs Overview document. Since the requirements were
considered mature enough, this MAF was promoted to ‘under
development’ and given in charge to the MDS subgroup.
f. Surveillance MAF
13759, James A.G. Annesley, James Orwell, Jim Aldridge, Kate Grant, AHG on Surveillance
MAF
49
13906, Eva Rodríguez, Jaime Delgado, IPMP and the Surveillance MAF (not presented
because authors were not available)
g. At this meeting, some good application scenarios have been provided.
However, technical requirements are still too vague, making impossible
to identify the possible tools to be included in this MAF. Further
progress on this MAF requires the more precise specification of
requirements. Since the surveillance application space is rather big, it
may happen than more than one MAF may have to be developed in this
area. Following this contribution and the BoG work during the week, the
application scenarios and requirements for this MAF have been revised
in the MAFs Overview document.
13760, H. Jean Cha, Herbert Thoma, AHG on Portable Video Player MAF
13995, H. Jean Cha, Tae Hyeon Kim, Refined requirements and technologies for Portable
Video Player MAF
13998, H. Jean Cha, Tae Hyeon Kim, Proposed working draft of Portable Video Player MAF
14017, H. Jean Cha, Proposed Work Plan for Portable Video Player MAF
h. Following these contributions, the application scenarios and
requirements for this MAF have been revised in the MAFs Overview
document. Since the requirements were considered mature enough, this
MAF was promoted to ‘under development’ and given in charge to the
Systems subgroup.
i. Digital Video/Cinema MAF
13862, Mayumi Koike, Takuyo Kogure, Hiroshi Yasuda, Adaptation to MPEG MAF of Digital
Video/Cinema file format
13863, Mayumi Koike, Takuyo Kogure, Hiroshi Yasuda, Requirement of Color Management
Information to MPEG-7 for Digital Video/Cinema
j. Following the requirements for this MAF and the close relation with the
Media Streaming MAF, it was decided to start addressing the
requirements for this MAF by adding the AVC video codec as one of the
‘service types’ in the Media Streaming MAF. Regarding possible
missing color management metadata, it was concluded that, based on the
vague requirements provided, it is not possible to identify if something is
missing from the MPEG-7 set of tools in this area. If something is really
missing, more detailed requirements will have to be brought to MPEG.
This MAF will still remain under consideration, especially because it is
also asking for scalable video coding and protection scalabilities and the
SVC standard is still under development. Further work on this MAF is
expected later in the future, notably including a more precise set of
requirements.
k. Digital Multimedia Broadcasting MAF
13858, Korean National Body, A T-DMB White Paper and a Introductory Movie (6 minutes)
13857, Korean National Body, Request on MAF standardization for DMB
50
13859, Munchurl Kim, Jeongyeon Lim, Hui Yong Kim, Hyon-Gon Choo, Yong Han Kim,
Jinhan Kim, Sung Ho Jin, Requirements for DMB MAF
13860, Munchurl Kim, Jeongyeon Lim, Hui Yong Kim, Hyon-Gon Choo, Yong Han Kim,
Jinhan Kim, Sung Ho Jin, Proposal for DMB Multimedia Application Format
l. Following these contributions, the application scenarios and
requirements for this MAF have been included in the MAFs Overview
document. Since the requirements and possible solutions were
considered mature enough, this MAF was promoted to ‘under
development’ and given in charge to the Systems subgroup. Since there
no consensus on the name of this MAF, the current name is considered
preliminary.
m. IPTV MAF
13929, Xin Wang, Proposal for Working on an IPTV MAF
n. This contribution proposed to develop an IPTV MAF based on the IPTV
requirements available, notably those brought to MPEG by the ATIS
Interoperable IPTV Forum. Since it was considered that some of the
existing MAFs, notably the Media Streaming and DMB MFAs, already
address the IPTV application scenario, it was decided to create an AHG
with the following tasks: i) identify relevant requirements from the list
provided; ii) collect more IPTV requirements; iii) check coverage of
relevant requirements by existing MAFs, notably MS and DMB MAFs,
and iv) provide plan to address gaps related to relevant requirements (if
we want to support them) (N8514).
o. Medical Imaging MAF
13966, Wo Chang, Proposed Medical Imaging MAF (MI MAF) for Preserving Medical
Imaging Records
The purpose of this contribution was to explore if MPEG experts have an interest on applying
the MAF concept to the medical imaging application domain. The contribution mentions the
relation with the Digital Imaging and Communication in Medicine (DICOM) standard which
has become one of the well-established and commonly used standards in medicine since 1983.
It was concluded that medical imaging is a very important area where MPEG did not have
much impact until now. MPEG experts interested in this application area should get together
to understand better what may be the requirements for a possible MAF and eventually make a
proposal.
p. DTV
13889, Hui Yong Kim, Jeong Hyun Yoon, Hee Kyung Lee, Han Kyu Lee, Sung Ho Jin, JaeSeok Jang, Yong Man Ro, Requirements for DTV MAF
This contribution was withdrawn.
q.
Summary on MAFs
The global MAF situation after the Hangzhou MPEG meeting is summarized in the MAFs
Overview document (N8510) as follows:
51
1. MAFs Finalized
a. Music Player MAF (done)
2. MAFs Under Development
a. Photo Player MAF (under Video) - FDIS
b. Protected Music Player MAFs (under MDS) - FCD
c. Musical Slide Show MAF (under Systems) - FCD
d. Media Streaming MAF (under MDS) - CD
e. Professional Archival MAF (under MDS) – WD
f. Open Release MAF (under MDS) – CD
g. Portable Video Player MAF (under Systems) – WD
h. Digital Multimedia Broadcasting MAF (under Systems) – WD
3. MAFs Under Consideration
a. Surveillance MAF
b. Digital Video/Cinema MAF
6. MPEG-C
Fixed-point Approximation of 88 IDCT Transform (joint with Video/ISG)
13930, CNNB, CNNB comments on the work of fixed-point 8x8 IDCT transform
This document was discussed in a joint session Requirements/Video/ISG. It was reaffirmed
that:
 It is essential that this activity provides the industry with something useful, with clear
benefits, which the industry will want to use.
 A single IDCT must be selected to get a drift free solution, if encoder and decoder use the
same IDCT.
 As stated in the CfP, the single IDCT solution must conform to previous standards, e.g.
MPEG-1, MPEG-2.
 The selection of the single IDCT solution will be based on an accuracy-complexity tradeoff.
Regarding the procedural question, WG11 experts raised no objection to the integration of all
available proposals in the technical selection process, including those proposals that may not
have fully followed the predefined rules.
Finally, 2 proposals from the 5 in the current WD were selected for further technical based on
performance results obtained since last meeting. The selection of the single solution to include
in the CD was given to the Video and ISG subgroups (see more details in the Video and ISG
reports).
7. Explorations
a. Free Viewpoint TV
13894, Masayuki Tanimoto, Toshiaki Fujii, Shigeyuki Sakazawa, Hideaki Kimata,
Requirements on Free Viewpoint Television (FTV) v.0
This contribution addresses the subject of free-viewpoint TV (FTV) / multi-view coding
(MVC). The document is not clear on the interfaces for which a standard approach is
requested and why. The discussion indicated that MPEG seems to have all the necessary tools
at the coding (MVC) and metadata (MPEG-7 or camera parameters in MVC) levels.
52
Regarding other requirements such as system layer requirements, the document is not clear
and thus further contributions are welcome with more precise requests.
b. Dual-Track Licensing Approach (N8071 e N8225)
13758, Chris Barlas, Takuyo Kogure, Andy Tescher, AHG on the Development of MPEG
standards
Two meetings ago, MPEG started an exploration activity to discuss the possibility to adopt a
Dual-Track Licensing Approach for developing MPEG standards. The major objective of the
new approach would be to extend the usage of MPEG specifications since it is perceived by
some delegates that many users are not adopting MPEG solutions because they find the
licensing conditions for technology in existing MPEG specifications too onerous for their
particular application, market etc. The dual-track approach for developing MPEG
specifications would than include:
i)
Track (a) is the long standing MPEG mode of operation where standards, based on
the best technology, are developed with technology made available under RAND
licensing conditions.
ii)
Track (b) will see the development of new specifications for video encoders and
decoders using a combination of
i. Existing technology which is no longer subject to usage restrictions due to
patents and other intellectual property rights etc, and
ii. Technology contributed to MPEG by companies willing to license that
technology on royalty-fee free terms for the implementation of a particular
specification
At the last meeting, it was decided to follow a three steps work plan described in N8225:
 Technology Survey – Identify potential RF standards (research external RF technologies
that might compete) and validate need.
 Process Design - Concurrently identify resources to design the process for development
and exploitation of RF standards.
 Standards development - Identify resources for making RF standards including other
bodies currently working on RF models and willing to collaborate.
Although there was some input since last meeting on the technology survey, there was no
input on the two other items above. In order to stimulate further contributions, it was decided
at this meeting to develop a straw man for the process, considering two main cases: IP Free
and IP bearing but Royalty Free Standards. It is intended to address in the future the case of
IP bearing but Royalty Free standard providing the Royalty Free core of a set of interdependent standards, some or all of which are royalty bearing.
The document with the straw man (N8513) developed at this meeting is intended as a starting
point for discussion. Nothing in the document may remain after debate and examination.
Readers are encouraged to criticize any part of it and propose their own improvements or
supply arguments as to why some or all of the statements herein are incorrect. The
Requirements subgroup kindly requests that companies and other interested parties study
N8513, which contains a straw man argument for IP Free and Royalty Free standardization
processes. Companies and other interested parties are invited to provide input in response to
this document at the 79th meeting.
53
8. 78th MPEG (Hangzhou) Agenda Requirements
9. Room: Ruiqi
TIME
TOPIC
ROOM
Monday
Opening Plenary Meeting
9:00-end
Lunch
First Discussion on New MAF Proposals
DMB MAF
13858, Korean National Body, A T-DMB White Paper and a Introductory Movie (6
minutes)
13857, Korean National Body, Request on MAF standardization for DMB
13859, Munchurl Kim, Jeongyeon Lim, Hui Yong Kim, Hyon-Gon Choo, Yong Han Kim,
Jinhan Kim, Sung Ho Jin, Requirements for DMB MAF
13860, Munchurl Kim, Jeongyeon Lim, Hui Yong Kim, Hyon-Gon Choo, Yong Han Kim,
14:30-17:00
Jinhan Kim, Sung Ho Jin, Proposal for DMB Multimedia Application Format
Reqs
DTV
13889, Hui Yong Kim, Jeong Hyun Yoon, Hee Kyung Lee, Han Kyu Lee, Sung Ho Jin,
Jae-Seok Jang, Yong Man Ro, Requirements for DTV MAF
IPTV MAF
13929, Xin Wang, Proposal for Working on an IPTV MAF
Dual-Track Standardization Approach
13758, Chris Barlas, Takuyo Kogure, Andy Tescher, AHG on the Development of
MPEG standards
HoDs Meeting
18:00-20:00
17:00-18:00
54
Reqs
HoD
Tuesday
Open Release MAF (Joint with MDS)
13915, Chun Hui Suen, Florian Schreiner, Klaus Diepold, File Format and Event
Reporting for Open Release MAF
13928, Xin Wang, Chris Barlas, Rights Enforceability in the Open Release MAF
MPEG-21 Profiles (joint with MDS)
13892, Kisong Yoon(ETRI), Taehyun Kim(DRM inside), Eva Rodriguez(DMAG-UPC),
Jaime Delgado(DMAG-UPC), Hogab Kang(DRM inside), Proposed MPEG-21
REL Open Release Profile
MPEG-7 (Joint with MDS)
9:00-12:00
13807, Tanya Beech, Proposal for improvements to Geographic Position in Mpeg7 Part
5
13870, Soo-Jun Park, Sung Min Kim, Chee Sun Won, Proposal for a new MPEG-7
input query format: Query-by-Layout
13875, Kyoungro Yoon, Hee-Cheol Seo, Hyunki Kim, Myung-Gil Jang,
Comparison of MP7QF Requirements and TV-Anytime Technology
13876, Hee-Cheol Seo, Hyunki Kim, Myung-Gil Jang, Kyoungro Yoon, Comparison of
XQuery and MPEG-7 Query Format
13891, Sang-Kyun Kim, Ryong Lee, Request for adding Query Requirements related to
data manipulation against MPEG-7 DB on current MPEG-7 Query Format
Requirement
13838, Munsi Haque and Addicam Sanjay, Use of UPnP Content Directory Service
(CDS) in MP7QF
Reqs
Surveillance MAF (Joint with MDS)
13759, James A.G. Annesley, James Orwell, Jim Aldridge, Kate Grant, AHG on
Surveillance MAF
13906, Eva Rodríguez, Jaime Delgado, IPMP and the Surveillance MAF
-
12:00-13:00
13:00-14:00
Lunch
MAF (joint with MDS/Video/Systems)
13967, Wo Chang, MAF to Industry
Portable Video Player MAF
13760, H. Jean Cha, Herbert Thoma, AHG on Portable Video Player MAF
13995, H. Jean Cha, Tae Hyeon Kim, Refined requirements and technologies for
Portable Video Player MAF
13998, H. Jean Cha, Tae Hyeon Kim, Proposed working draft of Portable Video Player
MAF
14017, H. Jean Cha, Proposed Work Plan for Portable Video Player MAF
Reqs
14:00-17:30
Digital Cinema MAF
13862, Mayumi Koike, Takuyo Kogure, Hiroshi Yasuda, Adaptation to MPEG MAF of
Digital Video/Cinema file format
13863, Mayumi Koike, Takuyo Kogure, Hiroshi Yasuda, Requirement of Color
Management Information to MPEG-7 for Digital Video/Cinema
DMB MAF
13858, Korean National Body, A T-DMB White Paper and a Introductory Movie (6
minutes)
13857, Korean National Body, Request on MAF standardization for DMB
13859, Munchurl Kim, Jeongyeon Lim, Hui Yong Kim, Hyon-Gon Choo, Yong Han Kim,
Jinhan Kim, Sung Ho Jin, Requirements for DMB MAF
55
13860, Munchurl Kim, Jeongyeon Lim, Hui Yong Kim, Hyon-Gon Choo, Yong Han Kim,
Jinhan Kim, Sung Ho Jin, Proposal for DMB Multimedia Application Format
DTV
13889, Hui Yong Kim, Jeong Hyun Yoon, Hee Kyung Lee, Han Kyu Lee, Sung Ho Jin,
Jae-Seok Jang, Yong Man Ro, Requirements for DTV MAF
IPTV MAF
13929, Xin Wang, Proposal for Working on an IPTV MAF
Medical Imaging MAF
13966, Wo Chang, Proposed Medical Imaging MAF (MI MAF) for Preserving Medical
Imaging Records
17:30-18:00
File formats (joint with MDS/Systems)
18:00-19:00
19:00-end
Liaison Meeting
Chairs Meeting
Reqs
Wednesday
09:00-End
plenary
Plenary Meeting
Laser (joint with Systems)
Reqs
11:30-12:00
14012, Sylvain Devillers, Renaud Cazoulat, Use case and requirement for LASeR
12:00-13:00
Feedback from Dual-Track Standardization Approach BoG
Reqs
Lunch
IDCT (Joint with Video & ISG)
14:00-16:00 13930, CNNB, CNNB comments on the work of fixed-point 8x8 IDCT transform
Other related issues
ISG
Various (joint with JVT)
16:00-18:00
13964, Thomas Wedi, Hideki Ohtaka, John Wus, Shun-ichi Sekiguchi, Intra-only
H.264/AVC profiles for professional applications
More on AVC profiles
SVC profiles
JVT
Social Event
Thursday
9:00-10:00
BoGs meeting
56
-
Various (joint with Video)
10:00-10:30 13986, Justin Ridge, Request new levels for MPEG-4 Simple Profile (withdrawn)
Reqs
13894, Masayuki Tanimoto, Toshiaki Fujii, Shigeyuki Sakazawa, Hideaki Kimata,
Requirements on Free Viewpoint Television (FTV) v.0
10:30-11:30
BoGs meeting
-
11:30-12:30
Possible CfP (joint with 3DGC)
3DGC
Lunch
14:00-14:30
Requirements on SAOC (joint with Audio)
Audio
14:30-15:00
BoGs
-
15:00-16:00
Feedback on Surveillance MAF and IPTV Requirements
Reqs
16:00-17:00
Feedback on SVC profiles (joint with JVT)
JVT
17:00-18:00
Feedback from Dual-Track Standardization Approach BoG
Reqs
18:00-18:30
MS MAF Requirements (joint with MDS)
Reqs
18:00-end
Chairs Meeting
Friday
Concluding MPEG-4
-
Concluding MPEG-7
MPEG-7 Requirements – Tanya
Query Formats Requirements - Wo
9:30-10:00
Query Formats Final Call for Proposals – Wo
AHG on MPEG-7 Query Formats
-
10:0010:30
Concluding MPEG-21
Concluding MPEG-A
MAFs Overview doc - Florian
MAFs Awareness Event – Wo
AHG on MAF Awareness Event
KNB on DMB MAF
AHG on IPTV Requirements – Xin
Liaison - Xin
Reqs
Reqs
Reqs
Reqs
Explorations
10:3011:00
Dual-track straw men doc – Chris
Rec on straw men doc - Chris
AHG on Dual-Track - Chris
CNNB on IDCT
Reqs
57
12:00 14:00
Lunch
14:00-end
plenary
Plenary Meeting
58
Annex F – Systems report
Source:
1
Systems Chair and Break-out group Chairs
Overview
The main outputs of the meeting from the Systems Sub-group perspective are:
No.
Title
X
8646
X
8647
8648
8649
8650
8651
X
8652
8653
8654
8655
X
8656
8657
X
8658
8659
8660
8661
X
8662
8663
X
8664
X
8665
8666
8667
8668
8669
8670
X
8671
8672
X
8673
8674
8675
X
8676
8677
X
14496-1 :2004/MPEG-4 Systems
Text of ISO/IEC 14496-1:2004/COR2 OD Dependencies
14496-4 MPEG-4 Conformance
Request of ISO/IEC 14496-4/Amd.24 File Format Conformance
Text of ISO/IEC 14496-4/PDAM.24 File Format Conformance
Request of ISO/IEC 14496-4/Amd.25 LASeR V1 Conformance
Text of ISO/IEC 14496-4/PDAM.25 LASeR V1 Conformance
WD1.0 of ISO/IEC 14496-4/Amd.27 LASeR V2 Conformance
14496-5 MPEG-4 Reference Software
DoC of ISO/IEC 14496-5/PDAM12 File Format Reference Software
Text of ISO/IEC 14496-5/FPDAM12 File Format Reference Software
WD1.0 of ISO/IEC 14496-5/Amd.17 LASeR Reference Software
WD1.0 of ISO/IEC 14496-5/Amd.16 Symbolic Music Representation Ref. Soft.
14496-11 :2005/MPEG-4 Scene Description
DoC on ISO/IEC 14496-11:2005/FPDAM5 Symbolic Music Representation
Text of ISO/IEC 14496-11:2005/FDAM5 Symbolic Music Representation
14496-12 ISO File Format
DoC on ISO/IEC 14496-12/FPDAM1 (Description of Timed Metadata)
Text of ISO/IEC 14496-12/FDAM1 (Description of Timed Metadata)
DoC on ISO/IEC 14496-12/PDAM2 (Flute Hint Track)
Text of ISO/IEC 14496-12/FPDAM2 (Flute Hint Track)
14496-15 AVC File Format
Draft DoC on ISO/IEC 14496-15/PDAM2 (SVC File Format)
Study Text of ISO/IEC 14496-15/PDAM2 (SVC File Format)
14496-18 Font Compression
Text of ISO/IEC 14496-18/COR1
14496-20 LASeR
DoC on ISO/IEC 14496-20/DCOR1
Text of ISO/IEC 14496-20/COR1
Study Text of ISO/IEC 14496-20/FPDAM1 (SVGT1.2 Support)
TuC for ISO/IEC 14496-20/Amd1
WD1.0 of ISO/IEC 14496-20 2nd Edition (1st Ed. + Cor + Amd.)
First ideas on MPEG-21 and LASeR
15938-7 MPEG-7 Conformance
DoC on ISO/IEC 15938-7/FPDAM2 Fast Access Extension Conformance
Text of ISO/IEC 15938-7/FDAM2 Fast Access Extension Conformance
23000-4 Musical Slide Show MAF
DoC on ISO/IEC 23000-4/CD (Musical Slide Show MAF)
Text of ISO/IEC 23000-4/FCD (Musical Slide Show MAF)
TuC for ISO/IEC 23000-4 (Musical Slide Show MAF)
23000-8 Portable Video Player MAF
Request of Subdivision of ISO/IEC 23000
WD1.0 of ISO/IEC 23000-8 (Portable Video Player MAF)
23000-9 Digital Multimedia Broadcasting MAF
59
8678
8679
X
8680
8681
8682
X
8683
8684
X
8685
8686
X
8687
X
8688
X
8689
X
8690
X
8691
8608
X
8692
8700
X
8693
8701
X
8694
X
8695
8696
2
Request of Subdivision of ISO/IEC 23000
WD1.0 of ISO/IEC 23000-9 (Digital Multimedia Broadcasting MAF)
23001-1 BinXML
Text on ISO/IEC 23001-1/COR1 (Editorial and technical clarifications)
Study of Text of ISO/IEC 23001-1/PDAM2 (Prefixes and of wild cards extensions)
MPEG-B Part 1 Reference software workplan
23001-2 Fragment Request Unit
DoC on ISO/IEC 23001-2/CD (Fragment Request Unit)
Text of ISO/IEC 23001-2/FCD (Fragment Request Unit)
23001-3 Binary to XML Mapping of IPMP-X
Request of subdivision of ISO/IEC 23001
Text of ISO/IEC 23001-3/CD (Binary to XML Mapping of IPMP-X)
23004-1 MPEG MultiMedia Middleware- Architecture
M3W White Paper : Multimedia Middleware Architecture
23004-2 MPEG MultiMedia Middleware- Multimedia APIs
M3W White Paper : Multimedia API
23004-3 MPEG MultiMedia Middleware- Component Model
M3W White Paper : Component Model
23004-4 MPEG MultiMedia Middleware – Resource and Quality Management
M3W White Paper : Resource and Quality Management
23004-5 MPEG MultiMedia Middleware – Component Download
M3W White Paper : Component Download
Text of ISO/IEC 23004-5/FCD Component Download
23004-6 MPEG MultiMedia Middleware – Fault Management
M3W White Paper : Fault Management
Text of ISO/IEC 23004-6/FCD Fault Management
23004-7 MPEG MultiMedia Middleware – Systems Integrity Management
M3W White Paper : System Integrity Management
Text of ISO/IEC 23004-7/FCD System Integrity Management
23004-8 MPEG MultiMedia Middleware- Reference Software
M3W Reference Software Plan
XXXXX-1 Media Streaming MAF Protocol
Request for New Project on Supplemental Media Technology
Text of ISO/IEC XXXXX-1/CD Media Streaming MAF Protocol
General issues
a.
General
The meeting report from Klagenfurt has been approved.
The following demonstrations have been made:
 M13962: Online database for 3D. Allow to access 3D content for testing and research
purpose. Upload functionality (from VRML, 3GStudioMax, …), converting to
MPEG-4 BIFS automatically.
 Mxxxx: Product demonstration (from France telecom and French Yellow Pages) using
large 3D environment navigation tools developed in MPEG-4 3DGC. AFX compliant.
Demonstration 2D / 3D BIFS integration.
b.
Pr
2
4
List of standards under development
Pt
1
1
Edit. Project
2000 Amd.2
200x Amd.3
Description
Carriage of Auxiliary Data
JPEG 2000 support in
Systems
60
CfP
WD
CD
FCD FDIS
06/04 06/07 07/01
06/04 06/07 07/01
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
5
5
4
5
4
4
4
5
12
15
4
A
A
20
4
8
A
9
B
1
B
B
1
1
B
B
2
3
E
E
E
E
1
2
3
4
E
E
E
5
6
7
E
X
8
1
2004 Amd.17 ATG Conformance
2004 Amd.22 Audio BIFS v3
conformance
2004 Amd.23 Synthesized Texture
conformance
2004 Amd.24 File Format Conformance
2007 Amd.25 LASeR V1 Conformance
2007 Amd.27 LASeR V2 Conformance
2007 Amd.26 Open Font Format
Conformance
2004 Amd.12 File Format Ref. Soft.
2007 Amd.14 Open Font Format Ref.
Soft
2007 Amd.16 Symbolic Music Rep. Ref.
Soft
2007 Amd.17 LASeR Ref. Soft
2005 Amd.2 Flute Hint Track
2005 Amd.2 SCV File Format
Extensions
2004 Amd.1 SVGT1.2 Support
200x 1st Ed.
Musical Slide Show MAF
200x 1st Ed.
Portable Video Player
MAF
200x 1st Ed.
Digital Multi.
Broadcasting MAF
200x Cor.1
Misc. Editorial and
technical clar.
200x Amd.1 Reference Soft. & Conf.
200x Amd.2 Exten. On encoding of
wild cards
st
200x 1 Ed.
Fragment Request Unit
st
200x 1 Ed.
Bin-to-XML Mapping for
IPMPX
st
200x 1 Ed.
Architecture
200x 1st Ed.
Multimedia API
st
200x 1 Ed.
Component Model
200x 1st Ed.
Ressource & Quality
Management
200x 1st Ed.
Component Download
st
200x 1 Ed.
Fault Management
st
200x 1 Ed.
System Integrity
Management
st
200x 1 Ed.
Reference Software
200x
Media Streaming MAF
Protocols
61
06/04 06/07 07/01
06/04 06/07 07/01 07/07
06/07 07/01 07/04 07/10
06/04
06/04
06/10
07/01
06/10
06/10
07/04
07/04
07/04
07/04
07/07
07/10
07/10
07/10
08/01
08/01
06/04 06/10 07/04
07/01 07/04 07/10 08/01
06/10 07/01 07/07 08/01
06/10 07/01 07/07 08/01
05/10 06/07 06/10 07/04
05/10 06/07 07/04 07/10
05/10 06/04 07/01
05/10 06/07 06/10 07/04
06/10 07/01 07/07 08/01
06/10 07/01 07/07 08/01
06/07 NAP
06/10
05/10 06/01 06/07 07/01
06/04 06/07 07/01 07/07
06/04 06/10 07/04
06/10 07/04 07/10
05/01
05/01
05/01
05/01
05/07
05/07
05/07
05/07
06/04
06/04
06/04
06/04
06/07
06/07
06/07
06/07
07/01
07/01
07/01
07/01
05/01 05/07 06/07 06/10 07/07
05/01 05/07 06/07 06/10 07/07
05/01 05/07 06/07 06/10 07/07
07/01 07/07 07/10 08/01
06/10 07/04 07/10
c.
Standing Documents
Pr
1
1
1
2
2
2
2
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
7
7
21
B
E
E
E
E
E
Pt
1
1
1
No.
N7675
N7676
N7677
Meeting
05/07 Nice
05/07 Nice
05/07 Nice
N7678
N7679
N7680
05/07 Nice
05/07 Nice
05/07 Nice
11
1
1
1
1
6
11
12
14
15
Documents
MPEG-1 White Paper – Multiplex Format
MPEG-1 White Paper – Terminal Architecture
MPEG-1 White Paper – Multiplexing and
Synchronization
MPEG-2 White Paper – Multiplex Format
MPEG-2 White Paper – Terminal Architecture
MPEG-2 White Paper – Multiplexing and
Synchronization
MPEG-2 White Paper – MPEG-2 IPMP
MPEG-4 White Paper – MPEG-4 Systems
MPEG-4 White Paper – Terminal Architecture
MPEG-4 White Paper – M4MuX
MPEG-4 White Paper – OCI
MPEG-4 White Paper – DMIF
MPEG-4 White Paper – BIFS
MPEG-4 White Paper – ISO File Format
MPEG-4 White Paper – MP4 File Format
MPEG-4 White Paper – AVC FF
N7503
N7504
N7610
N7921
05/07 Poznan
05/07 Poznan
05/10 Nice
06/01 Bangkok
06/04 Montreux
06/04 Montreux
05/10 Nice
06/04 Montreux
06/01 Bangkok
06/01 Bangkok
13
13
17
18
20
White Paper on MPEG-4 IPMP
MPEG IPMP Extensions Overview
White Paper on Streaming Text
White Paper on Font Compression and Streaming
Presentation Material on LASER
N7505
N6338
N7515
N7508
N6969
20
22
1
1
9
White Paper on LASeR
White Paper on Open Font Format
MPEG-7 White Paper - MPEG-7 Systems
MPEG-7 White Paper – Terminal Architecture
MPEG-21 White Paper – MPEG-21 File Format
N7507
N7519
N7509
N8151
N7925
05/07 Poznan
04/03 München
05/07 Poznan
05/07 Poznan
05/01 HongKong
05/07 Poznan
05/07 Poznan
05/07 Poznan
06/04 Montreux
06/01 Bangkok
X
X
MPEG-B White Paper – BinXML
MPEG Multimedia Middleware Context and
Objectives
1rst M3W White paper
2nd M3W White Paper : Architecture
Tutorial on M3W
M3W White Paper : Multimedia Middleware
Architecture
N7922
N6335
06/01 Bangkok
04/03 München
N7510
N8152
N8153
N8687
05/07 Poznan
06/04 Montreux
06/04 Monreux
06/10 Hanzhou
1
1
1
X
X
X
X
62
N8148
N8149
N7608
N8150
N7923
N7924
E
X
M3W White Paper : Multimedia API
N8688
06/10 Hanzhou
E
X
X
M3W White Paper : Component Model
M3W White Paper : Resource and Quality
Management
M3W White Paper : Component Download
M3W White Paper : Fault Management
M3W White Paper : System Integrity
Management
N8689
N8690
06/10 Hanzhou
06/10 Hanzhou
N8691
N8692
N8693
06/10 Hanzhou
06/10 Hanzhou
06/10 Hanzhou
E
E
E
E
X
X
X
63
d.
Mailing Lists Reminder
Topic
General
Systems
List
BiM
File
Format
LASeR
MAF
e.
Information
Liste Reflector : gen-sys@lists.uni-klu.ac.at
List-Subscribe:
http://lists.uni-klu.ac.at/mailman/listinfo/gen-sys
mailto:gen-sys-request@lists.uniklu.ac.at?subject=subscribe
List-Archive: http://lists.uni-klu.ac.at/pipermail/gen-sys
List-Help: mailto:gen-sys-request@lists.uniklu.ac.at?subject=help
Liste Reflector : mpeg7-sys@lists.uni-klu.ac.at
List-Subscribe:
http://lists.uni-klu.ac.at/mailman/listinfo/mpeg7-sys
mailto:mpeg7-sys-request@lists.uniklu.ac.at?subject=subscribe
List-Archive: http://lists.uni-klu.ac.at/pipermail/mpeg7-sys
List-Help: mailto:mpeg7-sys-request@lists.uniklu.ac.at?subject=help
Liste Reflector : mp4-sys@lists.uni-klu.ac.at
List-Subscribe:
http://lists.uni-klu.ac.at/mailman/listinfo/mp4-sys
mailto:mp4-sys-request@lists.uniklu.ac.at?subject=subscribe
List-Archive: http://lists.uni-klu.ac.at/pipermail/mp4-sys
List-Help: mailto:mp4-sys-request@lists.uniklu.ac.at?subject=help
Liste Reflector : mpeg-laser@lists.uni-klu.ac.at
List-Subscribe:
http://lists.uni-klu.ac.at/mailman/listinfo/mpeg-laser
mailto:mpeg-laser-request@lists.uniklu.ac.at?subject=subscribe
List-Archive: http://lists.uni-klu.ac.at/pipermail/mpeg-laser
List-Help: mailto:mpeg-laser-request@lists.uniklu.ac.at?subject=help
Liste Reflector : maf-sys@lists.uni-klu.ac.at
List-Subscribe:
http://lists.uni-klu.ac.at/mailman/listinfo/maf-sys
mailto:maf-sys-request@lists.uni-klu.ac.at?subject=subscribe
List-Archive: http://lists.uni-klu.ac.at/pipermail/maf-sys
List-Help: mailto:maf-sys-request@lists.uniklu.ac.at?subject=help
FAQ
The FAQ were updated as needed.
64
Kindly Managed
by
University of
Klagenfurt
University of
Klagenfurt
University of
Klagenfurt
University of
Klagenfurt
University of
Klagenfurt
f.
AOB
None.
65
3
MPEG-2 Systems (13818-1)
g.
13818-1:2005 Amd.2 Carriage of Auxiliary Data
i.
Topics
1.
ii.
Carriage of Auxiliary Data
Contributions
None.
Technical Work in Progress.
h.
13818-1:2005 Amd.3 Carriage of SVC
i.
Topics
1.
ii.
Transport of Scalable Video Coding
Contributions
None.
Technical Work in Progress.
4
MPEG-4 Systems (14496-1)
i.
14496-1:2005 Amd.3
i.
Topics
1.
ii.
JPEG 2000 Support in Systems
Contributions
None.
Technical Work in Progress.
j.
14496-1:2005 Cor. 2
i.
Topics
1.
ii.
OD dependencies in MPEG-4 Systems
Contributions
M13813: Summary of Voting on ISO/IEC 14496-1:2004/DCOR 2. Taken into account to finalized COR.
Technical Work Finalized.
5
MPEG-4 Conformance (14496-4)
k.
14496-4 Amd.17
i.
Topics
1.
ATG Conformance
66
ii.
Contributions
None.
Technical Work in Progress.
l.
14496-4 Amd.22
i.
Topics
1.
ii.
Audio BIFS Conformance
Contributions
None.
Technical Work in Progress.
m.
14496-4 Amd.23
i.
Topics
1.
ii.
Synthesized Texture Conformance
Contributions
None.
Technical Work in Progress.
n.
14496-4 Amd.24
i.
Topics
1.
ii.
File Format Conformance
Contributions
M13849 : Updated MP4 Conformance Files from Apple and M14026 Update of ENST ISO
File Format Conformance. Use as the basis for producing standard text. Thank you, and
thanks also to Jean Lefeuvre of ENST for his contributions.
There are some parts of the file format specification for which no conformance file and
reference software was produced :

Degradatation priority table

IPMP Control

Subsample information

Sample scaling

Fragment Random Access

AVC subsequences layering and switch tracks
The Systems subgroup encourage companies having interest in these elements to produce
conformance file oherwise they would be removed from the specification.
Technical Work in Progress.
o.
14496-4 Amd.25 LASeR V1 Conformance
i.
Topics
1.
LASeR Conformance
67
ii.
Contributions
 No progress at this meeting.
 New text will be submitted by the next meeting
 ENST and Streamezzo are exchanging bitstreams
Technical Work in Progress.
p.
14496-4 Amd.2x LASeR V2 Conformance
i.
Topics
1.
ii.
LASeR V2 Conformance
Contributions
M14024 : New sequences for LASeR Amd.1 Conformance. Used as the basis for producing the WD of
LASeR V2 Conformance.
Technical Work in Progress.
6
MPEG-4 Reference Software (14496-5)
q.
14496-5 Amd.12
i.
Topics
1.
ii.
ISO File Format Reference Software
Contributions
M13848 : Updated ISO Base Media File Format Reference Software and M13959: Contribution to ISO
Base Media File Format Reference Software. Used as basis for the production of Amd.12.
M13959: Michael on meta-data track s/w
We appreciate the work and input, and encourage others to follow this example.
Technical Work in Progress.
r.
14496-5 Software for JPEG2000 Support in MPEG-4 Systems
M13864: Implementation for JPEG2000 elementary stream support in MPEG-4 reference
software. Included in 14496-5 Amd.13.
s.
14496-5 Amd.1x
i.
Topics
1.
ii.
LASeR Reference Software
Contributions
M13971 : This contribution contains updated version of LASeR reference software (J2SE
codec software). This software will be uploaded to the CVS. LASeR utility software for
FDIS+COR1+AMD1 will be released by the end of this year.
68
M13896: Implementation of LASeR uDOM Interface in LASeR Player. To be integrated in
the utility software. This contribution presents the implementation of uDOM APIs. This
software will be uploaded to the CVS as a utility software
Technical Work in Progress.
7
MPEG-4 Scene Description (14496-11)
t.
14496-11:2005 Amd.5
i.
Topics
1.
ii.
Symbolic Music Notation
Contributions
M13772, M13854 : Summary of Voting on ISO/IEC 14496-11:2005/FPDAM 5: [SC 29 N 7752],
Editors' Study of ISO/IEC 14496-11/FPDAM5 : Taken as the basis to produce final text.
Technical Work Finalized
69
8
ISO File Format (14496-12)
u.
General
i.
M13757 AHG report
Approved.
v.
14496-12/Amd.1
i.
Topics
1.
ii.
Description of Timed Metadata
M13770 summary of voting on part 12 FPDAM 1
All comments accepted.
iii.
M13959 Meta-data track s/w
Accepted with thanks (and alignment with the revised amendment). We decided not to
implement MPEG-4 systems sample entries for now, as there is no streamType (or
objectTypes) for general meta-data streams.
iv.
Other issues
We adjusted the placement of fields and boxes to make parsing easier (maybe, possible).
Technical Work Finalized.
w.
14496-12/Amd.2
i.
Topics
1.
ii.
Flute Hint Track
M13855 editor's study on flute FF
The editor presented extensive editing as the result of offline sessions with the co-editors.
The item info entry probably needs sorting so it’s an extension. The URI field should say that
it’s “as used to identify the file in the FLUTE session” (not to find it). The content or transfer
length seem redundant with the item location box extent length fields?
This represents a major improvement on the text, with many items clarified.
iii.
M13828 summary of voting on part 12 pdam 2
Mostly accepted. There are lots of questions about rate-share coming up.
iv.
M13931 Tracks in MP4
We’d like to leverage existing tools to solve this interesting problem. We’d like to use track
references to indicate a hint track dependency (“do not send this enhancement unless you are
also sending its base”). Then the track selection mechanisms ought to solve the question of
which hint track(s)_to send.
v.
M13850 rate-share algorithm
Perhaps this is better, but it makes different usage of the numbers. Perhaps we need priority
information on the order of discard (or re-instatement, if the algorithm does that). The
algorithm needs to take into account the quantization, which the previous example did. It also
needs to consider track alternatives - clearly only one of them is included.
70
vi.
(see also M13932, rate share and SVC)
This was presented in full. It’s clear that this uses a different meaning of ‘operating point’
(the original uses this to mean different ranges of available bit-rate; this uses it to mean
different preferences for adaptation).
There is also a worrying degree in which this is dropping into general MPEG-21 DIA, which
is inappropriate, or is not leveraging techniques we already have in pre-computed adaptation,
such as extraction tracks.
vii.
Other issues
We agreed to move section 5.4 from part 15 to part 12. We deleted the mqua type, added
bitrate and frame-rate types, and clarified the table. The text will be adjusted to be in
‘standard-eze’ and remove SVC specificities.
We add a sentence clarifying the field order for fields smaller than a byte.
M13852: SMPTE KLV. The proposition to document the use of SMPTE KLV in the file
format was rejected. However, during the discussions, it appeared that KLV encoding is just
one example of a generic mechanism to identify value with a key, document by ISO (ISO
Labels). It was decided to enable the storage of metadata in such forms at the file format level.
It was also propose to investigate the mapping of SMPTE KLV to MPEG-7 metadata using
tools defined by MPEG-21. This may be investigated further depending of the interest of the
MPEG community. The issue was reported in the plenary.
M13877: AAC SBR timescales and sample rates
Indeed, though the spec. currently only says that ‘should’ match, we should clarify
1. that the timescale should be chosen to match the sampling rate, or be a multiple of it,
to enable sample-accurate timing
2. that if the codec has definitive information about the sampling rate, it must be taken as
definitive; in this case the sampling rate in the sample entry may be ignored, though a
sensible value should be chosen (probably the highest possible sampling rate)
3. that the sampling rate in the sample entry should be considered definitive only for
codecs that do not record their own sampling rate.
A DCOR will be prepared and issued at the next MPEG meeting.
Technical Work in Progress.
9
MPEG-4 AVC File Format (14496-15)
x.
14496-15:2004/Amd.2
i.
Topics
1.
ii.
SVC File Format Extensions
M13821 summary of voting on part 15 pdam 2
This was processed. Concerns over the timing of SVC FDAM and our FPDAM.
iii.
M13851 SVC suggestions
This question still pending: Will the SPS extension array ever be used here? The text should
perhaps clarify this point, in both 5.2.4.1 and perhaps in 5.3.5.1.2, where we have an
AVCConfigurationBox for a parameter set stream. Does it also need an
SVCConfigurationBox (optionally?).
71
Quality layers: pended for 13903/5.
iv.
M13903 On the PDAM2
Lots of good detailed work, all accepted in principle with minor variations. Thank you.
v.
M13905 On meta-data
We accept a lot of these nice clarifications, thank you. But for the use of meta-data outside
the file format, we pend this for further study and dialog with JVT.
vi.
M13955 SVC efficient protection
This raises an important point and we particularly appreciate the experiment and coding. We
certainly want to make sure that we document what indicates layer protection has been
applied, so that early implementers are at least required to notice that some layers are
protected, and not decode the layer as if it were unprotected.
Overall we like this. We’d like to require
 if the base layer (AVC) is protected, then the sample entry MUST be transformed
(4CC change and sinf added);
 if ANY layer is protected in a track, a sinf of some kind must eb added to the sample
entry (this is the ‘warning’ that some protection is in effect). the original format box is
probably still required but redundant as the 4CC need not change
 we allow sinfs in the scalable tier entry, roughly as proposed
 extractors may point to data in protected streams; the byte references are to data ‘on
disc’ (i.e. possibly protected). when protecting, if extractors are permitted are
permitted by the scheme in use, and the protection changes data sizes, then extractors
may need re-writing
 it is the responsibility of the scheme to document IV handling, data-size changes,
whether extractors are permitted, and so on.
vii.
M13932 On rate-share and SVC
We are concerned at the number of structures here, and they seem complex. Also, we wonder
how much of this can be solved with existing structures. Track selection already deals with
the problem outside of rate-share (do I want frame rate or quality?), and we could use tiers to
order the discard within one track. If the discard order must change, then we can build
different extractor tracks with different tier descriptions.
The current tier descriptions are required to be ordered by dependency, but for mutually
independent tiers, we are silent. It seems we should add the statement that it is
recommended (required?) that mutually independent tiers then be ordered by the
‘thinning walk’, that is, one thins/discards from the highest tiers down.
The parallel meta-data track that exists can be used to document how to thin an FGS layer for
consistent quality. It may well be that we need some new meta-data statements and
contributions are welcome.
Technical Work in Progress.
10.Font Compression (14496-18)
i.
Topics
1.
Miscellanea corrections
72
ii.
Contributions
M13822: Summary of Voting on ISO/IEC 14496-18:2004/DCOR 1. All comments where disposed of and
final text was produced.
Technical Work Finalized
73
11.LASeR (14496-20)
a.
General
M13756 AHG report  accepted.
M13794 Sophia AHG meeting report  accepted
b.
14496-20/Cor. 1
i.
Topics
1.
ii.
Misc. Corrigenda
Summary of Voting
M13761 Comments from France and Germany are accepted.
iii.
Schedule
It is decided to delay the publication of FDAM by one meeting because SVGT1.2 was not yet
reached stable status. It is expected to be reached PR by Janunary 2007. It SVGT1.2 is further
delayed AMD would be splited into two pieces to promote SVG independent technologies.
iv.
Contributions
M13773 This contribution contains draft of Study of DCOR1. Accepted as a base of DCOR
Dispositions on other contributions regarding COR1 are as follows:
Number
Topic
Dispositions
13781 
Media
Clipping


Accepted
DCOR
o Correction of namespace of MediaClipping
attribute
13775 
Time
Encoding


Accepted
DCOR
o Clarification on the time encoding (to be
relative)
o New event for “conditional execution time”
13969 
Time
Encoding


Accepted
DCOR
o New attribute in LASeRHeader to signal the
reference of time encoding
13787 
Various Items 

Accepted
DCOR
o Various items
13783 
<a> element & 
linked content 
accepted
DCOR
o Clarification on the replacement according to
74

13973 
<a> element & 
linked content
the type of linked contents
STUDY OF FPDAM
o “security” attribute to control the restriction to
the type of linked contents

DCOR
o Default usage of linked contents is replacement
of existing content
STUDY OF FPDAM
o New attribute about the usage of linked content
whether it is replacement or addition
13785 
Conditional
Execution

Refer to Time Encoding solutions
13782 
Full Screen


Accepted
DCOR
o Clarification on the “fullscreen” attribute
13895 
SAF Stream
Redefiniation

Agreed to remove note on TransientStreamHeader in
DCOR.
DCOR
o Proposed sentence as a normative statement.
o Clarification on repeating the packet with same
AU_SequenceNumber
o Clarification on using AU_SeuqnceNumber per
global SAF stream
o Change name to “SequenceNumber” with
clarifiation that this could be considered as
AU_SequenceNumber in SL point of view
TuC
o “SAFDecoderConfigurationUpdate”


13977 

14030 
Various Items 

Accepted
DCOR
rectClip

DCOR
o Transform does not applied to the clipping
rectangle itself
resume event


Accepted
DCOR
o Change “resume” to “play”
o Clarification of semantics based on the current
status
o Clarification on clipBegin/clipEnd
Technical Work Finalized
75
c.
14496-20/Amd. 1
i.
Topics
1.
ii.
Lightweight Application Scene Representation
Summary of Voting
M13796 Comment from France is accepted.
iii.
Contributions
M13774 This contribution contains draft of Study of FPDAM1. Accepted as a base of
FPDAM1
Dispositions on other contributions regarding AMD1 are as follows:
Number
Topic
Disposition
14011 
SDL


Accepted
STUDY OF FPDAM
o Modifications to SDL
o Text for 13.6 on extension decoding
13788 
animateScroll


accepted
STUDY OF FPDAM
o Extend the coverage of animateScroll to
any media
o Clarifications except “Spacing” attribute
13778 
Update Encoding
mechanism


Accepted
STUDY OF FPDAM
o New update class, ExtensionCommand
13786 
Main stream
identification


Accepted
STUDY OF FPDAM
o Definition of stream dependencies in
safConfiguration
13878, 
13789
Global Streams


Accepted
STUDY OF FPDAM
o globalID can only be assigned to
NonTrnasientStream
o Definition of global streams in
safConfiguration
o NonTransientStream may be released after
the terminal receives the LASeR
command “DiscardStream”
13777 
Animation with
external events

Refer to m13794
13879 
Animation with
external events

Refer to m13794
76
13776 
SAF stream
management

Refer to m13878
13790 
Waiting Tree

Refer to m13970
13970 
Waiting Tree

Accepted
o Add example to 6.7.1
o Add to existing note about “listner”
o Use “uDOM”
13974 
Various Items


Accepted
STUDY OF FPDAM
o “Screen orientation” event
o “Stop” event
o Definition of “activate” event
o Definition of Keys
o Modification of definition of SendEvent
o Definition of IDL for external events with
an example
13898 
GroupingDescriptor 

13920 
Media Events
13976 
cache management 



Accepted
STUDY OF FPDAM
Accepted to remove current streaming events
Final conclusion will be dependent on the
progress of Media Access Event
Accepted
TuC
M13779, M13975: Both contribution contains updated proposal on mini2 profile. Some items
are accepted to TuC. There was no final conclusion on the mini2 profile at this meeting.
2nd Edition:
M13801 This contribution contains draft integrated text for 2nd edition of LASeR
specifications. Accepted as a base document for 2nd edition.
Technical Work in Progress.
d.
LASeR Related Exploration
M14002 This contribution proposes transition effects for Music Slides Show MAF. Transition
effects defined in SMIL and SMPTE are proposed to be used in LASeR. Accepted to TuC for
the further consideration.
M13897 An exploration on MPEG-21 and LASeR. This contribution presents some ideas about the
usage of LASeR in MPEG-21 framework. Output document based on this contribution will be
produced and further investigated during AHG. This contribution is the first technical contribution
on the subject as a follow-up of the Klagenfurt meeting.
M14013 This contribution proposes new requirements and features for video adaptation.
Requirements are not accepted. Standardizing adaptation parameter is out of scope of LASeR
and there is an alternative solution to carry adaptation parameter as a part of xlink:href
77
12.MPEG-7 Conformance (15938-7)
a.
15938-7/Amd.2 Conformance
i.
Topics
1.
ii.
Fast Access Extension
Contributions
The summary of voting on the FPDAM (M13767) was reviewed.
The DoC of FPDAM was produced and reviewed (few technical and editorial fixes).
The final text for FDAM was produced.
Technical Work Finalized
13.MPEG-A MAF (23000)
a.
23000-4 Musical Slide Show MAF
i.
Topics
1.
ii.
Musical Slide Show MAF
Contributions
M13831: Summary of Voting on ISO/IEC CD 23000-4. All comments have been reviewed and disposed
of. See disposition of comments.
M14002: Usage of the transition element for Musical Slide Show MAF. Feature currently in TuC.
Consensus that it should be integrated in LASeR.
M14001: Proposed timed text formt for Musical Slide Show MAF. Proposition accepted. To be included in
the FCD text.
Technical Work in Progress.
b.
23000-8 Portable Video Player MAF
i.
Topics
1.
ii.
Portable Video Player MAF
Contributions
None.
Technical Work in Progress.
c.
23000-9 Digital Multimedia Broadcasting MAF
i.
Topics
1.
ii.
Digital Multimedia Broadcasting MAF
Contributions
None.
Technical Work in Progress.
78
14.MPEG-B
a.
23000-1 Cor.1
i.
Topics
1.
ii.
misc. Editorial and Technical Clarification
Contributions
The summary of voting on the DCOR (M13832) was reviewed: there was no comment to
address, so the DCOR was promoted to COR.
Technical Work Finalized
b.
23000-1 Binary Format Amd.1
i.
Topics
1.
ii.
Reference Software & Conformance
Contributions
None.
Technical Work in Progress.
c.
23000-1 Binary Format Amd.2
i.
Topics
1.
ii.
Extension on Encoding of Wild Cards
Contributions
Contribution M14010 (Editor’s Study of 23001-1/PDAM2) was reviewed and accepted. These are
essentially editorial fixes + missing functionalities in the syntax that were previously approved.
The Study was registered as output of Hangzhou.
Technical Work in Progress.
d.
23000-2 Fragment Request Unit
i.
Topics
1.
ii.
Fragment Request Unit
Contributions
Contribution M13886 (Editor’s study of 23001-2 FRU) was presented and discussed.
o
o
o
o
Informative sections have to be changed to Notes
It’s not clear what the systems layer comprises in this specification, sometimes it should
be understood as “application layer”, so this needs to be clarified in the spec.
FT/Orange feels that it is not possible to design such a system if we do not specify the
format/nature of the response, or at least strong constraints on it. Typically, how would we
define Conformance for this spec? 3 solutions were envisaged:
 there is no conformance, just an upstream message
 we mandate requirements on the response (based on infoSet): abstract definition
of downstream
 we mandate the format response as TeM.
It was decided to define usage of TeM for response to FRU as normative.
79
o
The streaming mode is still unclear, so it was decided to remove streaming functionality
from the current specification and start an amendment at next meeting.
o The sense of requesting the parent of an element should be clarified in the spec.
The draft DoC started in Klagenfurt was updated to produce the final DoC, inline with the results of the
discussions.
The specification was promoted to FCD with 2 month editing period.
Technical Work in Progress.
e.
23000-3 Binary to XML Mapping of IPMP-X
i.
Topics
1.
ii.
Binary to XML Mapping of IPMP-X
Contributions
M13945: Binary to XML Mapping of IPMP-X Messages. Accepted as the basis for the production of CD
text.
80
15.MPEG-E Multimedia Middleware (23004)
a.
Multimedia Middleware
i.
Topics
1.
ii.
MPEG Multimedia Middleware
Contributions
During its 78th meeting (Hangzhou, China, October 23-27, 2006) MPEG has reviewed the following
contributions it received as input to the meeting:
M13833: Summary of Voting on ISO/IEC CD 23004-5
M13834: Summary of Voting on ISO/IEC CD 23004-6
M13835: Summary of Voting on ISO/IEC CD 23004-7
Next to the contributions the following inputs have been available to the meeting:
Study of FCD for ISOIEC 23004 Part 1
Study of FCD for ISOIEC 23004 Part 2
Study of FCD for ISOIEC 23004 Part 3
Study of FCD for ISOIEC 23004 Part 4
Information from the UHAPI Forum regarding UHAPI Bylaws IPR position
Based on the input contributions the following agenda and mandates has been defined for the
M3W BoG meeting of the Systems group during the MPEG meeting in Hangzhou.
 Implement the (editorial) results from the studies of the FCD Parts 1 – 4 (for FDIS
stage)
 Promote Parts 5 – 7 to FCD
 Create White Papers for all 7 M3W parts
 Revise and update reference software WP
 Prepare liaisons statements to ITU-T Focus Group IPTV & UHAPI Forum
Based on these mandates the following tasks have been defined for the Systems M3W BoG during the MPEG
Hangzhou meeting:




Update the FCDs for ISO-IEC 23004 Part 1-4 according the Study results as a
preparation for the promotion from FCD to FDIS at the next MPEG meeting in
Marrakech. These documents are no output for the Hangzhou MPEG meeting. It
should be noted that the study only contained editorial changes to the documents.
Promote the CDs for ISO-IEC 23004 Part 5-7 from CD to FCD. No DoC has been
produced as output document and also no modifications have been done to the CD
documents as no comments or requests for changes have been received on the study
period of the CD. The ballot on all three CD parts (ISO-IEC 23004 Part 5 – 7) resulted
in 19 yes, 0 no and 4 abstentions.
Produce, on request of the convenor and chairs, white papers for all (7) parts of M3W.
Update the reference software work plan by adding sample applications for all the
functional and non-functional M3W interfaces that will be delivered as reference
software.
81


Prepare liaisons statement to ITU-T Focus Group IPTV to update the Focus Group on
the progress in the standardising process and inform them about the planned
extensions to the reference software.
Prepare liaisons statement to UHAPI Forum to update the forum on the progress in the
Standardisation process.
The activities on the Systems M3W BoG during the MPEG Hangzhou meeting have resulted in the following
results:













Text of ISO/IEC 23004-5/FCD Component Download
Text of ISO/IEC 23004-6/FCD Fault Management
Text of ISO/IEC 23004-7/FCD System Integrity Management
White Paper on M3W Part 1 – Architecture
White Paper on M3W Part 2 – Multimedia API
White Paper on M3W Part 3 – Component Model
White Paper on M3W Part 4 – Resource and Quality Management
White Paper on M3W Part 5 – Component Download
White Paper on M3W Part 6 – Fault Management
White Paper on M3W Part 7 – System Integrity Management
M3W Reference Software Plan (updated with sample app’s)
Liaison to ITU-T FG IPTV (WG 6)
Liaison to UHAPI Forum
The following table lists the status of the various M3W Parts & current work plan:
Part-1: Architecture FCD
Part-2: Multimedia API FCD
Part-3: Component Model FCD
Part-4: Resource and Quality Management FCD
Part-5: Component Download CD
Part-6: Fault Management CD
Part-7: System Integrity Management CD
The work-plan for parts 1 – 4 is as follows:
CD:
2006-04
FCD:
2006-07
FDIS:
2007-01
IS:
2007-04
The work-plan for parts 5 – 7 is as follows:
CD:
2006-07
FCD:
2006-10
FDIS:
2007-04
82
IS:
2007-07
Technical Work in Progress.
83
16.MPEG-X Supplementary Media Technology (xxxx)
a.
Media Streaming MAF Protocols
i.
Topics
1.
ii.
Media Streaming MAF Protocols
Contributions
M13946: Media Streaming MAF Protocols. Contribution taken as the basis to produce CD text.
Technical Work in Progress.
84
17.Latest References and Publication Status
Pr
Pt
2
1
2
Standard
No.
ISO/IEC 13818-1/Amd.7
2nd
00/12
1
1
1
1
ISO/IEC 13818-1:2000 (MPEG-2 Systems
1
1
ISO/IEC 13818-1:2000/Amd.2 (Support for IPMP on 2)
1
1
ISO/IEC 13818-1:2000/Amd.4 (Metadata Application CP)
ISO/IEC 13818-1:2000/COR3 (Correction for Field Picture)
2
1
1
1
2
1
ISO/IEC 13818-1:2006 (MPEG-2 Systems 3rd Edition)
2
1
ISO/IEC 13818-1:2006/Amd.1 (Transport of Streaming text)
N8369
2
11
1
1
ISO/IEC 13818-1:2003 (IPMP on 2)
N5607
N2501
N3054
2
2
2
2
2
2
2
2
2
4
4
Issue
Edition)
ISO/IEC 13818-1:2000/COR1 (FlexMux Descr.)
ISO/IEC 13818-1:2000/COR2 (FlexMuxTiming_ descriptor)
ISO/IEC 13818-1:2000/Amd.1 (Metadata on 2) & COR1 on Amd.1
ISO/IEC 13818-1:2000/Amd.3 (AVC Carriage on MPEG-2)
ISO/IEC 13818-1:2000/Amd.5 (New Audio P&L Sig.)
ISO/IEC 13818-1:2000/COR4 (M4MUX Code Point)
ISO/IEC 13818-1:2000/COR5 (Corrections related to 3rd Ed.)
ISO/IEC 14496-1 (MPEG-4 Systems 1st Ed.)
ISO/IEC 14496-1/Amd.1 (MP4, MPEG-J)
01/01 Pisa
01/12 Pattaya
03/07
Trondheim
03/03 Pattaya
03/07
Trondheim
04/10 Palma
04/07
Redmond
04/10 Palma
05/07 Poznan
06/01
Bangkok
N3844
N4404
N5867
N5604
N5771
N6847
N6585
N6845
N7469
N7895
06/xx
06/07
Klagenfurt
03/03 Pattaya
98/10 Atl. City
99/12 Hawaii
85
Status
Doc. with
Purpose
Published
2000/12
ISO
Award
Done
Published
Published
Published
Published
2000/12
2002/03
2002/12
2003/12
Proposed
N/A
N/A
Proposed
Published
Published
2004/03
XXXX
N/A
Proposed
FDAM
FDAM
ITTF
ITTF
to be published
to be published
N/A
N/A
COR
COR
COR
ITTF
ITTF
ITTF
to be published
to be published
to be published
N/A
N/A
N/A
Published
ITTF
FDAM
ITTF
Published
Published
Published
2003/12
1999/12
2001/11
TBP
to be published
TBP
Proposed
Done
Done
4
1
ISO/IEC 14496-1/Cor.1
N3278
4
ISO/IEC 14496-1:2001 (MPEG-4 Systems 2nd Ed.)
N3850
4
1
1
1
1
1
ISO/IEC 14496-1:2001/Cor.3
N4264
N5275
N6587
4
1
ISO/IEC 14496-1:2001/Amd.2 (Textual Format)
N4698
4
1
ISO/IEC 14496-1:2001/Amd.3 (IPMP Extensions)
N5282
4
ISO/IEC 14496-1:2001/Amd.4 (SL Extension)
4
1
1
ISO/IEC 14496-1:2001/Amd.7 (AVC on 4)
N5471
N5976
4
1
ISO/IEC 14496-1:2001/Amd.8 (ObjectType Code Points)
N6202
01/07 Sydney
02/10 Shangai
04/07
Redmond
02/03 Jeju
Island
02/10
Shanghai
02/12 Awaji
03/10
Brisbanne
03/12 Hawaii
4
1
ISO/IEC 14496-1:200x/Amd.1 (Text Profile Descriptors)
N7229
4
ISO/IEC 14496-1:200x/Cor4 (Node Coding Table)
4
1
1
ISO/IEC 14496-1 (MPEG-4 Systems 3rd Ed.)
N7473
N5277
4
1
ISO/IEC 14496-1:200x/Amd.1 (Text Profile Descriptors)
N7229
4
1
ISO/IEC 14496-1:200x/Cor.1 (Clarif. On audio codec behavior)
N8117
4
1
ISO/IEC 14496-1:200x/Amd.2 (3D Profile Descriptor Extensions)
N8372
4
1
ISO/IEC 14496-1:200x/Cor.2 (OD Dependencies)
N8646
4
4
4
00/03
Noordwijk.
01/01 Pisa
Published
2001/11
N/A
Published
Published
COR
COR
COR
2001/11
2002/10
ITTF
ITTF
ITTF
N/A
Done
N/A
N/A
N/A
AMD
ITTF
N/A
Published
2004-05
N/A
Published
Published
2003/12
2004-08
N/A
N/A
AMD
ITTF
to be published
N/A
05/04 Busan
PDAM
ITTF
N/A
05/07 Poznan
02/10
Shanghai
05/04 Busan
PDAM
IS
ITTF
ITTF
Final Text
Editing
to be published
to be published
PDAM
ITTF
COR
ITTF
PDAM
COR
ISO/IEC 14496-1:2001/Amd.1 (Flextime)
ISO/IEC 14496-1:2001/Cor.1
ISO/IEC 14496-1:2001/Cor.2
06/04
Montreux
06/07
Klagenfurt
06/10
Hangzhou
86
N/A
Proposed
N/A
ITTF
Final Text
Editing
Final Text
Editing
to be published
ITTF
to be published
N/A
N/A
N/A
4
ISO/IEC 14496-6:2000
4
6
8
11
4
11
ISO/IEC 14496-11/Amd.1 (AFX)
N5480
02/03 Jeju
05/01
HongKong
02/12 Awaji
4
11
ISO/IEC 14496-11/Amd.2 (Advanced Text and Graphics)
N6205
4
ISO/IEC 14496-11/Cor.1
4
11
11
N6203
ISO/IEC 14496-11/Cor.3 Valuator/AFX related correction N6594
4
11
ISO/IEC 14496-11/Amd.3 Audio BIFS Extensions
N6591
4
11
ISO/IEC 14496-11/Amd.4 XMT and MPEG-J Extensions
N6959
4
11
ISO/IEC 14496-11/Cor.3 (Audio BIFS Integrated in 3rd Edition)
N7230
4
11
ISO/IEC 14496-11/Cor.5 (Misc Corrigendum)
N8383
4
11
N8657
4
12
ISO/IEC 14496-11/Amd.5 Symbolic Music
Representation
ISO/IEC 14496-12 (ISO Base Media File Format)
4
12
ISO/IEC 14496-12/Amd.1 ISO FF Extension
N6596
4
12
N7232
4
12
ISO/IEC 14496-12/Cor.1 (Correction on File Type
Box)
ISO/IEC 14496-12/Cor.2 (Miscellanea)
4
12
ISO/IEC 14496-12/Amd.1 (Description of timed
metadata)
N8659
4
ISO/IEC 14496-8 (MPEG-4 on IP Framework)
ISO/IEC 14496-11 (MPEG-4 Scene Description 3rd
Edition)
N4712
N6960
2000/12
2004-05
SC29
FDAM
ITTF
03/12 Hawaii
FDAM
ITTF
03/12 Hawaii
04/07
Redmond
04/07
Redmond
05/01
HongKong
05/04 Busan
COR
COR
SC29
ITTF
FDAM
ITTF
FDAM
ITTF
COR
ITTF
COR
SC29
N/A
FDAM
ITTF
TBP
Published
2004-02
Proposed
FDAM
ITTF
FDAM 04/11/30
N/A
COR
ITTF
N/A
COR
ITTF
Final Text
Editing
Final Text
Editing
FDAM
ITTF
06/07
Klagenfurt
06/10
Hangzhou
02/10
Shanghai
04/07
Redmond
05/04 Busan
N5295
06/01
Bangkok
06/10
Hangzhou
N7901
87
N/A
Proposed
Proposed
Published
Published
FDIS
Final Text
Editing
Integration in 1st
Ed.
Integration in 1st
Ed.
st
Integration in 1
Ed.
Integration in 1st
Ed.
Integration in 1st
Ed.
Final Text
Editing
N/A
N/A
N/A
N/A
Proposed
N/A
N/A
N/A
N/A
4
13
ISO/IEC 14496-13 (IPMP-X)
N5284
4
14
ISO/IEC 14496-14 (MP4 File Format)
N5298
4
14
ISO/IEC 14496-14/Cor.1 (Audio P&L Indication)
N7903
4
15
ISO/IEC 14496-15 (AVC File Format)
N5780
4
15
ISO/IEC 14496-15/Amd.1 (Support for FREXT)
N7585
4
15
15
ISO/IEC 14496-15/Cor.1
ISO/IEC 14496-15/Cor.2 (NAL Unit Restriction)
N7575
N8387
17
18
18
N7479
N6215
N8664
02/10
Shanghai
02/10
Shanghai
06/01
Bangkok
03/07
Trondheim
05/10 Nice
4
19
20
18
4
22
ISO/IEC 14496-17 (Streaming Text)
ISO/IEC 14496-18 (Font Compression and Streaming)
ISO/IEC 14496-18/Cor.1 (Misc. corrigenda and
clarification)
ISO/IEC 14496-19 (Synthesized Texture Stream)
ISO/IEC 14496-20 (LASeR)
ISO/IEC 14496-20/Cor.1 (Misc. corrigenda and
clarification)
ISO/IEC 14496-22 (Open Font Format)
7
1
ISO/IEC 15938-1 (MPEG-7 Systems)
N4285
05/10 Nice
06/07
Klagenfurt
05/07 Poznan
03/12 Hawaii
06/10
Hangzhou
03/12 Hawaii
05/10 Nice
06/10
Hangzhou
06/07
Klagenfurt
01/07 Sydney
7
1
1
1
1
2
ISO/IEC 15938-1/Amd.1 (MPEG-7 Systems Extensions)
N6326
N6328
N7490
N7532
N4288
04/03 Munich
04/03 Munich
05/07 Poznan
05/10 Nice
01/07 Sydney
4
4
4
4
4
4
7
7
7
7
ISO/IEC 15938-1/Cor.1 (MPEG-7 Systems Corrigendum)
ISO/IEC 15938-1/Cor.2 (MPEG-7 Systems Corrigendum)
ISO/IEC 15938-1/Amd.2 (BiM extension)
ISO/IEC 15938-2 (MPEG-7 DDL)
N6217
N7588
N8666
N8395
88
to be published
Proposed
IS
ITTF
Published
2003-11
COR
ITTF
Published
2004-04
FDAM
ITTF
COR
COR
ITTF
ITTF
N/A
N/A
FDAM
Published
COR
ITTF
2004-07
ITTF
TBP
Proposed
N/A
Published
FDAM
COR
2004-07
Editor
ITTF
Proposed
TBP
N/A
FDAM
Editor
Published
2002/07
FDAM
COR
COR
FDAM
Published
ITTF
Editor
ITTF
ITTF
2002/02
Proposed
Final Text
Editing
N/A
Proposed
Final Text
Editing
Final Text
Editing
N/A
TBP
Done
FDAM 04/11/28
N/A
N/A
N/A
N/A
Done
7
7
ISO/IEC 15938-7/Amd.2 (Fast Access Ext. Conformance)
N8672
21
9
ISO/IEC 21000-9 (MPEG-21 File Format)
N6975
21
16
1
1
ISO/IEC 21000-16 (MPEG-21 Binary Format)
ISO/IEC 23001-1 (XML Binary Format)
ISO/IEC 23001-1/Cor.1 (Misc. Editorial and technical
clar.)
N7247
N7597
N8680
B
B
06/10
Hangzhou
05/01
HongKong
05/04 Busan
05/10 Nice
06/10
Hangzhou
89
N/A
FDAM
ITTF
FDIS
ITTF
FDIS 05/01/21
Done
FDIS
FDIS
COR
ITTF
ITTF
ITTF
FDIS 05/04/22
TBP
TBP
N/A
18.Resolutions of Systems
i.
Cf. WG11 resolution.
19.List of Reviewed Contributions
Title
N°
13756
AHG on Scene Representation
13757
AHG on MPEG File Formats
13761
13785
13786
13787
13788
13789
Summary of Voting on ISO/IEC 14496-20:2006/DCOR 1 [SC 29
N 7729]
Summary of Voting on ISO/IEC 15938-7:2003/FPDAM 2 [SC 29
N 7741]
Summary of Voting on ISO/IEC 14496-12:2005/FPDAM 1 and
ISO/IEC 15444-12:2005/FPDAM 1 [SC 29 N 7749]
Summary of Voting on ISO/IEC 14496-11:2005/FPDAM 5: [SC
29 N 7752]
Draft study of DCOR1, on going
Draft study of FPDAM1, on going
Time encoding issues for discussion at Sophia AHG
Stream management issues for discussion at Sophia AHG
Events and animation issues for discussion at Sophia AHG
Update extensibility issues for discussion at Sophia AHG
mini2 improvements for discussion at Sophia AHG
On SMIL MediaClipping in LASeR for discussion at Sophia AHG
On fullscreen video for discussion at Sophia AHG
On the usage of the a element in LASeR for discussion at Sophia
AHG
On LASeR Conditional Execution
On SAF Configuration
Comments on LASeR and SAF DCOR
On LASeR animateScroll
On SAF global streams
13790
On LASeR Waiting Tree
13794
13796
Report of LASeR AHG meeting in Sophia
Summary of Voting on ISO/IEC 14496-20:2006/FPDAM 1 [SC 29
N 7781]
Liaison Statement from W3C [SC 29 N 7782]
Draft LASeR 2nd edition (DCOR1 + FPDAM1)
Summary of Voting on ISO/IEC 14496-1:2004/DCOR 2
Summary of Voting on ISO/IEC 14496-15:2004/PDAM 2
Summary of Voting on ISO/IEC 14496-18:2004/DCOR 1
Summary of Voting on ISO/IEC 14496-12/PDAM 2 & 1544412/PDAM 2
Summary of Voting on ISO/IEC CD 23000-4
Summary of Voting on ISO/IEC 23001-1:2006/DCOR 1
Summary of Voting on ISO/IEC CD 23004-5
Summary of Voting on ISO/IEC CD 23004-6
13767
13770
13772
13773
13774
13775
13776
13777
13778
13779
13781
13782
13783
13798
13801
13813
13821
13822
13828
13831
13832
13833
13834
90
Authors
Young-Kwon Lim
Cyril Concolato
David Singer
Visharam Mohammed
SC 29 Secretariat
SC 29 Secretariat
SC 29 Secretariat
SC 29 Secretariat
Jean-Claude Dufourd
Jean-Claude Dufourd
Jean-Claude Dufourd
Jean-Claude Dufourd
Jean-Claude Dufourd
Jean-Claude Dufourd
Jean-Claude Dufourd
Jean-Claude Dufourd
Jean-Claude Dufourd
Jean-Claude Dufourd
Jean Le Feuvre
Jean Le Feuvre
Jean Le Feuvre
Jean Le Feuvre
Cyril Concolato
Jean Le Feuvre
Cyril Concolato
Jean Le Feuvre
Young-Kwon Lim
SC 29 Secretariat
W3C via SC 29 Secretariat
Jean-Claude Dufourd
SC 29 Secretariat
SC 29 Secretariat
SC 29 Secretariat
SC 29 Secretariat
SC 29 Secretariat
SC 29 Secretariat
SC 29 Secretariat
SC 29 Secretariat
13848
13849
13850
13851
13852
13854
Title
Summary of Voting on ISO/IEC CD 23004-7
SMPTE Liaison to JTC1/SC29 - SMPTE 421M ISO Base Media
File Format
Updated ISO Base Media File Format Reference Software
Updated MP4 Conformance Files from Apple
An improved rate-share algorithm for the ISO Base file format
Comments and suggestions on the SVC File Format draft
SMPTE KLV meta-data in ISO Base Media File format files
Editors' Study of ISO/IEC 14496-11/FPDAM5
13855
Study on 14496-12:2005/PDAM2 ALC/FLUTE server file format
13877
13878
On AAC SBR storage in ISO Media File
Discussion on SAF global streams
13879
On LASeR Fraction events
13885
Liaison Statement from FLO Folum [SC 29 N 7821]
13886
Editors input on 23001-2 FRU
13895
13896
On SAF streams redefinition
Implementation of LASeR uDOM Interface in LASeR Player
13897
An exploration on MPEG-21 and LASeR
13898
Improved text for GroupingDescriptor
13903
Comments on the AVC file format PDAM2 document
13905
Improvements of SVC file format meta data statements
13920
13931
On LASeR Events
Track relationship in file format
13932
Generic adaptation path in file format
13945
13946
13955
Binary to XML Mapping of IPMP-X Messages
Media Streaming MAF Protocols
Proposed Extension to SVC File Format for Efficient and Effective
Protection
N°
13835
13840
91
Authors
SC 29 Secretariat
SMPTE
David Singer
David Singer
David Singer
David Singer
David Singer
Pierfrancesco Bellini
Mauzio Campanai
Paolo Nesi
Per Fröjdh
Thorsten Lohmar
Miska Hannuksela
Imed Bouazizi
Jean Le Feuvre
Jean Le Feuvre
Jean-Claude Dufourd
Jean Le Feuvre
Cyril Concolato
FLO Folum via SC 29
Secretariat
Stephen Davis
Gerrard Drury
Jean Le Feuvre
YeSun Joung
Young-kwon Lim
Won-sik Cheong
Jihun Cha
KyungAe Moon
YeSun Joung
Young-kwon Lim
Won-Sik Cheong
Jihun Cha
KyungAe Moon
Youngjoo Song
Young-Kwon Lim
Jechang Jeong
Thomas Ragthen
Peter Amon
Andreas Hutter
Thomas Rathgen
Peter Amon
Andreas Hutter
Jean Le Feuvre
Miska M. Hannuksela
Ye-Kui Wang
Ye-Kui Wang
Miska M. Hannuksela
Hendry
Munchurl Kim
Sangjin Hahm
Keunsik Lee
Keunsoo Park
13959
Title
Contribution to ISO Base Media File Format Reference Software
13969
Final word on the encoding of times in LASeR
13970
Elements for the clarification of the waiting tree concept in LASeR
13971
13973
LASeR reference software release and status
On communication channels management with LASeR
13974
13975
Request for promotion of some TuC to LASeR AMD1
Update of the proposed LASeR mini2 profile
13976
On a new caching instruction for SAF
13977
14001
On a few missing fixes to LASeR docs
Proposed timed text formt for Musical Slide Show MAF
14002
Usage of the transition element for Musical Slide Show MAF
14010
Editor's Study of 23001-1 PDAM2
14011
14012
Fixes on LASeR Amd1
Use case and requirement for LASeR
14013
New feature for LASeR
14024
14026
New sequences for LASeR Amd.1 Conformance
Update of ENST ISO File Format Conformance
N°
92
Authors
Michael Ransburg
Hermann Hellwagner
Jean-Claude Dufourd
Nicolas Pierre
Elouan Le Coq
Cyril Concolato
Jean Lefeuvre
Jean-Claude Dufourd
Nicolas Pierre
Jean-Claude Dufourd
Jean-Claude Dufourd
Elouan Le Coq
Jean-Claude Dufourd
Jean-Claude Dufourd
Nicolas Pierre
Jean-Claude Dufourd
Elouan Le Coq
Jean-Claude Dufourd
Tae Hyeon Kim
H. Jean Cha
Tae Hyeon Kim
H. Jean Cha
Philippe de Cuetos
Gregoire Pau
Cedric Thienot
Philippe de Cuetos
Sylvain Devillers
Renaud Cazoulat
Sylvain Devillers
Renaud Cazoulat
Jean-Claude Dufourd
Jean Lefeuvre
Annex G – MDS report
Source: Ian S Burnett, PhD, Chair
1.0 Introduction
MDS commenced with an overview of the weeks planned activities:
Indications in green = Live content
Indications in white = Edit in master
Indications in blue
= Locked elements
Indications in black = Optional
elements
• Group name:
17pt Arial Regular, white
• IBM logo must not
be moved, added
to, or altered in
any way.
MPEG Multimedia Description Schemes (MDS) Sub-group
Maximum length: 1 line
• Presentation title:
28pt Arial Regular, black
Recommended maximum
length: 2 lines
Kick-off Multimedia Description
Schemes (MDS) Activities
78th MPEG Meeting
Hangzhou, CH
Ian S Burnett, Chair, MPEG MDS Group
October 23rd-27th, 2006
• Presentation subtitle:
20pt Arial Regular,
teal R045 | G182 | B179
Recommended
maximum length: 2 lines
Indications in green = Live content
Indications in white = Edit in master
Indications in blue
July 24th, 2005
Template release: Oct 02
For the latest, go to http://w3.ibm.com/ibm/presentations
• Group name:
14pt Arial Regular, white
Maximum length: 1 line
• Slide heading:
28pt Arial Regular,
blue R120 | G137 | B251
= Locked elements
Indications in black = Optional
elements
• Confidentiality/date
line:
13pt Arial
Regular,
white
MPEG Multimedia
Description
Schemes
(MDS)
Sub-group
Maximum length: 1 line
• Information separated by vertical strokes,
with two spaces on either side
• Disclaimer information may also be appear in this area. Place
flush left, aligned at bottom, 8-10pt Arial Regular, white
• Copyright: 10pt Arial
Regular, white
Overview of MDS Activities
• Background should
not be modified.
Maximum length: 2 lines
• Slide body:
18pt Arial Regular, black
Square bullet color:
teal R045 | G182 | B179
Recommended maximum
text length: 5 principal
points
MPEG-21 & MAFs:
• REL (“Profiles”) – Study of FPDAM/2, Open
Release
• RDD (COR/2)
• DIA FDAM/2
• Reference s/w (Study of CD)
• IPMP Components (FPDAM/1, Conf/Ref s/w)
• DI Streaming (FDIS CEs PDAM/1)
• MP MAF – (FCD)
• Media Streaming MAF – (CD)
• Audio Archival MAF (WD)
• MAFs – joint meetings with Reqts/Audio/systems
• MPEG-21 Schema Doc
July 19—24, 2004
Optional slide number:
10pt Arial Bold, white
|
69th MPEG Meeting
|
Redmond, WA USA
• Title/subtitle/confidentiality line: 10pt Arial Regular, white
Maximum length: 1 line
Information separated by vertical strokes,
with two spaces on either side
93
• IBM logo must not
be moved, added
to, or altered in
any way.
© 2003 IBM Corporation
• Copyright: 10pt
Arial
Regular, white
Indications in green = Live content
Indications in white = Edit in master
Indications in blue
• Group name:
14pt Arial Regular, white
Maximum length: 1 line
• Slide heading:
28pt Arial Regular,
blue R120 | G137 | B251
Maximum length: 2 lines
= Locked elements
Indications in black = Optional
Template release: Oct 02
For the latest, go to http://w3.ibm.com/ibm/presentations
elements
• IBM logo must not
be moved, added
to, or altered in
any way.
MPEG Multimedia Description Schemes (MDS) Sub-group
MPEG-21 & MPEG-A Timeline
St Pt Edit
Project
d
.
21 4 2004 Amd.1
21
4
2004 Amd.2
• Slide body:
18pt Arial Regular, black
21
5
2004 Amd.2
Square bullet color:
teal R045 | G182 | B179
21
7
2004 Amd.2
Recommended maximum
text length: 5 principal
points
21 8 200x
21 14 200x 1st Ed.
21 18 200x 1st Ed.
A
A
2
5
200x 2nd Ed.
200x
A
6
200x
Description
CfP
IPMP Components –
Base Profile
IPMP Components –
Media Streaming
Profile
REL profiles - the
DAC profile
Dynamic &
distributed adaptation
Reference software
Conformance
Digital Item
Streaming
Music player MAF
Media Streaming
MAF
Audio Archival MAF
WD
CD
FCD
FDIS
• Background should
not be modified.
06/07 07/10 07/04
06/07
07/04 07/07 08/01
06/01 06/07 07/01
05/10 06/04 06/10
03/10
06/07 06/10 07/04
06/04 06/10 07/04
05/10 06/04 06/10
06/04
06/07
06/07 06/10 07/04
07/04 07/07 08/01
Indications
content
06/07
07/04in green
07/07= Live
08/01
Indications in white = Edit in master
Indications in blue
19—24,
2004 | 69th MPEG Meeting | Redmond, WA USA
Template release:JulyOct
02
For the latest, go to http://w3.ibm.com/ibm/presentations
• Group name:
14pt Arial Regular, white
Maximum length: 1 line
• Slide heading:
28pt Arial Regular,
blue R120 | G137 | B251
= Locked elements
Indications in black
= Optional
© 2003
IBM Corporation
elements
• Title/subtitle/confidentiality line: 10pt Arial Regular, white
Optional slide number:
Maximum
length: 1 lineSchemes (MDS) Sub-group
10pt Arial Bold, MPEG
white Multimedia
Description
• Copyright: 10pt
Arial
Regular, white
Information separated
by vertical
strokes,
Major MDS goals
of the
week
with two spaces on either side
 MPEG-21 IPMP Components (Part 4):
•
•
Maximum length: 2 lines
• IBM logo must not
be moved, added
to, or altered in
any way.
• Background should
not be modified.
Discussions on Profiles
Output: FPDAM/1
 MPEG-21 REL (Part 5):
•
•
• Slide body:
18pt Arial Regular, black
Square bullet color:
teal R045 | G182 | B179
Profiles, Ref s/w, Open Rel?
Output: study of FDAM/2, Ref s/w plan
 MPEG-21 RDD (Part 6):
•
•
Recommended maximum
text length: 5 principal
points
Discussions, reviewing of inputs & finalise
Output: Corrigendum 2
 MPEG-21 Digital Item Adaptation (Part 7):
•
•
Review NB Comments, Inputs
Output: FDAM/2
Indications in green = Live content
Indications in white = Edit in master
Indications in blue
19—24,
2004 | 69th MPEG Meeting | Redmond, WA USA
Template release:JulyOct
02
For the latest, go to http://w3.ibm.com/ibm/presentations
• Group name:
14pt Arial Regular, white
Maximum length: 1 line
• Slide heading:
28pt Arial Regular,
blue R120 | G137 | B251
Maximum length: 2 lines
• Slide body:
18pt Arial Regular, black
Square bullet color:
teal R045 | G182 | B179
Recommended maximum
text length: 5 principal
points
= Locked elements
Indications in black
= Optional
© 2003 IBM Corporation
elements
• Title/subtitle/confidentiality line: 10pt Arial Regular, white
Optional slide number:
Maximum
length: 1 lineSchemes (MDS) Sub-group
Description
10pt Arial Bold, MPEG
white Multimedia
• Copyright: 10pt
Arial
Regular, white
Information
separatedgoals
by verticalof
strokes,
Major MPEG-21
& MAF
the week (cont.)
with two spaces on either side

MPEG-21 Ref s/w 2nd edn (Part 8):

MPEG-21 Conformance (Part 14):

MPEG-21 DI Streaming (Part 18)

MPEG-21 Schemas output document

MAF – Consolidated Music Player
New s/w: DID v2, DIP, DIP amd/1, DII amd/1, REL Profiles, IPMP Components, RDD Utility s/w, DIA Amds, ER, FID,
FF(?), DIS: Study of CD?
•
•
•
•
•
•
•
•

Discussion of CE results/inputs, NB Comments
Output: FDIS, updated TuC, CE workplan, Create AMD/1
Host on ITTF site
Working Document – output kept up to date (DIA, DIS)
Consolidated Music Player MAF CD
Output:FCD, Ref s/w workplan
MAF – Media Streaming
•
•

Inputs
Output: Study of CD ?
Inputs, AHG inputs
Output: CD, Ref s/w workplan, TuC
MAF – Audio Archival
•
•
Inputs, AHG inputs
Output: WD, Ref s/w workplan
July 19—24, 2004
Optional slide number:
10pt Arial Bold, white
|
69th MPEG Meeting
|
Redmond, WA USA
• Title/subtitle/confidentiality line: 10pt Arial Regular, white
Maximum length: 1 line
Information separated by vertical strokes,
with two spaces on either side
94
© 2003 IBM Corporation
• Copyright: 10pt
Arial
Regular, white
• IBM logo must not
be moved, added
to, or altered in
any way.
• Background should
not be modified.
2.0 Notes on discussions on Input Documents
During the meeting, MDS created three Break Out groups to better manage activities. These
considered DIA matters (finalizing the FDAM/2), REL Profile matters and DIS matters (finalizing
the FDIS and creating the new Amendment 1). Other inputs were dealt with as follows:
Treatment of MDS Input Documents
MPEG-21 RDD (14h30 - 15h30)
MDS Room Wenqi
13836
Jaime Delgado Eva Rodriguez Final Comments on the Ontological Analysis of the
Study of DCOR/2 of ISO/IEC 21000-6
Marc Gauvin
13837
ESNB position paper: Problems with the
inconsistency of the MPEG-21 Rights Data
Dictionary
Francisco Morán
Inputs:
This input considered the implementation of a proposal to solve issues from the previous meeting. It
found that 2 issues remained and that the changes proposed resulted in some further issues. These
were solved using an ad hoc solution. One concern is that the semantic consequences of the
achievement of a consistent RDD are not fully understood in terms of the semantic of the RDD.
The Spanish NB position paper supports the input 13836
Actions:
Add the three triples that have been proposed in 13836. This will create a ‘consistent’ RDD. MDS
then agreed to add text to the RDD stating that “The Terms in ISO/IEC 21000-6 are presented in the
form of an ontology, however, ISO/IEC 21000-6 does not intend to explicitly express all
relationships between Terms.”. The Corrigenda will be completed with this.
MDS will also propose that Requirements consider new work to address a new Rights Ontology
which will standardize implementation and messaging interfaces. Within this work, we expect that
further definition of the term ‘Rights’ will be required. MDS expects the scope of this work to be
substantially different in focus to that of the current 21000-6 RDD.
MDS proposed the Requirements Chair consider the work and he invited inputs from the requesting
parties to the 79th meeting.
13991 Thomas Skjølberg Peder Drege
Delivery of dynamic resources in Digital Item Streaming
Inputs:
Input on the concepts of Live Sources and possible changes to the DI model. Proposes changes to
document using BiM operators – Insert, Delete, Replace. There is an issue with the carouselling in
the current ref s/w. There needs to be a way to handle changing resources. The input has draft XML
for the mechanism.
Actions:
95
Concerns were raised about exactly what was being proposed. It was clarified teat instructions are
for the BBL processor and that the real issue is with what should be standardized. It was decided
that use-cases are needed to fully understand this work.
MAFs (16h00 - 17h30)
13913
MDS Room Wenqi
Study Text on ISO/IEC CD 23000-2 MPEG-A Music
Player 2nd edition
Harald Fuchs
Input
This input gives improved study text for the Music Player 2nd edition. The playlist format section
was improved.
Actions
The reorganization of the document which clarifies and improves intelligibility was accepted.
13943
AHG on MAFs Under Development
13945
AHG on MAFs Under Development
13946
AHG on MAFs Under Development
Proposal of Updated Working Draft of ISO/IEC 23000-5
Media Streaming Player
Proposal of Updated Working Draft of IPMP Extensions
XML Messages
Proposal of Updated Working Draft of Media Streaming
MAF Technologies
Input
13945 was transferred to Systems following the Chairs meeting.
13943 was considered following the discussion of the AHG on the Sunday prior to the meeting.
A discussion was had on MPEG-21 DI mapping to the ISO fileformat and interoperability of MAF
files. Further discussion was held over to the joint meeting on the File format.
Actions
One question was whether the MAF is specifying too much – beyond a Music Player MAF? It was
suggested that the standard part of this MAF need to be clearly delimited and separated. The BoG
will consider the reorganization of the document and the documents (both 13943 and 13946) will be
considered again on Wednesday am.
Tuesday
For the duration of Tuesday, joint meetings were held with Requirements discussing MPEG-21
Profiles, new MPEG-7 descriptors and MAFs. See the Requirements report for details on these
activities.
Wednesday
MPEG-21 DIA - BS SCHEMA (11h30 - 12h00)
13963
Joint with Video in MDS
Davy De Schrijver
Wesley De Neve Davy
Van Deursen Saar De
Zutter Rik Van de
Walle
96
An MPEG-21 BS Schema for the scalable
extension of MPEG-4 AVC version 6 (Joint
Scalable Video Model 6)
Input
This input overviewed the contribution and gave a demo of the BSDL schema and the
implementation. It gives a schema for AVC JSVM 6. The input used context attributes from the
FDAM, and other extensions. The input used a modified version of the BSDL 1.2.1 software with
context attributes and emulation prevention bytes. The input also uses STX stylesheets for extra
adaptation.
Actions
Proposing schema and style sheets be added to the repository. These will be added to the reference
software.
Joint meeting with Audio on MAFS – see the Audio report for this section (also refer to later
MDS sessions on the MP MAF and the Professional Archival MAF)
Chris Poppe Saar De Zutter Rik Van de Contribution to Utility Software for ISO/IEC 21000-10
13965 Walle
DIP/AMD 1
Input
This input is an example of the C++ bindings in the DIP amendment. It demonstrates te invokement
and execution on a Java based terminal.
Actions
Include in the MPEG-21 reference software as Utility software.
Saar De Zutter Frederik De Keukelaere
Gerrard Drury Christian Timmerer Xin Editor’s input to ISO/IEC 21000-8 Reference Software (Second
13968 Wang
Edition)
Input
This input gives editorial changes to the Reference software, fixed the layout and added new inputs
– 21000-12, 21000-15, 21000-17
Actions
MDS accepted the editors input.
Saar De Zutter Sylvain Devillers Thomas
13972 DeMartini Andrew Tokmakoff
Editor's input to ISO/IEC 21000-14 Conformance Testing
Input
This input added new ER conformance software.
Actions
Include in the Conformance FCD from this meeting.
97
Saar De Zutter Davy De Schrijver Rik
13978 Van de Walle
Saar De Zutter Chris Poppe Davy De
13979 Schrijver Rik Van de Walle
Update to Reference Software for Conformance to ISO/IEC
21000-10
Update to Reference Software for Conformance to ISO/IEC
21000-10/Amd 1
Input
This input restructured the location of conformance streams and gave extra explanation on the use
of the reference software.
Actions
Include in the Reference Software Study of CD from this meeting.
13980 Saar De Zutter Rik Van de Walle
Contribution to summary and 1-pager of Enhanced
Interoperability for MPEG-21 Session Mobility using DIP
Input
This input gives a Session Mobility summary /1 pager.
Actions
Include on the one pager WWW site.
Saar De Zutter Davy De Schrijver Rik
13981 Van de Walle
Update to summary of Digital Item Technologies: Digital
Item Processing
Input
This input gives a Digital Item Processing summary /1 pager.
Actions
Include on the one pager WWW site.
Saar De Zutter Chris Poppe Davy De
13982 Schrijver Rik Van de Walle
Contribution to summary and 1-pager of Digital Item
Technologies: Digital Item Processing Amd 1
Input
This input gives a new Digital Item Processing Amd 1 summary /1 pager.
Actions
Include on the one pager WWW site.
Saar De Zutter Davy De Schrijver Rik
13983 Van de Walle
Contribution to summary and 1-pager of Conformance: MPEG21 Digital Item Processing
Input
98
This input gives a Digital Item Processing Conformance summary /1 pager.
Actions
Include on the one pager WWW site.
Saar De Zutter Chris Poppe Davy De
13984 Schrijver Rik Van de Walle
Contribution to summary and 1-pager of Conformance: MPEG21 Digital Item Processing Amd 1
Input
This input gives a new Digital Item Processing Amd 1 Conformance summary /1 pager.
Actions
Include on the one pager WWW site.
Saar De Zutter Davy De Schrijver Rik
13985 Van de Walle
Update to summary and 1-pager of Reference Software: MPEG21
Input
This input gives an updated Reference Software summary /1 pager.
Actions
Include on the one pager WWW site.
Michael Eberhard Michael Sablatschan
13988 Christian Timmerer
gBSDtoBin (MPEG-21 DIA) reference software update
Input
This input implements the DIA AMD/2 gBSDtoBin updates in the reference software and also
provides utility software.
Actions
Include this in the Study of CD for Reference Software.
Hyon-Gon Choo Filippo Chiariglione
13949 Bum-Suk Choi
Proposed Working Draft of ISO/IEC 21000-4/Amd 2 Media
Streaming Profile
Input
This input proposes improvements to the new profile of IPMP Components for media Streaming.
The input had been improved at the weekend AHG meeting.
Actions
Concerns were raised over whether this was really a profile. The contributors were asked to discuss
the issue with other experts and bring some proposal back to a meeting of MDS on Thursday.
13956 Hendry Takafumi Ueno
Editor’s Study of ISO/IEC 21000-4/PDAM 1: IPMP Base Profile
99
13957 Hendry Munchurl Kim
Contribution to ISO/IEC 21000-4/PDAM 1: IPMP Base Profile
Reference Software
Input
These inputs give the editors input on the IPMP Base Profile and a further contribution of the
reference software for this profile.
Actions
The editorial changes were accepted.
MDS also considered the DoC.
The reference software for the IPMP Base profile was accepted and will be added to the Study of
Reference software and has already been uploaded to the reference software.
Joint meeting with Systems on LASeR and MPEG-21 – see Systems report
Thursday
Joint meeting with Video, ISG, Systems on RVC – see Video Report
Joint meeting with Systems on Metadata Conversion – see Systems Report
Session on MS MAF.
The MDS group looked at the proposed IPMP Components Profile and asked that a clear set of
Requirements for changes/improvements in IPMP Components be created. These were to be
reported back to MDS on Thursday pm. Two requirements were created but another was required
and further work was then undertaken before a joint meeting with Requirements. The Requirements
for the MS MAF usage of IPMP Components were clearly established (see Requirements report and
output). MDS decided to generate three outputs for the MS MAF:
1. CD of MS MAF Player
2. TuC of MS MAF IPMP technologies
3. Reference software workplan
It was noted that the TuC IPMP technologies needed to be studied by MDS experts between
meetings to determine how these technologies should be treated. Possibilities are Corrigenda on
IPMP Components, Amendments to IPMP Components, a separate set of Technologies specific to
MS MAF, a common MAF set of IPMP technologies.
Professional Archival MAF
The Professional Archival MAF /audio was discussed in detail. Five tasks were identified:
1. MimeType Issues
2. MPEG-7Issues
3. Format of the Digital Item
4. Mapping to the ISO FF
5. Audio Specific Issues
The Professional Archival MAF/ Audio will be discussed further in an AHG between meetings. A
WD will support that discussion.
100
3.0 MDS Output Documents and Resolutions – Klagenfurt 77th Meeting
MPEG
No.
Title
General
Metadata Conversion – Problem and High Level Solution
8555
Statement
MPEG-7
No.
Title
15938-5 Multimedia Description Schemes
Request for Amendment 3 of ISO/IEC 15938-5 Improvements to
8556
Geographic Position Descriptor
ISO/IEC 15938-5/PDAM 3 Improvements to Geographic Position
8557
Descriptor
No.
TBP Available
N
06/10/27
TBP Available
N
06/10/27
N
06/10/27
Title
15938-7 Conformance
Request for Amendment 4 of ISO/IEC 15938-7 New Geographic
8558
Position Descriptor Conformance
ISO/IEC 15938-7/PDAM 4 New Geographic Position Descriptor
8559
Conformance
TBP Available
No.
TBP Available
Title
15938-10 Schema definition
DoC on ISO/IEC 15938-10:2005/DCOR 1 Multimedia content
8560
description interface — Part 10: Schema definition
ISO/IEC 15938-10:2005/COR 1 Multimedia content description
8561
interface — Part 10: Schema definition
N
06/10/27
N
06/10/27
N
06/10/27
N
06/10/27
1.1.3.
The MDS subgroup thanks National Body of Japan for its comment on
ISO/IEC 15938-10/DCOR1.
1.1.4.
The MDS subgroup nominates the following as editors of ISO/IEC
15938-10:2005/Cor.1: Robert O’Callaghan, Akio Yamada.
MPEG-21
No.
Title
21000 General
8562 Schema Files for MPEG-21 standards v.5
101
TBP Available
Y
06/10/27
1.1.5.
The MDS subgroup notes that the document N8562 is a new version of
an ongoing working document containing the ‘electronic’ versions of schemas for the
current MPEG-21 parts at IS/FDIS. The MDS subgroup requests that the versions of
the schemas be updated on the ITTF WWW site at the same URL as previous versions.
No.
Title
21000-4 IPMP Components
DoC for ISO/IEC 21000-4/PDAM 1: MPEG-21 IPMP Components
8563
Base Profile
ISO/IEC 21000-4/FPDAM 1: IPMP Components Base Profile
8564
TBP Available
N
06/10/27
N
06/10/27
1.1.6.
The MDS subgroup thanks the National Bodies of Italy, Korea and Spain
for their comments on the ISO/IEC 21000-4/PDAM 1 IPMP Components Base Profile.
1.1.7.
The MDS subgroup nominates Hendry and Takafumi Ueno as editors of
21000-4 AMD 1: IPMP Components Base Profile.
No.
Title
21000-5 Rights Expression Language
Request for Amendment 3 of ISO/IEC 21000-5 ORC Open Release
8565
Content
8566 ISO/IEC 21000-5/PDAM 3 ORC Open Release Content
TBP Available
N
06/10/27
N
06/10/27
1.1.8.
The MDS subgroup requests that Xin Wang, Chris Barlas, Jaime Delgado
be recorded as editors of ISO/IEC 21000-5 AMD 1; MAM Profile.
1.1.9.
The MDS subgroup nominates Jaime Delgado, Tae Hyun Kim, Chris
Barlas and Florian Schreiner as editors of ISO/IEC 21000-5 AMD 3: Open Release
Content.
No.
8567
Title
21000-6 Rights Data Dictionary
DoC on ISO/IEC 21000-6/DCOR 2 Rights Data Dictionary
8568 Text of ISO/IEC 21000-6/COR 2 Rights Data Dictionary
TBP Available
N
06/10/27
N
06/10/27
1.1.10.
The MDS subgroup thanks the National Bodies of Spain and the UK for
their comments on the ISO/IEC 21000-6/COR 2 Rights Data Dictionary.
102
No.
Title
21000-7 Digital Item Adaptation
8569 Disposition of Comments on ISO/IEC 21000-7/FPDAM 2
Text of ISO/IEC 21000-7/ FDAM 2 Dynamic and Distributed
8570
Adaptation
8571 MPEG-21 DIA Reference Software and Status Work Plan
TBP Available
N
N
06/10/27
07/01/14
N
06/10/27
1.1.11.
The MDS subgroup thanks the National Bodies of Austria and France for
their comments on the ISO/IEC 21000-7/FPDAM 2 Dynamic and Distributed
Adaptation.
No.
Title
21000-8 Reference Software
8572 Study of ISO/IEC CD 21000-8: Reference Software Second Edition
TBP Available
No.
TBP Available
Title
21000-14 Conformance
8573 DoC on ISO/IEC CD 21000-14: Conformance Testing
8574 ISO/IEC FCD 21000-14: Conformance Testing
N
N
06/10/27
06/10/27
06/10/27
1.1.12.
The MDS subgroup thanks the National Body of Japan for its comments
on the ISO/IEC CD 21000-14 Conformance Testing.
No.
8575
8576
8577
8578
8579
8580
Title
21000-18 Digital Item Streaming
DoC of ISO/IEC FCD 21000-18 Digital Item Streaming
Text of ISO/IEC 21000-18 Digital Item Streaming
TuC v5.0 for ISO/IEC 21000-18 Digital Item Streaming
Workplan for Core Experiment on DI Streaming Technologies
under Consideration
Request for Amendment 1 of ISO/IEC 21000-18 Digital Item
Streaming: Simple Fragmentation Rule
ISO/IEC 21000-18/PDAM/1 Digital Item Streaming
TBP Available
N
N
N
N
06/10/27
07/01/10
06/11/03
06/10/27
N
06/10/27
N
06/12/01
1.1.13.
The MDS subgroup kindly reminds National Bodies of the need for any
patent statements regarding ISO/IEC 21000-18 Digital Item Streaming to be lodged
with the ISO Central secretariat.
1.1.14.
The MDS subgroup thanks the National Bodies of Australia and Norway
for their comments on the ISO/IEC 21000-18/FCD Digital Item Streaming.
1.1.15.
The MDS subgroup nominates Gerrard Drury and Thomas Rørvik
103
Skjølberg ISO/IEC 21000-7/FPDAM 2 Dynamic and Distributed Adaptation.
MPEG-A
No.
Title
23000-2 MPEG-A Music Player
8581 DoC on ISO/IEC CD 23000-2 MPEG-A Music Player 2nd edition
8582 ISO/IEC FCD 23000-2 MPEG-A Music Player 2nd edition
Reference Software Workplan for MPEG-A Music Player 2nd
8583
edition
TBP Available
N
N
N
06/10/27
06/10/27
06/10/27
1.1.16.
The MDS subgroup thanks the National Bodies of Australia, Germany,
Japan and the US for their comments on the ISO/IEC CD 23000-2 MPEG-A Music
Player 2nd edition.
1.1.17.
The MDS subgroup nominates Harald Fuchs, Stefan Krägeloh, Schuyler
Quackenbush and Hendry as editors of ISO/IEC 23000-2 MPEG-A Music Player 2nd
edition.
No.
Title
23000-5 Media Streaming Player MAF
8584 ISO/IEC CD 23000-5 Media Streaming Player
8585 TuC for Media Streaming Player IPMP Technologies
Reference Software Workplan for ISO/IEC CD 23000-5 Media
8586
Streaming Player
TBP Available
No.
TBP Available
Title
23000-6 Professional Archival MAF
Request for Name Change of subdivision 23000-6 to Professional
8587
Archival MAF
8588 Professional Archival MAF Under Development Workplan
8599 WD of 23000-6 Professional Archival MAF - Audio
Y
N
N
06/10/27
06/10/27
06/10/27
N
06/10/27
N
N
06/10/27
06/10/27
Promotion
No.
8600
8601
8602
8603
8604
Title
21000 General
MPEG-21 Session Mobility One Pager
MPEG-21 Digital Item Processing Amendment 1 One Pager
MPEG-21 Conformance to Digital Item Processing One Pager
MPEG-21 Conformance to Digital Item Processing Amendment 1
One Pager
MPEG-21 Reference Software One Pager
104
TBP Available
Y
Y
Y
Y
06/10/27
06/10/27
06/10/27
06/10/27
Y
06/10/27
AHGs
i. AHG on MPEG-21 DIS
1. Carry out the Core Experiments on DI Streaming TuC and provide
recommendations to the MDS subgroup.
2. Finalise the text of ISO/IEC 21000-18.
3. Discuss and provide inputs on Reference S/W and Conformance for DIS.
4. Discuss usage of DIS in the context of streaming from dynamic sources
5. Discuss the relationship between DIS and the Media Streaming MAF
Gerrard Drury (gerrard*at*enikos.com)
Chair:
Peder Drege (peder.drege*at*adactus.no)
Duration: Until the 79th meeting.
Meetings: AHG meeting will be held on the weekend prior to 79th meeting. Other business will
be conducted by e-mail or telephone conference.
Reflector: mpeg21-uma_at_merl.com
Subscribe: To subscribe send email to avetro_at_merl.com (Anthony Vetro).
N8605
Mandate:
N8606
Mandate:
AHG on MDS MAFs Under Development
1. To conduct MAF under development workplans
2. Make recommendations to WG11 regarding the MAFs under
development standardisation
Chairman: Stefan Kraegeloh, Filippo Chiariglione and Noboru Harada
Duration: Until 79th Meeting
Meetings: AHG meeting will be held on the weekend prior to the 79th meeting. Other
work will be conducted by email/telephone conference
Reflector: mpeg-maf-dev@lists.uni-klu.ac.at
Subscribe: To subscribe follow the instructions on
http://lists.uni-klu.ac.at/mailman/listinfo/mpeg-maf-dev
105
4.0 MDS Final Schedule – Hangzhou 78th Meeting
v1.5
MPEG MDS Chair: Ian S Burnett
MPEG-7, MPEG-21, MAF
v1.4
Source
Title
Number
Monday Morning
(9h00-13h00)
MPEG Plenary
Plenary room
Monday Afternoon
(13h30-20h00)
Kick-off of MPEG
MDS activities
(13h30-14h00)
MDS Room Wenqi
Agenda, Goals and Issues for the Week for
MDS Group
Review of AHG
resolutions, CE results
and action points
(13h45-14h20)
Ian S Burnett
MDS Room Wenqi
13754
Gerrard Drury Peder Drege
13884
Thomas Skjølberg Peder Drege
Gerrard Drury Joseph Thomas-Kerr Report of CE on DIS TuC
13802
SC 29 Secretariat
13810
SC 29 Secretariat
13829
SC 29 Secretariat
13830
SC 29 Secretariat
13987
Christian Timmerer Michael Ransburg on Austrian NB comments on ISO/IEC 21000-7
behalf of the ANB
FPDAM
AHG on MPEG-21 DIS
Summary of Voting on ISO/IEC 210007:2004/FPDAM 2 [SC 29 N 7784]
Summary of Voting on ISO/IEC FCD 21000-18
[SC 29 N 7802]
Summary of Voting on ISO/IEC 210004:2006/PDAM 1
Summary of Voting on ISO/IEC CD 23000-2
[2nd Edition]
Define BoGs and
Mandates (14h2014h30)
MDS Room Wenqi
BoG1 = Zhuiyun, BoG2 = Tanyun
Virtual & Physical DoC production on FPDAM/2
BoG1 Time TBD
DIA
IPMP Components
DIS
IPMP Tuesday TBD BoG1
DoC Production PDAM/1 TuC CE BoG 2 Monday
4pm-, Tuesday 10am-
Music Player MAF
Virtual - Skype
Media Streaming MAF
CD TuC BoG1
REL
Tuesday 10am - 11am BoG1
MPEG-21 RDD (14h30
- 15h30)
13836
MDS Room Wenqi
Jaime Delgado Eva Rodriguez Marc
Gauvin
106
Final Comments on the Ontological Analysis of
the Study of DCOR/2 of ISO/IEC 21000-6
13837
ESNB position paper: Problems with the
inconsistency of the MPEG-21 Rights Data
Dictionary
Francisco Morán
MPEG-21 DIS (15h30
- 16h00)
13991
MDS Room Wenqi
Thomas Skjølberg Peder Drege
MAFs (16h00 17h30)
MDS Room Wenqi
13913
Harald Fuchs
13943
AHG on MAFs Under Development
13945
on MAFs Under Development
13946
on MAFs Under Development
Tuesday Morning
(9h00-13h00)
MPEG-21 Profiles
(9h00 - 9h30)
13892
Kisong Yoon(ETRI) Taehyun
Kim(DRM inside) Eva
Rodriguez(DMAG-UPC) Jaime
Delgado(DMAG-UPC) Hogab
Kang(DRM inside)
Tanya Beech
13876
13891
Sang-Kyun Kim Ryong Lee
13875
MAFs (11h00 - 12h00)
13915
13928
13759
13906
Proposed MPEG-21 REL Open Release Profile
Reqts
Soo-Jun Park Sung Min Kim Chee
Sun Won
Kyoungro Yoon Hee-Cheol Seo
Hyunki Kim Myung-Gil Jang
Hee-Cheol Seo Hyunki Kim MyungGil Jang Kyoungro Yoon
13870
Study Text on ISO/IEC CD 23000-2 MPEG-A
Music Player 2nd edition
Proposal of Updated Working Draft of ISO/IEC
23000-5 Media Streaming Player
Proposal of Updated Working Draft of IPMP
Extensions XML Messages
Proposal of Updated Working Draft of Media
Streaming MAF Technologies
Reqts
MPEG-7 (9h30 11h00)
13807
Delivery of dynamic resources in Digital Item
Streaming
Proposal for improvements to Geographic
Position in Mpeg7 Part 5
Proposal for a new MPEG-7 input query format:
Query-by-Layout
Comparison of MP7QF Requirements and TVAnytime Technology
Comparison of XQuery and MPEG-7 Query
Format
Request for adding Query Requirements related to
data manipulation against MPEG-7 DB on current
MPEG-7 Query Format Requirement
Reqts
Chun Hui Suen Florian Schreiner
Klaus Diepold
Xin Wang Chris Barlas
File Format and Event Reporting for Open
Release MAF
Rights Enforceability in the Open Release MAF
James A.G. Annesley James Orwell
Jim Aldridge Kate Grant
Eva Rodríguez Jaime Delgado
AHG on Surveillance MAF
107
IPMP and the Surveillance MAF
Tuesday Afternoon
(14h00-18h00)
Portable Video Player
MAF
Reqts
13760
H. Jean Cha Herbert Thoma
13995
H. Jean Cha Tae Hyeon Kim
13998
H. Jean Cha Tae Hyeon Kim
14017
H. Jean Cha
AHG on Portable Video Player MAF
Refined requirements and technologies for
Portable Video Player MAF
Proposed working draft of Portable Video Player
MAF
Proposed Work Plan for Portable Video Player
MAF
Digital Cinema MAF
13862
13863
Reqts
Mayumi Koike Takuyo Kogure
Hiroshi Yasuda
Mayumi Koike Takuyo Kogure
Hiroshi Yasuda
Digital Multimedia
Broadcasting MAF
Adaptation to MPEG MAF of Digital
Video/Cinema file format
Requirement of Color Management Information
to MPEG-7 for Digital Video/Cinema
Reqts
A T-DMB White Paper and a Introductory Movie
(6 minutes)
Request on MAF standardization for DMB
13858
13857
Korean National Body
Korean National Body
13859
Munchurl Kim Jeongyeon Lim Hui
Yong Kim Hyon-Gon Choo Yong Han
Requirements for DMB MAF
Kim Jinhan Kim Sung Ho Jin
Digital TV MAF
13889
Reqts
Hui Yong Kim Jeong Hyun Yoon Hee
Kyung Lee Han Kyu Lee Sung Ho Jin
Requirements for DTV MAF
Jae-Seok Jang Yong Man Ro
IPTV MAF
13929
Reqts
Xin Wang
Proposal for Working on an IPTV MAF
Medical Imaging MAF
13966
Reqts
Proposed Medical Imaging MAF (MI MAF) for
Preserving Medical Imaging Records
Wo Chang
File Formats (17h3018h00)
MDS, Systems, Reqts in Reqts
Wednesday Morning
(09h00-13h00)
MPEG Plenary (9h0011h00)
Plenary room
MPEG-21 DIA - BS
SCHEMA (11h30 12h00)
Joint with Video in MDS
108
13963
Davy De Schrijver Wesley De Neve
An MPEG-21 BS Schema for the scalable
Davy Van Deursen Saar De Zutter Rik extension of H.264/MPEG-4 AVC version 6
(Joint Scalable Video Model 6)
Van de Walle
Update on MS MAF
activity (12h0012h30)
MDS Room Wenqi
Wednesday Afternoon
(14h00-17h45)
Audio Archival MAF
(14h00 - 15h00)
Audio
13881
Noboru Harada Takehiro Moriya
Yutaka Kamamoto
13913
Harald Fuchs
MPEG-21 Ref s/w
Conformance and 1
pagersI (15h00 16h00)
Proposed text for WD of Audio Archival MAF
Study Text on ISO/IEC CD 23000-2 MPEG-A
Music Player 2nd edition
MDS Room Wenqi
13968
Chris Poppe Saar De Zutter Rik Van
de Walle
Saar De Zutter Frederik De
Keukelaere Gerrard Drury Christian
Timmerer Xin Wang
13988
Michael Eberhard Michael Sablatschan
Christian Timmerer
Editor’s input to ISO/IEC 21000-8 Reference
Software (Second Edition)
gBSDtoBin (MPEG-21 DIA) reference software
update
Saar De Zutter Sylvain Devillers
Thomas DeMartini Andrew
Tokmakoff
Saar De Zutter Davy De Schrijver Rik
Van de Walle
Saar De Zutter Chris Poppe Davy De
Schrijver Rik Van de Walle
Editor's input to ISO/IEC 21000-14
Conformance Testing
Update to Reference Software for Conformance
to ISO/IEC 21000-10
Update to Reference Software for Conformance
to ISO/IEC 21000-10/Amd 1
13965
13972
13978
13979
13980
13981
13982
13983
13984
13985
Saar De Zutter Rik Van de Walle
Saar De Zutter Davy De Schrijver Rik
Van de Walle
Saar De Zutter Chris Poppe Davy De
Schrijver Rik Van de Walle
Saar De Zutter Davy De Schrijver Rik
Van de Walle
Saar De Zutter Chris Poppe Davy De
Schrijver Rik Van de Walle
Saar De Zutter Davy De Schrijver Rik
Van de Walle
MPEG-21 REL & IPMP
(16h00 - 16h30)
Contribution to Utility Software for ISO/IEC
21000-10 DIP/AMD 1
Contribution to summary and 1-pager of
Enhanced Interoperability for MPEG-21 Session
Mobility using DIP
Update to summary of Digital Item
Technologies: Digital Item Processing
Contribution to summary and 1-pager of Digital
Item Technologies: Digital Item Processing Amd
1
Contribution to summary and 1-pager of
Conformance: MPEG-21 Digital Item Processing
Contribution to summary and 1-pager of
Conformance: MPEG-21 Digital Item Processing
Amd 1
Update to summary and 1-pager of Reference
Software: MPEG-21
MDS Room Wenqi
109
13949
Hyon-Gon Choo Filippo Chiariglione Proposed Working Draft of ISO/IEC 210004/Amd 2 Media Streaming Profile
Bum-Suk Choi
13956
Hendry Takafumi Ueno
13957
Hendry Munchurl Kim
Editor’s Study of ISO/IEC 21000-4/PDAM 1:
IPMP Base Profile
Contribution to ISO/IEC 21000-4/PDAM 1:
IPMP Base Profile Reference Software
MPEG-21 and LaSeR (16h30 - 17h30)
Systems
Thursday Morning (09h00-13h00)
RVC (09h00 - 10h00)
with Systems, ISG, Video in Video
Conversion between Metadata Systems (10h00 - 11h00)
with Systems
MS MAF Report Back (11h00 - 12h00)
MDS Room Wenqi
REL Report Back / Open Release MAF (12h00 -)
MDS Room Wenqi
Thursday Afternoon (14h00-19h00)
Audio Archival MAF (14h00-15h00)
Plenary MDS and Reports of BoG (15h00 - 17h00)
MPEG-21
MPEG-21
Further review of Output documents, AHGs, CEs, DoC,
Std (17h00+++)
MDS Room Wenqi
MDS Room Wenqi
MPEG-21
Issues
(15h00 16h00)
MDS Room Wenqi
MAF
Issues
(16h00 17h00)
MDS Room Wenqi
MDS Room Wenqi
Friday Morning (09h00-13h00)
Wrapping up (09h00 - 13h00)
MDS Room Wenqi
Approval
of
resolution
s, AHGs
and
Output
documents
Friday Afternoon (14h00-21h00)
MPEG Plenary
Plenary room
Contact: Ian S Burnett
x
110
Annex H – Video report
Source: Jens-Rainer Ohm, Gary Sullivan (Video), Miroslaw Bober (MPEG-7 Visual)
1. MPEG-1 and MPEG-2 Conformance
The video subgroup has analyzed the situation of conformance specifications related to video
standards prior to MPEG-4. In particular, it was found that ISO/IEC 11172-4 does not contain
any concrete specification of video conformance bitstreams, nor any such bitstreams at all.
Therefore, the video subgroup has issued a resolution to ask companies or individuals being in
possession of MPEG-1 video bitstreams appropriate for conformance testing, to donate them for
the purpose of an updated specification.
For MPEG-2, an initial investigation indicates that conformance test bitstreams for profiles other
than the Main Profile hardly exist. As a longer-term action, it might be considered to remove
profiles that are apparently unused.
2. New Colour Spaces
FDAM documents of the amendments related to inclusion of new colour space code points in
MPEG-2 Video (13818-2), MPEG-4 Visual (14496-2) and MPEG-4 AVC (14496-10) were
approved for release. All comments that were received during the ballot and by liaison
communication were accommodated. In particular, out-dated references to non-MPEG standards
were agreed to be updated.
Input documents reviewed
13762
SC 29 Secretariat
13763
SC 29 Secretariat
13766
SC 29 Secretariat
13780
ITU-R SG 6/WP 6J via SC 29
Secretariat
Summary of Voting on ISO/IEC 13818-2:2000/FPDAM 2
[SC 29 N 7736]
Summary of Voting on ISO/IEC 14496-2:2004/FPDAM 3
[SC 29 N 7737]
Summary of Voting on ISO/IEC 14496-10:2005/FPDAM 1
[SC 29 N 7740]
Liaison Statement from ITU-R SG 6/WP 6J [SC 29 N 7764]
Output documents
No.
Title
13818-2 Video
8445
Disposition of Comments on ISO/IEC 13818-2:2000/FPDAM 2
8446
Text of ISO/IEC 13818-2:2000/FDAM 2 Support for Colour Spaces
14496-2 Visual
8447
Disposition of Comments on ISO/IEC 14496-2:2004/FPDAM3
8448
Text of ISO/IEC 14496-2:2004/FDAM 3 Support for Colour Spaces
14496-10 Advanced Video Coding
8450
Disposition of Comments on ISO/IEC 14496-10:2005/FPDAM1
8451
Text of ISO/IEC 14496-10:2005/FDAM 1 Support for Colour
Spaces and Aspect Ratios
111
TBP Available
No
No
06/10/27
06/11/10
No
No
06/10/27
06/11/10
No
No
06/10/27
06/11/10
3. MPEG-7 Visual
a. MPEG-7 Visual related work in Hangzhou
The MPEG-7 breakout group was active during the whole week. Input documents related to the
Visual parts in 15938-3, 15938-6, 15938-7 and Photo Player MAF (23000-3) are listed in the
table below. All of these documents were reviewed and discussed.
13767
SC 29 Secretariat
13768
SC 37 via SC 29 Secretariat
13812
SC 29 Secretariat
13823
SC 29 Secretariat
13825
13856
SC 29 Secretariat
Hae Kwang Kim
Weon Geun Oh
Eun Ku Jung
Hae Kwang Kim
Sangki Kim
Sangyoun Lee
Myung Gil Jang
Jeong Hur
Weon Geun Oh
Hyeong yong Jeon
Jung Sub Shin
Chi Jung Hwang
Maeng Sub Cho
Ik-Hwan Cho
Seok-Kyoo Shin
Weon Geun Oh
Dong-Seok Jeong
Soo-Jun Park
Seon Hee Park
Soo-Jun Park
Seon Hee Park
13861
13867
13868
13869
13871
13872
13873
Ryoma Oami
13882
Weon Geun Oh
Donggyu Sim
isha1012@kw.ac.kr SueKyung
Park
Sang-Kyun Kim
Yong-Ju Jung
Yong Man Ro
Paul Brasnett
Miroslaw Bober
Paul Brasnett
Miroslaw Bober
Robert O'Callaghan
Miroslaw Bober
Akio Yamada
Wo Chang
Robert O'Callaghan
Miroslaw Bober
Sang-Kyun Kim
Akio Yamada
13887
13890
13936
13937
13950
13951
13952
Robert O'Callaghan
13953
Robert O'Callaghan
Summary of Voting on ISO/IEC 15938-7:2003/FPDAM 2
[SC 29 N 7741]
Liaison Statement from SC 37/WG 3 [SC 29 N 7742] (Face
Recognition)
Summary of Voting on ISO/IEC FCD 23000-3
Summary of Voting on ISO/IEC 159383:2002/Amd.1:2004/DCOR 2
Summary of Voting on ISO/IEC 15938-7:2003/PDAM 3
Survey on visual identifier technologies
An Image Data Management System for MPEG-7 VCE-6
CE Report for VCE-5
An Image Identifier Based on Singular Value
Decomposition and Feature Point
The Category and Complexity based Test Image Extraction
Method on MPEG-7 VCE-6
Report of Core Experiment: VCE-3 - Person-Identity-based
clustering, indexing and retrieval of images
Dataset for VCE-3 by ETRI, Version3
A proposal for a referencing mechanism of person
information for MPEG-A Photo Player
Request of Amendment in VCE-6 Specifications
Face detection
CE Report on Person-Identity based photo clustering and
indexing (VCE-3)
Experimental results on an image identifier (VCE-6)
Experimental dataset for VCE-6
Editors' input: FDIS 23000-3 (Photo-Player MAF)
Editors' input: TR 15938-8 DAM3 (Technologies for digital
photo management)
Defect Report: ISO/IEC 15938-3 Amd.2 (Perceptual 3D
Shape Descriptor)
UKNB comments on the text of ISO/IEC 15938-7 PDAM3 &
112
13954
13995
(on behalf of the UKNB)
Robert O'Callaghan
(on behalf of the UKNB)
H. Jean Cha
Tae Hyeon Kim
15938-6 PDAM2
UKNB comments on the text of ISO/IEC TR 15938-8
PDAM3
Refined requirements and technologies for Portable Video
Player MAF
Summary of key work items:
Part-3:
– Corrigendum work ISO/IEC 15938-3:2002/Amd.1/COR2
– Tools for Version 3 - Current Core Experiments:
– VCE-3 –Person-Identity-based clustering, indexing and retrieval of images
– VCE-5 -Evaluation of MPEG-7 Face Recognition Technology on IR Images
– VCE-6 -Visual Identifier
Part-6: Software amendment for Perceptual 3D Shape – FPDAM 2
Part-7: Conformance amendment for Perceptual 3D Shape – FPDAM 3
Photo Player MAF: ISO/IEC FCD 23000-3
– Addressing NB comments –all done
– Minor bug fixes
– External resources –legacy formats (e.g. paper photographs)
– Software amendment: PDAM 1
MPEG-7 Visual continues to run a series of CE’s related to Visual tools and DS for image or
photo-libraries with the key objective to develop new visual Description Schemes and other
algorithms for use with digital image libraries, such as personal collections of photos from digital
cameras. VCE-3 on person-identity-based clustering, indexing and retrieval of images will
continue aiming to improve ID-based clustering mechanism, and selection of the optimum usage
scenario. More experimental data are needed –in particular personal photo collections with many
faces. MPEG participants are also encouraged to help testing the developed technology on their
photo collections. VCE-5 on Evaluation of MPEG-7 Face Recognition Technology on IR Images
will continue with the objectives to compare performance of various algorithms, further extend
the Yonsei University database and use of other existing databases (e.g. Equinox, U. of NotreDame) to evaluate the applicability and performance of the Advanced Face Recognition
Descriptor on IR images and video. For VCE-6, the performance evaluation will continue.
Currently, 3 methods are under testing: Local Gradient Histogram, Local Gaussian Curvature via
Hessian matrix and Trace-transform. New stringent testing conditions with 1ppm false positive
rate on 10 billion images were defined as limit, and new image deformation types were added.
Based on the results of experiments, decision about a new extension of part 3 and possible
timeline will be made by the 79th meeting.
The Photo Player MAF specification was reviewed and all NB comments addressed. Work also
continued on the conformance testing and reference s/w.
b. Output documents related to MPEG-7 Visual
No.
8246
8247
8461
8462
Title
15938-3 Visual
Text of ISO/IEC 15938-3:2002/Amd.1/DCOR2
Description of Core Experiments for MPEG-7 New Visual Extensions
Disposition of Comments on ISO/IEC 15938-3:2002/Amd.1/DCOR2
Text of ISO/IEC 15938-3:2002/Amd.1/COR2
15938-6 Reference Software
113
TBP Available
No
No
No
No
06/07/21
06/07/21
06/10/27
06/10/27
8465
8466
8467
8468
Disposition of Comments on ISO/IEC 15938-6:2003/PDAM2
Text of ISO/IEC 15938-6:2003/FPDAM2 (Perceptual 3D Shape)
15938-7 Conformance testing
Disposition of Comments on ISO/IEC 15938-7:2003/PDAM3
Text of ISO/IEC 15938-7:2003/FPDAM3 (Perceptual 3D Shape)
No
No
06/10/27
06/10/27
No
No
06/10/27
06/10/27
c. Output documents related to MPEG-7 Part 8
No.
8469
8470
Title
15938-8 Extraction and Use of MPEG-7 Descriptions
Disposition of Comments on ISO/IEC TR 15938-8:2002/DAM3
Text of ISO/IEC TR 15938-8:2002/FDAM3 (Technologies for
digital photo management using MPEG-7 visual tools)
TBP Available
No
No
06/10/27
06/12/20
d. Output documents related to MPEG-A Photo Player MAF
No.
8471
8472
8473
8474
Title
23000-3 Photo Player Application Format
Disposition of comments on ISO/IEC FCD 23000-3
Text of ISO/IEC FDIS 23000-3
Request for ISO/IEC 23000-3/Amd.1: Reference Software for Photo
Player MAF
Working Draft 2 of ISO/IEC 23000-3/Amd.1
TBP Available
No
No
No
06/10/27
06/12/20
06/10/27
No
06/10/27
4. 23002 MPEG-C Video Technologies
e. 23002-1
A request for the first amendment to part 1 as well as the PDAM text were issued, containing a
software package suitable to perform the conformance tests as described in the standard. Draft
software was provided that appeared to be in generally very good condition as the basis for such
a reference software amendment.
No.
8477
8478
Title
23002-1 Accuracy specification for implementation of integer-output
IDCT
Request for ISO/IEC 23002-1/Amd.1 Software for Integer IDCT
Accuracy Testing
Text of ISO/IEC 23002-1/PDAM1 Software for Integer IDCT Accuracy
Testing
TBP Available
No
06/10/27
No
06/11/13
f. 23002-2 Fixed-point DCT/IDCT
i. Project status and overview of input contributions
The fixed-point IDCT/DCT project had shown very active ad-hoc group interest, with a large
amount of contribution and discussion. N8255 was the prior working draft, containing 5 non-final
candidate algorithms. N8256 had defined a workplan and metrics to be considered for the
114
selection of a proposal for creation of a CD at this meeting. N8257 contained a software testbed
for experiments and metric measurements.
Proponents had been asked to submit all final refinements of the 5 identified candidate algorithms
by 1 Sept 2006. The deadline was agreed by email to be extended to 8 Sept 2006 to accommodate
a personal emergency for one of the proponents. The proponents submitted the following in
response to the request for final refinements, with procedural aspects noted as follows:
– M13784 (Qualcomm/IBM/Zhejiang Univ.) contained no technical change relative to the
corresponding method in the WD (the submission is informative rather than a change of
proposal).
– M13791 (Connex) contained a small simplification relative to the corresponding method in
the WD.
– M13797 (CAS) contained a small refinement relative to the corresponding method in the
WD to make it more precise.
– M13799 (FastVDO) had been the subject of extensive discussion on the AHG email reflector
as to whether its content was consistent with the planned procedures for CD candidate
algorithms. It had been agreed by email that the technical evaluation conducted in the AHG
would use M13799 rather than the algorithm in the WD and the decision whether to consider
M13799 as a CD candidate or not would be left to MPEG to determine in Hangzhou.
– M13800 (Aveiro Univ.) contained a small simplification relative to the method in the WD,
and an accidental problem in a parameter value. M13803 was then submitted a week after
the agreed deadline, reportedly due to the discovery of an incorrect value of a parameter in
M13800. The algorithm and structure were not changed in M13803. No objection was
voiced to performing the subsequent technical evaluation using M13803 rather than M13800.
Three sets of evaluation results of the 5 submitted algorithms were provided: M13916, M13990,
and M13941.
The DCT/IDCT testbed software had been updated as reflected in N8257. N8257 also includes
software for ISO/IEC 23002-1 tests and invertibility and linearity tests. Subsets of this software
were also submitted as input contributions M13804, M13805, M13806, and M13999.
Additional information on proposed algorithms and proposed core experiments were submitted in
M13846, M13847, M13993, M14000, M14005, M14003, M13914, and M13997.
A cross-check of an algorithm in the WD (prior to its modification) was provided in M13992.
New information about dynamic range requirements for IDCT operation was provided in
M14004. New information about drift in IDCT operation was provided in M13912, M13927, and
M13934.
Information on existing IDCT designs was provided in M13996 and M14006.
The Chinese NB provided comments on precision, complexity, and proposal procedural issues in
M13930.
The AHG Recommended the following:
– To consider and resolve the procedural issues raised in AHG email and NB comments as
outlined above.
– To proceed with evaluation of candidate algorithms, algorithm selection, and creation of CD.
– To study new information contributions and proposed core experiment descriptions for postCD experiment development.
– To consider new information with potential impact on requirements.
115
–
To consider the (eventual) creation of formal ISO/IEC 23002-1 reference software (possibly
based on this testbed software and possibly including ISO/IEC 23002-2 reference software
on the same schedule), perhaps as ISO/IEC 23002-4. (See section e above.)
ii. Severe drift artifact studies
Drift artifact studies were reported in M13912 and M13934 with the following conditions and
results:
– Tested the 5 candidate IDCTs
– Testbeds: MPEG-2 & H.263+
– Test sequence: "MPEG-4 World News"
– Encoder using D-P floating point IDCT
– Found artifacts with QP=1, 2, 3
– Example obvious isolated artifacts were shown by the 57th frame with QP=1 for MPEG-2
with the M13971 (Connex) proposal; somewhat later with higher values of QP. Similar
behavior was reported for M13799 – in 50 frames there were reportedly already obvious
serious artifacts.
– Another example using H.263+ showed much more serious artifacts already by about frame
30 with M13971 (Connex).
A participant asked whether the encoder following the recommendation, stated in a nonnormative note in the MPEG-2 standard, to check for an all-zero reconstruction block (subclause
7.4.4 "Mismatch control" Note 2) ? This was called "the John Morris test" by one participant
(apparently after the person who reported the phenomenon to MPEG). During the meeting some
of these tests were reportedly performed again after including this recommended test, without a
major improvement in the outcome.
Such artifacts were found with the TM5 reference software's fixed-point IDCT approximation.
But with the M13784 method, such artifacts were not observed.
PSNR curves shown were for MPEG-2. For QP=1 and 2, M13791 (Connex) and M13799
(FaxtVDO) showed significantly more drift than most. M13797 (CAS) and M13784
(Zhejiang/IBM/QCOM) showed very little drift (roughly none) for QP=1, 2, 3. For QP=1 and 2,
the others were roughly grouped together. For QP=3 there was more separation, with M13803
(Aveiro) showing the best behaviour other than that of M13797 and M13794, next was
Broadcom's original proposal (not a CD candidate), and next TM5's fixed-point (also not a CD
candidate).
For peak-pixel-error (PPE), M13791 (Connex), BCOM, M13799, TM5, and M13803 all
reportedly showed large errors.
It was remarked that other test sequences should be tested and that the Akiyo and News
sequences in particular should perhaps not be used in MPEG in the future. Some suggested video
sequences included: Paris, Silent Voice, Irene, Deadline, Mother & Daughter, etc. Later in the
meeting it was remarked that some other test sequences seemed to show similar behaviour,
although perhaps not quite as bad in most cases.
The contributor of M13912 recommended not to adopt candidate algorithms that performed
poorly in such a test.
For testbed encoding algorithm, for QP=1, 2, 3:
– M13971 (Connex) and M13799 had severe artifacts.
– M13803 (Aveiro) better but still has some obvious artifacts.
– M13797 (CAS) and M13784 (Zhejiang/IBM/QCOM)
116
It was remarked that these values of QP are too small to represent typical practical use. However
there was a reply questioning the wisdom of hypothetically making a standard in which we would
add informative notes within the standard saying not to use it under some circumstances (e.g.,
with small values of QP).
It was noted that we had agreed that the target is an appropriate compromise trade-off of
complexity and precision.
Running the tests again with inclusion of the “John Morris test” reportedly didn't help much.
It was remarked that adding some noise prior to encoding seems to make the problem go away
(although requiring encoders to do such a thing does not seem like a satisfactory approach to the
issue).
It was remarked that this phenomenon seems to be the same basic phenomenon that is tested in
the linearity test and perhaps the DC test. It was noted that the methods that failed these tests
were also the methods that sometimes exhibited obvious drift artifacts on still areas of video
content.
iii. Study of “Anti-IDCT” behaviour
M13927 suggested to consider the behavior of an IDCT that is similarly distant from the ideal as
the IDCT under test, but "hostile" in its behavior relative to the IDCT under test.
This was modeled by having equal error of opposite sign as the IDCT under test. This should
behave the same way as negating all inputs, performing the IDCT, and negating the result, which
would be an actual implementation of the "anti-IDCT".
Software testbed testing was reportedly performed (for MPEG-2 and H.263) of such "anti-IDCT"
behavior. The CAS proposal was reported to have the best outcome, then Zhejiang, then Aveiro
in the middle, then (much trailing) FastVDO and Connex. For the News sequence there was
reportedly a 10 dB drift range by the 100th frame when using QP = 1.
M13934 reported completely identical cross-verification check of the results reported in M13927.
iv. Study of dynamic range requirements
M14004 reported on a topic that, it was pointed out in group discussion, had previously been
published as:
M. Zhou and J. De Lameillieure, "IDCT output range before clipping in MPEG video
coding", Signal Proc.: Image Communication, Vol. 11, No. 2, Dec 1997, pp. 137-145.
That paper also refered to a prior MPEG document M265 (July 1995) by E. Linzer.
Zhou reported a possible range of [-1805, 1805] in the spatial domain for MPEG-2 with TM5
quantization, and by email he reported [-1706, 1706] for H.263 with TMN 3 quantization.
It was illustrated that actual image prediction error content can excite this phenomenon when an
encoder is following ordinary encoding practices (e.g., TM5 quantization). It was shown that
many permutations of such prediction error content are possible.
117
An imperfect forward DCT or special encoding tricks such as dead-zone expansion and
individual non-zero coefficient removal or attenuation (some of which are well known) could
potentially aggravate the problem.
This suggests to consider supporting a dynamic range requirement of 12 bits, or at least ranges
beyond the +/-384 mentioned explicitly in current standards and for our current project metric
specifying a +/-512 test range.
The contribution suggested as a possibility to consider adding syntax to indicate a maximum
dynamic range requirement for decoding. It suggested some sort of “supplemental enhancement
information” indicator to be provided for an encoder to signal the potential dynamic range
requirement of a decoder.
The group conjectured as to how often such high-dynamic range cases might arise. One
suggestion was to require clipping (which would increase the total operation count requirements).
Another was to just add sufficient extra bits to cover +/-2048 dynamic range (two bits for those
that pass the 512 test and three bits for those that don't).
It was remarked that we never saw such a phenomenon in our tests. However it was noted that we
never tested with high values of QP. This dynamic range issue is a high-QP phenomenon, while
we had been assuming that low values of QP were more critical.
After discussion, the group agreed that we probably would not want to allow overflow to occur
when conforming to ISO/IEC 23002-2 for the anticipated +/-2048 range of values.
The plan was thus formed that for purposes of further discussion we would consider the
following approach to the issue:
– For the 16-bit methods, clipping is the only feasible of the two approaches (as extra dynamic
range would defeat their purpose).
– For the others, adding two bits of dynamic range seems the appropriate obvious approach (as
it avoids increasing the number of operations).
The precise impact analysis of the issue was left open for the moment.
v. Evaluations of candidate proposals with agreed metrics
M13916 reported that the dynamic range requirements for passing the +/-384 output range test
were as follows (the two 16-bit methods don't pass the 512 range test)
– M13784: 24
– M13791 / Connex: 16 (with muls)
– M13797: 29
– M13799: 16 / 17 (with muls)
– M13803: 27 [near DC error = 1, non-linear]
Reportedly, the two "16 bit" algorithms were not 16 bit for multiplierless operation (open to
further discussion). A supporting remark was that M13799 follows multiplication by a mid-range
offset, then a right shift – that addition requires an extra bit of dynamic range.
Three proposals (M13791 / Connex, M13799, and M13803) had an error of 1 on the near-DC
test; while M13784 and 13979 had zero error.
Similarly, M13784 and M13797 "pass" the Sarnoff linearity test; the others do not.
118
In terms of the number of required bit adds, it was reported that M13799 needed substantially
more than all others, then M13791 (Connex), then others. In terms of the number of required
shifts, the same basic characteristics were reported.
On video coding drift tests, M13791 (Connex) and M13799 reportedly performed relatively
poorly, then M13803 in middle, and the other two had quite good behavior but with M13797
being better than M13784.
M13990 contained results that were reportedly consistent with those reported in M13916.
For “16 bit” algorithms M13990 reported that M13799 (FV) had somewhat less drift than
M13791 (Connex). As an example, in some test result, the PPE was 1 for the two lowest-drift
methods, 3 for M13803 (Aveiro), and 21 or more for the two 16 bit methods.
M13990 (from Aveiro) reported that the M13803 (Aveiro) proposal had the lowest number of
total operations by a particular measure. The rough ratio of complexity to drift magnitude was
reported in M13990 (from Averio) to be the best for M13803 (Aveiro), then M13784, then
M13791 (Connex) and M13797, and finally M13799.
M13941 reported that overall objective precision metrics were generally better for M13797 than
M13784 (although both were very good relative to the others). Complexity metrics were reported
to be generally somewhat worse for M13797 than M13784, although reportedly not dramatically
so.
There was some discussion of some missing results (“Table 2”) in M13941 and some potential
minor inconsistencies, although the reported results seemed generally consistent with those in
other contributions. After discussion and further investigation, there seemed to be no significant
disagreements among the reported test results, although there was some expression that it would
have been desirable to have a more consistent style of result reporting to enable easier
comparison.
vi. Complexity consideration contributions
M13993 considered typical processor technologies (MMX, SSE, SSE II, XScale, Wireless MMX,
"DaVinci"). It contained a focus on parallel multiply-sum-shift operation support.
For multiplications, it was reported that sometimes a programmer can choose to store the upper
16 bits of a result or the bottom 16 bits of the result, but not some other subset of bits. For
multiply-adds, the operation availability seemed less constrained, but it was reported that one
must do multiply and add immediately and cannot do as many in parallel. Using a dot-product
with shift and round, one can reportedly do a sort of half-butterfly in one instruction: multiply,
add, then rounding offset and shift.
According to M13992, the proposal contained in M13799 (FV) is not friendly to the dot product
instruction. It would reportedly need to use parallel multiply, but that has an alignment problem,
so some key operations need to be broken into two components.
It was remarked that we cannot be sure whether this is true without going through a more
complete effort of full implementation of architecture-specific optimized implementation of
algorithms.
According to M13992, the proposal contained in M13791 (Connex) avoids a rounding stage with
a rounding cancellation trick that is friendly to such parallel architecture implementation. Also
119
M13791 (Connex) was reported to have another trick for performing adds prior to offset and shift
that makes it particularly "friendly".
M13992 generally suggested to examine particular architecture limitations when evaluating
candidate designs.
M13914 investigated hardware architectures for 4 candidate algorithms, as follows::
– M13784 (IBM) (prescaling, etc.)
– M13791 and M13799 (no pre-scaling, just left shift) – same basic arch
– M13797 (CAS) cascaded multiplication prescaling
It was questioned whether this study properly accounted for the full wordlength of intermediate
results in the 16 bit methods. It was agreed that this may not have been done.
For butterfly area, M13914 reports that M13784 was the smallest in Xilinx Vertex 4 FPGA,
M13784 & M13791 smallest in Synopsis, M13799 and M13797 were reportedly higher
complexity (which is worst reportedly depended on FPGA vs Synopsis).
For upscaling area, the M13797 upscaling part reported 7 times larger than M13784.
For total area, M13797 was reportedly worst (by a factor of 2.5), with the others roughly equal to
each other.
It was agreed that these were not fully optimized implementations.
It was remarked that the Xilinx platform has hardware multipliers, but this was implemented in a
multiplierless fashion. However, it was noted that M13997 reports similar results with use of
multipliers.
The contribution noted that pre-scaling can sometimes be combined with inverse quant and
pruning techniques.
The Aveiro design had not been included in the reported comparison.
M13997 focused on M13784 (IBM/QCOM/ZJU), M13791 (Connex), and M13799 (FV) based
on some typical computer architectures. The contribution considered cycle counts, latency, and
pruning.
M13997 asserted that M13784 takes far fewer cycles than either M13791 (Connex) or M13799
on basic measures of typical cycle counts and latency. However, the estimate does not account
for parallelization opportunities or impact of details of special instructions.
With pruning ("K" = 5 assumed), particularly on Pentium 4, M13997 asserted lower latency for
M13784, with the next slower being M13791 (Connex), and finally M13799 (FV) on such
measures.
M13997 reportedly used textbook computer architecture measures with "carry chain adder" and
"carry lookahead adder" and "school method multiplier" components with some optimistic (not
quite valid) assumptions in favor of the 16-bit schemes, and reported a much lower (3x) circuit
size and lower (with carry lookahead adder) or roughly equal (with charry chain adder) latency
for M13784 ZJU/IBM/QCOM than the 16-bit candidates (with multipliers assumed needed for
the M13799 FV proposal).
120
vii. Survey of industry implemented techniques
M13996 shows that some easily-found software decoder and encoder implementations (freeware
and otherwise) provide an option to use high-precision (particularly double-precision floating
point) IDCT when operating. Some implementations use this (or single-precision floating point)
all the time when running on PCs, due to the lack of significant complexity penalty for doing so.
One remark that arose in the discussion was whether there could be significant coding efficiency
differences for using different DCTs/IDCTs in encoders. After discussion the group concludd
that there probably would not be, and that we would assume not, unless some evidence is
provided otherwise.
M14006 lists seven places where the full details of some fixed-point IDCTs are publicly
available: H.263 Annex W (16 bit, but not MPEG-2 conforming), MPEG-2 TM5 software (32
bit), TI (16 bit Chen non-scaled for low-power devices), Motorola (16 bit Chen scaled), Intel IPP
(has 16 bit – MPEG-2 conforming?), XVID open source (various methods), Flask open source (9
selectable methods). Some of these may be targeted for H.263 or MPEG-4 Simple profile rather
than high-quality MPEG-2 implementation (note that H.263 has less stringent conformance
requirements).
viii. Core experiment suggestions
M13846 and M13847 reportedly suggested an emphasis on QP values 9-25 as the most
reasonable to consider, and noted that none of the proposals have major drift at those QP values.
It seems generally agreed that drift behavior (without overflow) is not prominent in such a QP
range.
M13846 reported that a number of adjustments can be made within a given design structure to
tune it for various purposes such as higher accuracy, lower complexity on various metrics,
Lifting-based variants, etc.
M13847 seemed to have been uploaded with the wrong document content – it contained no core
experiment suggestions (despite its reported title).
M14005 (cross-checked by M14003) discusses factorizations used in proposals of three
proponents, and provides remarks on common design methodology (e.g., pre-scaling with
subsequent butterflies). (It was remarked that the CAS contribution number was incorrect in this
contribution.) M14005 suggested experimenting with fine-tuning of accuracy, and fined-tuning of
constant factors for other reasons (e.g., storage in 8 bits, multiply-free computations,
minimization of number of shifts, bit depth constraints, LLM11 vs. LLM12 vs. AAN, 6 multiply
or 44 add implementation). There was some questioning as to whether it would be reasonable to
consider so many variations in CEs.
M14003 contains a cross-check of some of the experiment results in M14000. Used source code
that was used in the experiments and the testbed N8257, it was confirmed that all variants
described in M14000 met the criteria in ISO/IEC 23002-1. It was identified that four of the six
variants tested in M14000 did not pass the linearity test.
ix. Conclusions for fixed-point IDCT/DCT work
There was some discussion in which it was expressed that although some problems had arisen in
the interim period, there was no objection to considering all five of the current proposals as
candidates for technical evaluation and CD selection. The five proposals were thus agreed to be
considered as having equal status as candidates for technical evaluation purposes.
121
Based on the latest available information, there were two candidates (M13797 from CAS and
M13784 from Zhejiang/IBM/QCOM) that performed well on the following four basic criteria:
– linearity test
– near-DC test
– lack of serious visible artifacts under any identified conditions
– good statistical behavior on drift experimental tests
Thus it was agreed to base the CD on one of those two candidates.
As described above in section iv, we had agreed to add two bits of dynamic range requirement to
the selected method (M13797 or M13784) for the CD design.
The following considerations were then noted and agreed in the further discussion:
– Both M13784 an M13797 had excellent performance on experimental drift testing (each
generally exhibited no more than PPE = 1 for the vast majority of video sequence tests with
very small quantization step sizes and very long periods of drift accumulation, or perhaps
PPE = 2 in a couple of cases). Thus both seemed entirely acceptable in terms of measurable
drift. Although M13797 seemed somewhat statistically better by such measures, neither ever
exhibited a quantity of drift that seemed likely to be visible.
– In overall estimated implementation complexity, M13797 appeared to have significantly
high computational complexity requirements (perhaps roughly 30% higher) than M13784.
After extensive testing of algorithms contained in the previous working draft, consensus was thus
reached to adopt one single design for CD, based on the LLM11-factored M13784 algorithm with
two added bits of dynamic range requirement to prevent overflow. This selection establishes a
good compromise between complexity and accuracy, and was incorporated into the draft N8479
which was agreed to be progressed to CD.
The following core experiments were established for further study and possible improvements as
described in N8480:
– CE on reducing complexity of IDCT
– CE on support for extended dynamic range
A new version of the software testbed for fixed-point DCT/IDCT V 5.0 was issued as N8481,
including the various updates received for the meeting.
Documents reviewed:
13784
Yuriy Reznik
13791
Lazar Bivolarski
13797
13799
13800
13803
13804
13805
13806
13846
Honggang Qi
Wen Gao
Debin Zhao
Siwei Ma
Trac D. Tran
Navarro
Reznik
Silva
Navarro
Reznik
Silva
Arianne T. Hinds
Zhibo Ni
Yuriy Reznik
Trac D. Tran
updated IDCT algorithm for CD selection
Updated Connex Proposal of Low Complexity IDCT for
CD Selection
AAN IDCT Design for CD Selection
FastVDO IDCT proposal for CD
Improved IDCT
Improved IDCT- Replacing M13800
Updated MPEG-4 testbed
Updated MPEG-2 testbed
Updated H.263+ testbed
FastVDO 16-bit IDCT Proposal for CD: Performance and
122
13927
Lijie Liu
Pankaj Topiwala
Trac D. Tran
Lijie Liu
Pankaj Topiwala
Zhibo Ni
Lu Yu
Dandan Ding
Zhibo Ni
Lu Yu
Zhibo Ni
Cixun Zhang
Lu Yu
Lu Yu
13930
CNNB
13934
Honggang Qi
13941
Honggang Qi
13847
13912
13914
13916
13992
Antonio Navarro
Antonio Silva
Antonio Navarro
13993
Lazar Bivolarski
13996
Arianne T. Hinds
13997
Arianne T. Hinds
13999
Arianne T. Hinds
14000
Lazar Bivolarski
14003
Antonio Navarro
14004
Yuriy Reznik
14005
Yuriy Reznik
14006
Yuriy Reznik
13990
14018
14019
Jianguo Liu
Guoyou Wang
Shengkui Dai
Pingping Zhu
Xinjian Meng
Jianhua Zheng
Jianguo Liu
Guoyou Wang
Shengkui Dai
Pingping Zhu
Xinjian Meng
Jianhua Zheng
Comparison
Core Experiments for IDCT
Drift Problem of Fixed-Point IDCT on News Sequence
Analysis of Hardware Implementation Cost of Fixed-Point
IDCT
Test Results for Technical Selection of Committee Draft of
ISO/IEC 23002-2 Fixed-Point IDCT
Anti-IDCT for IDCT Drift Test
CNNB comments on the work of fixed-point 8x8 IDCT
transform
Crosscheck for proposal m13927
Test Results for Selection of Committee Draft of ISO/IEC
23002-2 Fixed-Point IDCT
Performance in MPEG-4 of five submitted integer IDCTs
for CD
Crosschecking an integer 16 bit IDCT (M13791)
On implementation of IDCTs on existing 16-bit
architectures
On the Usage of High Precision IDCTs in Existing MPEG
Products
On the Cost and Performance of IDCT Implementations in
Hardware
Updated T.83 testbed for IDCT testing
On the Complexity Analysis of IDCT Algorithms for CD
Selection
Cross check of proposed additional (CE-stage) IDCT
designs
On clipping and dynamic range of variables in IDCT
designs
Additional information on IDCT CD candidates and
proposed core experiments
Examples of existing fixed-point IDCTs
DSP implementations of 24-bit AAN algorithms
16-bit high precision scaled AAN for fixed-point IDCT
Output Documents:
No.
8479
8480
8481
Title
23002-2 Fixed point implementation of DCT/IDCT
ISO/IEC CD 23002-2 Fixed point IDCT and DCT
Description of Core Experiments on Fixed-Point DCT/IDCT
Software Testbed for fixed-point DCT/IDCT V 5.0
123
TBP Available
No
No
No
06/10/27
06/10/27
06/12/01
g. 23002-3 Representation of Auxiliary Video and Supplemental
Information
A Study of FCD was issued making various editorial improvements, and adding a suite of
conformance streams. A change of the title was requested by a resolution.
Documents reviewed:
13843
Arnaud Bourge
Proposed WD for ISO/IEC 23002-3 Conformance
Output document:
No.
Title
23002-3 Auxiliary Video Data Representation
8482
Study Text of ISO/IEC FCD 23002-3 Representation of Auxiliary
Video and Supplemental Information
TBP Available
No
06/11/13
h. 23001-4 and 23002-4 Reconfigurable Video Coding (RVC)
The purpose of the previous exploration on RVC is to provide a framework allowing a dynamic
development, implementation and adoption of standardized video coding solutions with features
of higher flexibility and reusability. At the 78th meeting, it was decided to
– 23001-4 Codec Configuration Representation, which does not contain any video-specific
elements
– 23002-4 Video Tool Library
23001-4 will contain the following specifications:
1) Description Language
– Specification of decoding rules for decoder representation
– Encoded video data bitstream syntax and rules for demultiplexing of the decoded
bitstream
– Connection
of
functional
units
(scheduling implicit when data-flow oriented language is used for describing FUs)
2) Abstract model
– describes behaviour of system in a way that conformance of the implementation can be
checked
– allows to generate a running model from description e.g. using simulation tools or C-code
generation
23002-4 is planed to consist of:
1) Collection of Functional Unit Descriptions (textual, normative)
– Based on formal description such that interfaces and internal behaviour are uniquely
specified
2) Definition of the formal description language for functional units
3) RSM implementation (not normative)
WD documents related to both of these parts were produced by splitting and updating the
previous WD 1 of RVC into the appropriate subsections, and updating contents by the proposals
in input documents 13908 and 14021. Contributions related to CEs were reviewed accordingly.
The most remarkable achievements were as follows:
– The “dedicated” description method that had evolved from the previous VCTR
exploration now uses a Decoding Hierarchy Table (DHT) to map the hierarchical
structure of bitstream syntax, and is better capable to control the scheduling (M13910);
124
–
The “generic” method based on CAL is mapped into an XML dialect which naturally
gives a hierarchical description, and could be further compressed using BiM tools for
better compactness (M13942).
These results indicate that a certain process of convergence between the two methods has been
achieved, where however it will be necessary to further investigate the advantages and
disadvantages to find the right combination of the two approaches. A major step forward was
achieved by the common understanding that an abstract dataflow model is a prerequisite for
successful generic description and unique and reproducible derivation of any implementation /
running device. A first “proof of concept” for such an abstract model will be the applicability for
existing solutions of today (in particular AVC) as expected from the updated work plan and
ongoing CEs.
An issue discussed in a joint meeting with the Systems and MDS subgroups, which also needs
further study in CEs is the relationship with BSDL. From the current BSDL approach, probably it
is possible to describe the syntax of table-based VLC codes. For arithmetic decoding in particular,
definition of dedicated Functional Units would most probably be needed. The link between the
code syntax and semantics (i.e. code de-multiplexing and decoder operations to be performed) is
currently unresolved in BSDL. Furthermore, the question remains open whether it would be
possible to define a generic parsing unit, or whether parsing would at least partially need to be
performed by defining dedicated functional units. These issues will further be investigated in the
core experiments.
Documents reviewed:
13907
13908
13909
13910
13919
13942
Euee S. Jang
Sunyoung Lee
Alex Chungku Yie
Eunkyung Kwak
James S.G. Yoo
Rana Lee
Sunyoung Lee
Hyungyu Kim
Hyunsoo Ahn
Sinwook Lee
Jaebum Jun
Giseok Son
Chungku Yie
Euee S. Jang
Hyungyu Kim
Sunyoung Lee
Hyunsoo Ahn
Sinwook Lee
Jaebum Jun
Giseok Son
Chungku Yie
Euee S. Jang
Hyunsoo Ahn
Sunyoung Lee
Hyungyu Kim
Sinwook Lee
Jaebum Jun
Giseok Son
Chungku Yie
Euee S. Jang
Sung-Wen Wang
Chung-Yi Weng
Wei-Kai Steve Su
Marco Mattavelli
Reshaping Digital Media Business Models by
Reconfigurable Video Coding
Proposed Updates of RVC Working Draft 1.0
RVC CE1 : RVC based Inter Coding Implementaion
Proposal on scheduling over RVC framework
RVC CE2: Extensibility of FUs and Interfaces between
CAL and C++
Report on results of RVC CE 1.2 Formalize XML-based
125
13944
13947
13948
14021
Jorn Janneck
Dave Parlour
Marco Mattavelli
Joseph Thomas-Kerr
Jorn Janneck
Dave Parlour
Marco Mattavelli
Andrew Kinane
Christophe Lucarz
Jorn Janneck
Dave Parlour
Marco Mattavelli
Christophe Lucarz
Andrew Kinane
Marco Mattavelli
Jorn Janneck
description of configuration of FUs.
Report on results of RVC CE 1.1 Implement flexible FUs
according to the processing mechanism in RVC WD using
CAL.
Report on results of RVC CE 2.1 Reshape the current
MPEG-4 SP CAL decoder according to the current FU
interface in RVC WM.
Report on results of RVC CE 2.2 Explore the extensibility
of FUs
Proposition for update of the RVC WD
Output Documents:
No.
8475
8476
8483
8484
8485
8486
8487
8488
Title
23001-4 Codec Description Representation
Request for Subdivision: ISO/IEC 23001-4 Codec Description
Representation
WD 2 of ISO/IEC 23001-4
23002-4 Video Tool Library
Request for Subdivision: ISO/IEC 23002-4 Video Tool Library
WD 2 of ISO/IEC 23002-4
White Paper on Reconfigurable Video Coding (RVC)
Description of Core Experiments in RVC
RVC Simulation Model (RSM) V2.0
RVC Work Plan
TBP Available
No
06/10/27
No
06/10/27
No
No
Yes
No
No
No
06/10/27
06/11/06
06/11/06
06/11/06
06/11/06
06/10/27
5. JVT Report
The Joint Video Team (JVT) of ITU-T Q.6/16 and ISO/IEC JTC 1/SC 29/WG 11 held its 21st
meeting during October 20-27, 2006 in Hangzhou, China. The JVT meeting was held under the
chairmanship of Dr. Gary Sullivan (Microsoft/USA) and Dr. Jens-Rainer Ohm (RWTH
Aachen/Germany), and under the associate chairmanship of Dr. Thomas Wiegand (Fraunhofer
HHI/Germany). The other JVT associate chairman, Dr. Ajay Luthra (Motorola/USA), was unable
to attend this meeting. The JVT meetings opened at approximately 2:30 p.m. on Friday October
20, 2006 and closed at approximately 1:40 p.m. on Friday October 27, 2006. Approximately 195
people attended the JVT meetings (as recorded on a sign-in sheet passed at the meeting) and
approximately 160 input documents were discussed. The meetings took place in a co-located
fashion with a meeting of ISO/IEC JTC 1/SC 29/WG 11. The subject matter of these activities
consisted of work on video coding.
i. Documents of the JVT meeting
i. Input documents
1. Administrative input contributions
JVT-U000 List of documents of Hangzhou meeting
126
JVT-U001 [G.J. Sullivan, J.-R. Ohm, A. Luthra, T. Wiegand] AHG Report: Proj mgmt and
errata
JVT-U002 [T. Wiegand, K. Suehring, A. Tourapis, K.P. Lim] AHG Report: JM text and ref soft
JVT-U003* [T. Suzuki] AHG Report: Bitstreams & conformance
JVT-U004* [J. Vieron, M. Wien, H. Schwarz, L. Bivolarski] AHG Report: JSVM s/W and new
func. integ.
JVT-U005* [J. Reichel, H. Schwarz, M. Wien] AHG Report: JSVM & JD text
JVT-U006* [S. Sun, A. Segall, J. Reichel] AHG Report: Spatial scalability resampling
JVT-U007* [Y.-K. Wang, S. Pateux, P. Amon, T. Schierl] AHG Report: High-level syntax, err
resil
JVT-U008* [M. Wien, H. Schwarz] AHG Report: Coding eff & JSVM perf test cond
JVT-U009* [T. Suzuki] AHG Report: Study of 4:4:4 functionality
JVT-U010* [J. Vieron] AHG Report: SVC interlaced coding
JVT-U011* [J. Ridge, D. Marpe, G. Sullivan] AHG Report: SVC quantization, CAVLC,
CABAC
JVT-U012* [M. Mathew, J. Li, H. Schwarz] AHG Report: Bitstream extractor
JVT-U013* [H. Schwarz, Y. Bao] AHG Report: Complexity reduction
JVT-U014* [S. Kamp, X. Wang] AHG Report: AR-PR and PR slices
JVT-U015* [A. Vetro, Y. Su] AHG Report: MVC H-L syntax & buffer mgmt
JVT-U016* [H. Kimata, A. Smolic, Y. Su, A. Vetro] AHG Report: JMVM text editing
JVT-U017* [P. Pandit, A. Vetro] AHG Report: JMVM soft & new func integ
2. Input liaison statements
JVT-U018* [SMPTE] LS: Constraints on High 10 profile (WG 11 input document M13841)
JVT-U019* [SMPTE] LS: New profile for production (WG 11 input document 13842)
3. Non-administrative input contributions
JVT-U020* [J. He , Y. Yan, Y. Prieto] Disabling SVC chroma deblocking
JVT-U021-L [W. Yao, Z. G. Li, S. Rahardja] Balanced inter-layer prediction
JVT-U022* [H. Yu, G. Sullivan] Proposed 4:4:4 draft changes
JVT-U023* [D.T. Nguyen, J. Ostermann] Error concealment in the NAL
JVT-U024* [Y. Yan, J. He, Y. Prieto] On CE4: Dyadic spatial resampling
JVT-U025* [E.Francois, V.Bottreau, J.Vieron] Modified inter-layer prediction for ESS
JVT-U026* [P. Pandit, Y. Su, P. Yin] Comments on High-level Syntax for MVC
JVT-U027* [D. Sim, S.N. Park] CE11: MB-based illumination comp.
JVT-U028* [D. Sim, S.N. Park] CE11 Sejong/ETRI's illum. comp. JVT-XXXX
JVT-U029-M [A. Leontaris, A.M. Tourapis, K. Suehring] ME & MC Enhancements to JM ref
soft
JVT-U030-L [A.M. Tourapis, K. Suehring, G.J. Sullivan, A. Leontaris] Revision of JM ref
software manual
JVT-U031-L [J.-H. Yang] CE11: Illum. comp. consistent pred.
JVT-U032-L [Z. Lu, J. Zheng, W. Lin, S. Rahardja] Percept. Deblock Filter for ROI SVC
JVT-U033* [K. Shimauchi] Inter-layer estimation for SVC
JVT-U034-L [B.-K. Lee] CE3: Improved context modeling PR slices
JVT-U035* [S. Wittmann, T. Wedi] Post-filter hint SEI
JVT-U036* [P. Onno, F. Le Leannec, X. Hinocq, J. Takeda] Quality layer SEI for virtual
resolutions
JVT-U037* [F. Le Leannec, P. Onno, X. Henocq, J. Takeda] CE2: Switching PR slices
JVT-U038* [F. Le Leannec, P. Onno, X. Henocq] CE2: Cross-verif ETRI/Sejong JVT-U050
JVT-U039-L [F. Le Leannec, P. Onno, X. Henocq] CE6: Cross-verif Thomson's JVT-U025
JVT-U040-L [S.-H. Lee, S.-H. Lee, N.-I. Cho, J.-H. Yang] MVC: Disparity vector prediction
JVT-U041* [A. Segall, L. Kerofsky and S. Lei] Tone Mapping SEI Message: New results
127
JVT-U042* [A. Segall, J. Zhao] CE4: Texture Upsampling with 4-tap Cubic Spline
JVT-U043* [A. Segall] CE8: SVC-to-AVC bitstream rewriting for CGS
JVT-U044-L [A. Segall] Transcoding in Scalability Info SEI
JVT-U045 [withdrawn] withdrawn
JVT-U046-L [W.S. Shim, H.S. Song, Y.H. Mun, J.B. Choi] High-level syntax for flexible I
frame position
JVT-U047* [H. Yan, J. Huo, Y. Chang, S. Lin, P. Zeng, L. Xiong] Regional Disparity Est/Comp
for MVC
JVT-U048* [S. Lin, P. Zeng, J. Zhou, Q. Xie, C. Hu, L. Xiong] MVC high level syntax: Camera
Parameters
JVT-U049* [Y. Gao, Y. Wu] Apps & Reqs for color bit depth SVC
JVT-U050* [J. Jia, H.K. Kim, H.C. Choi, J.G. Kim] CE2: Tool 1 SP Picture for SVC Switching
JVT-U051* [J. Jia, H.K. Kim, H.C. Choi, J.G. Kim] CE2: Tool 2 Verif Canon JVT-U037 Sw PR
JVT-U052* [Y.-L. Lee, J.-H. Hur, S.H. Cho, N.H. Hur, J.W. Kim, Y. Su, P. Yin, C. Gomila, J.H. Kim, P.-L. Lai, A. Ortega] CE11: Illumination compensation
JVT-U053* [Y.-L. Lee, J.-H. Hur, S.H. Cho, N.H. Hur, J.W. Kim] CE11 Kwangwoon
University Illum Comp
JVT-U054* [Y.H. Tan] CE3: Modif CABAC for MC for M-R FGS
JVT-U055* [S. Kamp, M. Wien] CE5: Results for JVT-U062
JVT-U056 [withdrawn] withdrawn
JVT-U057* [S. Rane, P. Baccichet, B. Girod] CE9: On error prot redundant slices
JVT-U058* [Q. Chen, Z. Chen] Modif scene info SEI message
JVT-U059* [Z. Chen, Q. Chen, X.D. Gu] SEI for functional app
JVT-U060* [H. Nakamura, M. Ueda] MVC H-L syntax parallel proc
JVT-U061* [A. Vetro, S. Yea, P. Pandit, Y. Su] MVC ref software implementation plan
JVT-U062* [A. Vetro, S. Yea] On MVC DPB management
JVT-U063* [S. Yea, A. Vetro] CE10: View synthesis prediction
JVT-U064* [V. Bottreau] CE4: Verif Sharp inter-layer JVT-U042
JVT-U065* [S. Sun, V. Bottreau] CE4: Texture upsampling results
JVT-U066* [P. Symes, H. Yu] Simple Intra profile for prof apps
JVT-U067* [G.J. Sullivan] Position Calc for SVC Upsampling
JVT-U068* [K. Ugur, J. Lainema, M.M. Hannuksela, H. Liu] On parallel encoding/decoding of
MVC
JVT-U069* [K. Ugur, J. Lainema] On common conditions for MVC
JVT-U070* [I. Amonou, N. Cammas, S. Kervadec, S. Pateux] On SVC performance and profiles
JVT-U071* [K. Sohn, Y. Kim, J. Seo, J. Yoon, J. Kim] Encoder optimization of MVC
JVT-U072* [K. Sohn, Y. Kim, J. Seo, J. Yoon, J. Kim] CE11: Verif LG/SNU JVT-U031-L
Illum comp
JVT-U073* [G. Park, S. Jeong, M. Park, D. Suh, K. Kim, K. Moon, J. Hong] CE5: Tool1 results
JVT-U021-L
JVT-U074* [S. Jeong, M. Park, G. Park, K. Kim, D. Suh] CE5: Verif Aachen JVT-U055 & T1
vs T2
JVT-U075* [D.Y. Suh, G.H. Park, J. Oh, M. Park] CE9: JVT-S028 extension redundant pic
(withdrawn)
JVT-U076-L [X. Ji] CE5: Improv FGS for low-delay
JVT-U077-L [X. Ji] Block based FGS for low-delay
JVT-U078-L [L. Zhang, X. Ji, D. Zhao, W. Gao] Adapt. spatial & transform domain FGS
JVT-U079* [K.B. Kim, M.-C. Hong] Search range for fast ME
JVT-U080* [B. Lee, J. Lim, M. Kim, S. Hahm, B. Kim, K. Lee, K. Park] SVC NAL unit types
for online extraction
JVT-U081* [J. Lim, P. Chen, B. Lee, M. Kim, S. Hahm, B. Kim, K. Lee, K. Park] Optimal SVC
bitstream extraction
JVT-U082-L [H. Kirchhoffer, D. Marpe, T. Wiegand] CE3: Improved CABAC for PR slices
128
JVT-U083-M [H. Kirchhoffer, D. Marpe, T. Wiegand] CE3: Verif JVT-U034-L PR context
model
JVT-U084-L [D. Marpe, G. Marten, T. Wiegand] Fast CABAC renorm for H.264/MPEG4-AVC
JVT-U085* [A. Eleftheriadis] Clarif Nesting Temporal Levels
JVT-U086-M [A. Eleftheriadis] Prop SVC profile for videoconf
JVT-U087-M [B.K. Lee] CE3: Verif JVT-U082-L PR slice CABAC
JVT-U088-L [W.-J.Han, B.-K.Lee] CE5: Verification of ETRI JVT-U073
JVT-U089 [withdrawn] withdrawn
JVT-U090-L [S.-W. Park, B.-Y. Jeon] Usage of store_base_rep_flag
JVT-U091-L [H.-S. Koo, Y.-J. Jeon, B.-Y. Jeon] MVC motion from neighbor view
JVT-U092 [withdrawn] withdrawn
JVT-U093* [H. Kimata, S. Shimizu, M. Tanimoto, T. Fujii] CE10: MVC view interpolation pred
JVT-U094-M [S. Jeong, K. Moon, J. Hong] CE5: Verif Tool 3 JVT-U076-L L-D FGS
JVT-U095-L [J. Xu] CE4: Improv inter-layer pred
JVT-U096-L [J. Xu] CE5: Verif JVT-U077-L and JVT-U076-L
JVT-U097-L [E. Francois] CE6: Verif Nokia JVT-U130 ESS
JVT-U098* [V. Bottreau] SVC MB layer for EI slices
JVT-U099-L [S. Sekiguchi, Y. Yamada, K. Asai] Advanced 4:4:4 profiles
JVT-U100* [Y. Ho, K. Oh, C. Lee, P. Park, B. Choi] Global Disparity Comp for MVC
JVT-U101* [Y. Ho, K. Oh, C. Lee, P. Park] Reference Frame for MVC
JVT-U102* [Y. Ho, C. Lee, S. Yoon, K. Oh, B. Choi] View Interpolation for MVC
JVT-U103-L [Y.-K. Wang, Y. Chen, M. M. Hannuksela] Comments to JMVM 1.0
JVT-U104-L [Y.-K. Wang, Y. Chen, M. M. Hannuksela] Time-first coding for MVC
JVT-U105-L [Y. Chen, Y.-K. Wang, M. M. Hannuksela] MVC reference picture management
JVT-U106-L [Y. Guo, Y.-K. Wang, M. M. Hannuksela, H. Li] Discardable data adaptation
JVT-U107-L [Q. Shen, Y.-K. Wang, H. Li] Adaptive inter-layer prediction
JVT-U108 [Q. Shen, Y.-K. Wang, M. M. Hannuksela, H. Li] Ref pic marking for temporal SVC
(withdrawn)
JVT-U109-L [Y.-K. Wang, M. M. Hannuksela] On SVC high-level syntax
JVT-U110* [M. M. Hannuksela, Y.-K. Wang] AVC SEI semantics in SVC context
JVT-U111-L [Y.-K. Wang, M. M. Hannuksela] SVC HRD
JVT-U112-M [Y. Chen, Y-. K. Wang, M. M. Hannuksela] SVC ref pic list construction
JVT-U113-M [Y. Guo, Y.-K. Wang, H. Li] CE9: Verif JVT-U057 redund slices
JVT-U114-M [C. Zhu, Y.-K. Wang, H. Li] Adaptive redundant picture coding
JVT-U115* [T.C. Thang, T.M. Bae, Y.M. Ro, J.W. Kang, J.-G. Kim] AR-FGS with motion
refinement
JVT-U116* [A. Eleftheriadis, S. Cipolli, J. Lennox] Err resil frame nums in key pics
JVT-U117-L [H. Schwarz] CE8: Verif JVT-U043 SVC-to-AVC
JVT-U118-L [J. Jia, H.K. Kim, H.C. Choi, J.G. Kim] Terms for SVC access unit def
JVT-U119* [Y. Bandoh, S. Takamura, K. Kamikura, Y. Yashima] Sep luma/chroma comp. in
SVC
JVT-U120* [T. Wedi, H. Ohtaka, J. Wus, S. Sekiguchi] Intra-only profile for prof apps
JVT-U121* [V. Bottreau] CE4: Verif LG JVT-U089 interl
JVT-U122-L [S. Sun] Verif QCOM JVT-U126 smooth ref
JVT-U123* [S. Regunathan, S. Srinivasan, C. Tu, S. Sun, G. Sullivan] Flexible 4-tap spat SVC
upsamp
JVT-U124-L [S. Kamp, M. Wien] Low-delay leaky base layer
JVT-U125* [Y. Bao, Y. Ye, M. Karczewicz, P. Sagetong] CE1: Results PR slice improve
JVT-U126* [Y. Ye, Y. Bao] CE4: L-C smooth ref spat SVC
JVT-U127-L [J. Ridge] Mobile profile for SVC
JVT-U128-L [J. Ridge, X. Wang] CE1: Improve FGS VLC
JVT-U129-L [J. Ridge, X. Wang] Component separation FGS
129
JVT-U130* [X. Wang, J. Ridge] CE6: ESS Inter-layer pred
JVT-U131-M [X. Wang] Verif RWTH-Aachen JVT-U055
JVT-U132* [M. Karczewicz, R. Panchal] Refinement coef coding
JVT-U133-M [S.-T. Hsiang] Intra subband/wavelet framework
JVT-U134-L [H. Kimata, S. Shimizu] On direct mode for MVC anchors
JVT-U135-M [S.-T. Hsiang] CE1: Verif Nokia JVT-U128-L
JVT-U136-L [S. Sekiguchi] Prop changes to 4:4:4 draft
JVT-U137* [B. Haskell] Simple SVC profile
4. Late-registered input contributions
JVT-U138-L [T. Senoh, T. Aoki, H. Yasuda, T. Kogure] CE10: Inter-camera prediction
JVT-U139-M [P. Amon, T. Rathgen, D. Singer] SVC file format
JVT-U140-M [M. Wien, R. Cazoulat, A. Graffunder, A. Hutter, P. Amon] R-T SVC streaming
syst
JVT-U141-M [M. Wien, H. Schwarz, T. Oelbaum] SVC performance analysis
JVT-U142 [T. Suzuki] Prop DCOR AVC/FRExt conformance (withdrawn)
JVT-U143-M [T. Suzuki] Level definitions for prof apps
JVT-U144-L [I. Amonou, N. Cammas, S. Kervadec, S. Pateux] R-D extract quality layers SVC
JVT-U145-L [H. Schwarz, D. Marpe, T. Wiegand] SVC overview
JVT-U146-L [E. Francois, J. Vieron, V. Bottreau] Interlaced coding in SVC
JVT-U147-L [T. Tran, L. Liu, P. Topiwala] Down/up-sampling filter for SVC
JVT-U148 [T. Tran, L. Liu, P. Topiwala] Filtering for ESS (withdrawn)
JVT-U149-L [M. Mathew, B.K. Lee] CE4: Verif JVT-U095-L
JVT-U150-L [J. Xu] 3D wavelet SVC coding scheme
JVT-U151-M [Y.-K. Wang, M.M. Hannuksela, S. Pateux, A. Eleftheriadis] SVC System &
Transport Interface
JVT-U152-M [S. Wenger, Y.-K. Wang, T. Schierl] SVC in IP networks
JVT-U153-L [X. Ji] CE1: Verif JVT-U125 PR slice
JVT-U154* [ITU-R SG6/WP 6J] LS: Colour spaces
JVT-U155-M [Y. Bao] CE4: Verif JVT-U123 upsampling
JVT-U156-L [S. Sun, G. Sullivan] Scalable Coding Solutions Based on Various Sub Sequence
Structures
JVT-U157* [ITU-T SG 9] LS: On MVC
JVT-U158-M [P. Topiwala] Requirements for HD/SD SVC
JVT-U159-M [L. Cieplinski] Verif JVT-U132 coef coding
JVT-U160-M [A. Eleftheriadis] On telescopic mode decision
JVT-U161-M [J. Ridge] Verif JVT-U147 resampling
ii. Major output documents
(Dates listed are planned dates of availability.)
JVT-U200 Meeting report of the 21st JVT meeting [06/11/20] (included in WG 11 parent
body report)
JVT-U201 Joint Draft 8: Scalable Video Coding [06/11/10] (WG 11 N 8455)
JVT-U202 Joint Scalable Video Model (JSVM) 8 [06/12/08] (WG 11 N 8456)
JVT-U203 JSVM 8 Software [07/01/05] (WG 11 N 8457)
JVT-U204 Joint Draft 5: 4:4:4 coding [06/11/14] (WG 11 N 8452)
130
JVT-U205 Joint 4:4:4 Video Model (JFVM) 5 [06/11/14] (WG 11 N 8453)
JVT-U206 JFVM 5 Software [06/11/14] (WG 11 N 8454)
JVT-U207 Joint Multi-view Video Model (JMVM) 2 [06/11/10] (WG 11 N 8459)
JVT-U208 JMVM 2 Software [06/11/17] (WG 11 N 8460)
JVT-U209 Joint Draft 1: Multiview Video Coding [06/11/10] (WG 11 N 8458)
JVT-U210 ITU-T Rec. H.264 | ISO/IEC 14496-10 Advanced Video Coding Defect Report
[07/01/10] (WG 11 N 8449)
iii. JVT internal output documents
JVT-U211 Common conditions for MVC [06/10/27]
iv. SVC core experiment output documents
Submission of final description: next meeting start - 3 weeks
Submission of final software and results: next meeting start - 2 weeks
JVT-U301: CE 1: Refinement coding to find whether a) something is broken and the
adaptation should be removed. b) whether macroblock-adaptive signaling should be used.
Based on JVT-U132. [M. Karczewicz, Qualcomm, Nokia, HHI, TI, Mitsubishi]
JVT-U302: CE 2: Switching (SP pictures): Based on JVT-U050. [J. Jia, Nokia, HHI, BT,
MS, Qualcomm]
JVT-U303: CE 3: Resampling: Based on JVT-U024*, JVT-U123, JVT-U147, JVT-R070. [S.
Sun, MS, HHI, RWTH, Siemens, Nokia, Sharp, Freescale, Qualcomm, Motorola, Ericsson,
Mitsubishi]
JVT-U304: CE 4: Enhancement layer complexity. Based on H.241 RCDO, JVT-U020, JVTU123. [S. Sun, MS, HHI, Nokia, Freescale, Huawei, Sharp, Ericsson, Mitsubishi,
Qualcomm, Motorola, RWTH]
Study complexity aspects of SVC, including deblocking and MC. CE participants will implement
H.241 RCDO and use this as an additional anchor. Compare the various combinations. Make
results and software available by Dec 10, 2006.
JVT-U305: CE 5: Subband techniques. Based on JVT-U095 and JVT-U133. [J. Xu, MS,
HHI, Nokia, Freescale, Motorola, Qualcomm, Sharp, Huawei, ICTCAS, Ericsson, Orange]
JVT-U306: CE 6: AR PR slices. Based on JVT-U076, JVT-U077. [X. Ji, ICTCAS, HHI,
RWTH, KHU, Nokia, MS, Motorola, Qualcomm, ICU]
131
JVT-U307: CE 7: Inter-layer prediction. Based on JVT-U107. [Y.-K. Wang, Nokia, HHI,
Sharp, RWTH, I2R, Motorola, Mitsubishi, Qualcomm]
JVT-U308: CE 8: Rewriting. Based on JVT-U043. [A. Segall, Sharp, HHI, Nokia, BT,
Motorola, Mitsubishi, Siemens, Orange, Thomson, NEC, MS, LG, RWTH, Ericsson,
Qualcomm]
v. Error resilience core experiment output documents
None.
vi. MVC core experiment output documents
JVT-U309: CE 9: Illumination compensation. Based on JVT-U052, JVT-U031. [Y. Su,
Thomson, HHI, Nokia, KHU, KWU, Samsung, Huawei, MS]
JVT-U310: CE 10: View synthesis. based on JVT-U063, JVT-U093, JVT-U102 [A. Vetro,
Mitsubishi, HHI, GIST, Nokia, Nagoya, KETI, LG, Orange, Samsung, NTT, KHU,
Qualcomm, Huawei, Sharp, MS]
JVT-U311: CE 11: Disparity and motion vector coding. Based on JVT-U040, JVT-U091
[H.-S. Koo, LG, Sejong U, Nokia, HHI, Huawei, KHU, Tsinghua, ETRI, KETI, Samsung,
Yonsei U, KWU, SNU, MS]
j. JVT administrative and liaison topics
i. Meeting opening remarks by the chairmen
Opening remarks: The chair remarked that there have been many late and badly-formatted
document uploads. A better method of handling document submissions is needed.
The chair also expressed concern regarding his perception of a lack of sufficient editorial
competence and dedication for draft amendments and the draft AVC corrigendum work.
The chair indicated that perhaps the highest priority for this meeting is to finalize the work on the
new 4:4:4 profiles to prepare for ITU-T Consent next month. We will start the major work on
that topic on Sunday morning. Two other high priorities include progressing beyond JVT-T210
toward a mature corrigendum and progressing the work and assessing the status of our SVC
project. One thing that is critical to all of those projects is the great need for editorial diligence
for clarity and consistency. I see that as a critical need for the JVT at this time. MVC, of course,
is another major focus although on a somewhat longer time-scale.
ii. JVT working practices
JVT documents are available at http://ftp3.itu.int/av-arch/jvt-site.
132
These can also be accessed via ftp with the site name ftp3.itu.int, user ID avguest and password
Avguest. Upon login, documents are found in the directory "jvt-site". Uploading of
contributions is done by upload via ftp protocol to the "jvt-site/dropbox" directory.
JVT email lists are managed through the site http://mailman.rwth-aachen.de/mailman/options/jvtxyz, and to send email to one of these reflectors, the email address is "jvt-xyz@lists.rwthaachen.de", where "xyz" is
– "experts" for general experts group discussions
– "bitstream" for bitstream exchange activities
– "svc" for SVC work
– "mvc" for MVC work (new starting at this meeting)
iii. Scheduling notes
The meetings on Friday 20 October 2006 ran approx 2:30 p.m. to 7 p.m.
Started Saturday at 9 a.m., ran to 8:30pm.
Continued resampling Saturday.
SVC High-level syntax intended for Saturday, but key proponents not present.
Sunday starting at 9 a.m.:
MVC
4:4:4
CE9 work planned not before Sunday.
Revisiting planned for Tues a.m.
…
iv. Closing session notes
At the closing session, there were no requests to review the outcome of non-normative issues.
Thanks were expressed by the JVT to the meeting host and to WG11 for holding the JVT
meeting under its auspices.
The meeting was closed at 1:40 pm on Friday 27 October 2006.
v. IPR policy reminder
Participants were reminded of the IPR policies established by the parent organizations of the JVT
and were referred to the parent body web sites for further information. The IPR policies were
summarized for the participants.
Participants were particularly reminded of the need to supply a completed JVT IPR status
reporting form in all technical proposals for normative standardization. Participants were also
reminded of the need to formally report patent rights to the top-level parent bodies (using the
twin text form on the database found below) and to make verbal and/or document IPR reports
within the JVT as necessary in the event that they are aware of unreported patents that are
essential to implementation of a standard or of a draft standard under development.
Some relevant links for organizational and IPR policy information are provided below:
133
–
–
–
–
http://ftp3.itu.int/av-arch/jvt-site (JVT contribution template for each meeting)
http://www.itu.int/ITU-T/studygroups/com16/jvt/index.html (JVT founding charter)
http://www.itu.int/ITU-T/dbase/patent/index.html (ITU-T IPR database)
http://www.itscj.ipsj.or.jp/sc29/29w7proc.htm (SC29 Procedures)
The chair invited participants to make any necessary verbal reports of previously-unreported IPR
in draft standards under preparation and opened the floor for such reports: No such verbal reports
were made.
vi. Late documents
No objections were voiced to the consideration of the late documents. Documents not listed in
this report with a "*" were classified as late. Those documents will only be considered as
information documents only (unless agreed otherwise by the group) if time permits, and
consideration of them may be shifted to the end of the meeting as determined appropriate by the
group. Documents suffixed by "-L" below were the least late and were available by the first
meeting day; and those suffixed by "-M" were more late than that.
No objections voiced at opening session.
JVT-U021-L [W. Yao, Z. G. Li, S. Rahardja] Balanced inter-layer prediction
JVT-U029-M [A. Leontaris, A.M. Tourapis, K. Suehring] ME & MC Enhancements to JM ref
soft
JVT-U030-L [A.M. Tourapis, K. Suehring, G.J. Sullivan, A. Leontaris] Revision of JM ref
software manual
JVT-U031-L [J.-H. Yang] CE11: Illum. comp. consistent pred.
JVT-U032-L [Z. Lu, J. Zheng, W. Lin, S. Rahardja] Percept. Deblock Filter for ROI SVC
JVT-U034-L [B.-K. Lee] CE3: Improved context modeling PR slices
JVT-U039-L [F. Le Leannec, P. Onno, X. Henocq] CE6: Cross-verif Thomson's JVT-U025
JVT-U040-L [S.-H. Lee, S.-H. Lee, N.-I. Cho, J.-H. Yang] MVC: Disparity vector prediction
JVT-U044-L [A. Segall] Transcoding in Scalability Info SEI
JVT-U046-L [W.S. Shim, H.S. Song, Y.H. Mun, J.B. Choi] High-level syntax for flexible I
frame position
JVT-U076-L [X. Ji] CE5: Improv FGS for low-delay
JVT-U077-L [X. Ji] Block based FGS for low-delay
JVT-U078-L [L. Zhang, X. Ji, D. Zhao, W. Gao] Adapt. spatial & transform domain FGS
JVT-U082-L [H. Kirchhoffer, D. Marpe, T. Wiegand] CE3: Improved CABAC for PR slices
JVT-U083-M [H. Kirchhoffer, D. Marpe, T. Wiegand] CE3: Verif JVT-U034-L PR context
model
JVT-U084-L [D. Marpe, G. Marten, T. Wiegand] Fast CABAC renorm for H.264/MPEG4-AVC
JVT-U086-M [A. Eleftheriadis] Prop SVC profile for videoconf
JVT-U087-M [B.K. Lee] CE3: Verif JVT-U082-L PR slice CABAC
JVT-U088-L [W.-J.Han, B.-K.Lee] CE5: Verification of ETRI JVT-U073
JVT-U090-L [S.-W. Park, B.-Y. Jeon] Usage of store_base_rep_flag
JVT-U091-L [H.-S. Koo, Y.-J. Jeon, B.-Y. Jeon] MVC motion from neighbor view
JVT-U094-M [S. Jeong, K. Moon, J. Hong] CE5: Verif Tool 3 JVT-U076-L L-D FGS
JVT-U095-L [J. Xu] CE4: Improv inter-layer pred
JVT-U096-L [J. Xu] CE5: Verif JVT-U077-L and JVT-U076-L
JVT-U097-L [E. Francois] CE6: Verif Nokia JVT-U130 ESS
JVT-U099-L [S. Sekiguchi, Y. Yamada, K. Asai] Advanced 4:4:4 profiles
JVT-U103-L [Y.-K. Wang, Y. Chen, M. M. Hannuksela] Comments to JMVM 1.0
JVT-U104-L [Y.-K. Wang, Y. Chen, M. M. Hannuksela] Time-first coding for MVC
134
JVT-U105-L [Y. Chen, Y.-K. Wang, M. M. Hannuksela] MVC reference picture management
JVT-U106-L [Y. Guo, Y.-K. Wang, M. M. Hannuksela, H. Li] Discardable data adaptation
JVT-U107-L [Q. Shen, Y.-K. Wang, H. Li] Adaptive inter-layer prediction
JVT-U109-L [Y.-K. Wang, M. M. Hannuksela] On SVC high-level syntax
JVT-U111-L [Y.-K. Wang, M. M. Hannuksela] SVC HRD
JVT-U112-M [Y. Chen, Y-. K. Wang, M. M. Hannuksela] SVC ref pic list construction
JVT-U113-M [Y. Guo, Y.-K. Wang, H. Li] CE9: Verif JVT-U057 redund slices
JVT-U114-M [C. Zhu, Y.-K. Wang, H. Li] Adaptive redundant picture coding
JVT-U117-L [H. Schwarz] CE8: Verif JVT-U043 SVC-to-AVC
JVT-U118-L [J. Jia, H.K. Kim, H.C. Choi, J.G. Kim] Terms for SVC access unit def
JVT-U122-L [S. Sun] Verif QCOM JVT-U126 smooth ref
JVT-U124-L [S. Kamp, M. Wien] Low-delay leaky base layer
JVT-U127-L [J. Ridge] Mobile profile for SVC
JVT-U128-L [J. Ridge, X. Wang] CE1: Improve FGS VLC
JVT-U129-L [J. Ridge, X. Wang] Component separation FGS
JVT-U131-M [X. Wang] Verif RWTH-Aachen JVT-U055
JVT-U133-M [S.-T. Hsiang] Intra subband/wavelet framework
JVT-U134-L [H. Kimata, S. Shimizu] On direct mode for MVC anchors
JVT-U135-M [S.-T. Hsiang] CE1: Verif Nokia JVT-U128-L
JVT-U136-L [S. Sekiguchi] Prop changes to 4:4:4 draft
vii. Withdrawn document registrations
The following document contribution registrations were withdrawn by the request of their
registrants.
JVT-U045
JVT-U056
JVT-U075* [D.Y. Suh, G.H. Park, J. Oh, M. Park] CE9: JVT-S028 extension redundant pic
(withdrawn)
JVT-U089
JVT-U092
JVT-U108 [Q. Shen, Y.-K. Wang, M. M. Hannuksela, H. Li] Ref pic marking for temporal SVC
(withdrawn)
JVT-U142 [T. Suzuki] Prop DCOR AVC/FRExt conformance (withdrawn)
JVT-U148 [T. Tran, L. Liu, P. Topiwala] Filtering for ESS (withdrawn)
viii. Administrative documents
JVT-U000 List of documents of Hangzhou meeting
JVT-U001 [G.J. Sullivan, J.-R. Ohm, A. Luthra, T. Wiegand] AHG Report: Proj mgmt and
errata
JVT-U002 [T. Wiegand, K. Suehring, A. Tourapis, K.P. Lim] AHG Report: JM text and ref
soft
135
JVT-U003* [T. Suzuki] AHG Report: Bitstreams & conformance
JVT-U004* [J. Vieron, M. Wien, H. Schwarz, L. Bivolarski] AHG Report: JSVM s/w and
new func. integ.
JVT-U005* [J. Reichel, H. Schwarz, M. Wien] AHG Report: JSVM & JD text
JVT-U006* [S. Sun, A. Segall, J. Reichel] AHG Report: Spatial scalability resampling
Surveys the contributions relating to spatial scalability resampling, including topics in CE 4, 6,
and 7. Review and further study recommended.
JVT-U007* [Y.-K. Wang, S. Pateux, P. Amon, T. Schierl] AHG Report: High-level syntax,
err resil
JVT-U008* [M. Wien, H. Schwarz] AHG Report: Coding eff & JSVM perf test cond
JVT-U009* [T. Suzuki] AHG Report: Study of 4:4:4 functionality
See section JVT 4:4:4 coding normative modifications of this report.
JVT-U010* [J. Vieron] AHG Report: SVC interlaced coding
JVT-U011* [J. Ridge, D. Marpe, G. Sullivan] AHG Report: SVC quantization, CAVLC,
CABAC
JVT-U012* [M. Mathew, J. Li, H. Schwarz] AHG Report: Bitstream extractor
JVT-U013* [H. Schwarz, Y. Bao] AHG Report: Complexity reduction
JVT-U014* [S. Kamp, X. Wang] AHG Report: AR-PR and PR slices
JVT-U015* [A. Vetro, Y. Su] AHG Report: MVC H-L syntax & buffer mgmt
JVT-U016* [H. Kimata, A. Smolic, Y. Su, A. Vetro] AHG Report: JMVM text editing
JVT-U017* [P. Pandit, A. Vetro] AHG Report: JMVM soft & new func integ
136
ix. JVT Liaison communications
Two incoming liaison statements were received from SMPTE. They are discussed below in
section 13 of this report. A liaison reply to each of those incoming liaison statements was sent by
the MPEG parent body as documented below.
Additional liaison statements arrived from ITU-R SG6/WP 6J and ITU-T SG 9 as described
below in this section.
JVT-U154* [ITU-R SG6/WP 6J] LS: Colour spaces
ITU-R Working Party 6J would like to draw to the attention of ISO/IEC JTC 1/SC 29/WG 11 to
the concerns that ITU-R Working Party 6J has with respect to a number of documents – including
in particular, the amendment to AVC for enhanced colour space support.
Good input – to be taken into account during final editing process. JVT decision: Agreed.
Liaison reply sent by MPEG parent body as documented below.
JVT-U157* [ITU-T SG 9] LS: On MVC
ITU-T Study Group 9 thanks ISO/IEC JTC1/SC29/WG11 (MPEG) for its liaison letter informing
about the work on Multi-view Video Coding (MVC) and ISO/IEC 23002-3. SG9 informs that
they have started the study to develop a draft Question on the Free-viewpoint TV (FTV) system
toward future standardization especially from a view point of transport system aspect. Since they
are reportedly in the very early stage of the study, they are surveying requirements and
technologies of a whole FTV system. They are expecting MVC to be a potential technology to
encode video signals for FTV system. They report that they would appreciate the provision of
information of the FTV system. And they also look forward to receiving further information
regarding the MVC specification and its progression. They plan to keep ISO/IEC
JTC1/SC29/WG11 (MPEG) informed of their progress on this issue.
No JVT action needed – liaison reply sent by MPEG parent body as documented below.
k. JVT SVC normative modifications
i. CE 1 & related docs: PR slice VLC
JVT-U125* [Y. Bao, Y. Ye, M. Karczewicz, P. Sagetong] CE1: Results PR slice improve
CE1 combines the proposal JVT-T086 and JVT-T087 into one software. In JVT-T086, an
adaptive VLC scheme was presented, and in JVT-T087, a block-based FGS coder for the purpose
of reducing the complexity is presented.
Fri 20 presentation postponed.
Sat 21 presentation postponed
From JVT-T087, Cycle-aligned fragment (aligned with fragment boundaries). Claim to improve
the error resilience. Macroblock header in PR slice similar to CGS case.
CAF has only little impact on compression performance.
From JVT-T086, two changes: Adaptive VLC and special EOB
Significance coding in JSVM6.8: special EOB after run max has an additional space
unnecessarily, position only depends on number of remaining zeros.
137
Av. Bit rate reduction by specal EOB and ad. VLC:
4CIF; -2, -2.5%
CIF -1.1, -1.2
QCIF -0.1, -0.1
In average over all, only a small difference was reported.
JVT-U153-L [X. Ji] CE1: Verif JVT-U125 PR slice
The purpose of this report is to verify proposal JVT-U125 titled ‘Report of core experiment on
PR slice improvements (CE1)’ from Qualcomm Inc. As a verification task, coding performance
check was carried out. The results presented by Qualcomm in JVT- U125 were reported
confirmed.
JVT decision: Adopt CAF
JVT-U128-L [J. Ridge, X. Wang] CE1: Improve FGS VLC
This contribution reviews contribution JVT-T086, reportedly providing some insights on the
source of gain or loss in the claimed results. Results suggest that adaptivity in the VLC of PR
slices actually leads to a loss in performance, with most of the claimed gain coming from the
special end of block (SEOB). A possible improvement to JVT-T086 is introduced, which is
claimed to eliminate loss at QCIF and provide a relatively significant improvement at 4CIF (from
0.81% in JVT-T086 to 1.41%), with less text modification.
Results indicate that consideration of methods from JVT-T086 is not worthwhile. Operation
points where small improvements could be possible, same can be achieved by non-normative
tools. Does not contribute to reduce the complexity of FGS.
Note: Currently, considering software runtime, FGS is allegedly three times as complex as base
layer decoder (containing MC, DF + residual decoding). There is a claim that the factor is <3, but
no number was provided.
Non-normative suggestion developed to provide intended benefit.
Suggests not to adopt normative changes previously under consideration.
Some (relative) improvement shown with alternative method developed in interim period.
However, overall impact reported to be very minor.
Contribution noted. No action taken (as recommended in contribution).
JVT-U129-L [J. Ridge, X. Wang] Component separation FGS
Currently in SVC, there is the provision for color components to be separated in a PR slice, so
that all luminance data is decoded prior to all chrominance data. This reportedly greatly assists
low complexity editing or analysis operations. However, reportedly due to an oversight, entropy
decoding must still be performed since there is no separation marker between the color
components. This proposal would add such a separation marker, reportedly with negligible
impact on efficiency.
The separation marker would be like a start code.
138
Remark: Why this particular marker? – we could think about a variety of boundary delineation
markers that would have hypothetical value in some scenario – is this the rational and
conceptually-consistent choice? Use an SEI message to indicate boundaries in data?
JVT decision: Adopt SEI message as described in (JVT-U129r1-L).
Byte alignment between the luma and chroma components requires to flush the CABAC engine
(typically around 3 bits) plus on average 4 bits for the alignment itself (JVT decision: adopt this).
The separate decoding capability could then be achieved by an SEI message, still to be worked
out.
What about enforcing byte alignment between luma and chroma and between chroma
components? JVT decision: Agreed.
JVT decision: Adopt as in r1.
JVT-U132* [M. Karczewicz, R. Panchal] Refinement coef coding
This contribution proposes to replace the adaptation used to select VLC table for coding of the
refinement coefficient by signaling to the decoder which table should be used for which
macroblock type (Inter or Intra). The proposal aims to reduce decoder complexity and ensure
proper table selection when both macroblock types are present within one slice. The contribution
further proposes to extend the ideas presented in document JVT-T077 to increase coding
efficiency of VLC refinement coefficients coding. The proposed changes mainly affect Intra
coded slices and for those slices the improvements are in the range 3-7% for 3 FGS layers.
Proponent not present.
Sat 21 Presentation postponed
First part: adaptation removal
Second part: mb-level refinement coefficient signaling
Current adaptation adds complexity and may decrease compression performance. Proposal to
remove the adaptation in case of mixed mode slices. Using method from JVT-T077 gives 5-7%
reduction at highest rate point.
Seems that something is broken in the adaptation.
JVT decision: Establish CE to find whether a) something is broken and the adaptation should be
removed. b) whether macroblock-adaptive signaling should be used.
JVT-U159-M [L. Cieplinski] Verif JVT-U132 coef coding
Verifier mainly supports the first part (removal of adaptation)
Verifies JVT-U132. Verifier suggests to adopt adaptation removal and has no opinion about mblevel refinement coefficient signaling.
JVT-U135-M [S.-T. Hsiang] CE1: Verif Nokia JVT-U128-L
139
Verifies JVT-U128-L.
Reportedly verified without closely looking into it.
ii. CE 2 & related docs: Switching
JVT-U037* [F. Le Leannec, P. Onno, X. Henocq, J. Takeda] CE2: Switching PR slices
In the current JSVM, the coding efficiency of PR enhancement layers can be optimized by
adjusting the AR-PR leaky factors. However, the choice of a leaky factor that would provide
good coding performance may increase the potential drift that can be obtained when switching
from a PR layer to an upper one on the decoder side. This tool aims at gathering the high coding
efficiency obtained with AR-PR leaky factors chosen for a good coding efficiency, together with
the ability to quickly recover video quality when increasing the decoded PR rate. To do so, a
switching PR signal is introduced between two successive layers (base layer and a PR layer on
top of it), in which a residual signal between a current frame at a target quality level and a
reference frame calculated from a decoded reference picture at a lower quality level is calculated.
This calculated residual signal is then encoded conforming to the Progressive Refinement slice
syntax.
This is called "Tool 2" of CE2.
Additional slice type – "SPR slices". Intended to enable switching between FGS layers.
Remark: Assumes instantaneous rate switching decision & feedback information availability at
sending side. Is this realistic? Reply: Yes, that assumption is presumed valid, at least for some
unicast scenarios.
Remark: Also some amount of switching data overhead needed. Some presented comparisons
may not fully account for that.
Remark: AR-FGS seems rather complex, this further complicates it. Reply: This is not using
new things on top of AR-FGS, but rather ordinary FGS to encode switching data.
Remark: Relation to profile decision?
Requested to provide further analysis of overhead and delay latency.
Text changes drafted? Has anyone reviewed them?
Make presentation deck available.
Additional notes:
Takes up to 20 frames after full quality of next-higher FGS layer is reached with current ARFGS. New type of switching PR slice (to be transmitted at time of switching) would allow to
reach full rate immediately (rate adaptation). Other application in case of packet losses where
switching slices are periodically inserted depending on packet loss rate. Gain reported in
particular for case of 2 FGS layers. Average frame PSNR gain reported 0.35-0.4 dB.
Questionable points: Realistically, sending switching slices on request could not be done in real
time, a delay would occur that is not considered in the PSNR figures. For case of packet losses,
the rate is increased. No information is given about the overhead. In principle, it would also be
possible to use AR-FGS with alpha > 2 to achieve faster recovery (but then also increased rate).
Realistically, a comparison would only be possible at same rates (switching frame overhead or
140
AR-FGS with higher alpha). Therefore, in both cases the reported PSNR gains would clearly be
lower in a realistic application scenario.
Requires a fair comparison with AR-FGS when using a smaller leak factor.
Proponent shows that rate is increased by up to (sometimes above) 10% for the case of higher
packet loss where the main advantage is claimed.
JVT decision: Continue CE. Bring more convincing comparison where the scalable codec is also
run in a mode that has better resilience/faster recovery against packet losses (larger alpha for
more leak, with comparable increase of bit rate). As scalable codecs have graceful degradation
property under data losses if operated correctly, it is questionable whether switching slice type
would be necessary (note: switching slices hardly used even in non-scalable codecs).
JVT-U051* [J. Jia, H.K. Kim, H.C. Choi, J.G. Kim] CE2: Tool 2 Verif Canon JVT-U037
Sw PR
This document reports the cross-check results for the SVC CE2 Tool 2 “Switching PR Slices”
proposal by Cannon as described in JVT-U037. The verification has been performed by decoding
the coded bitstreams which are provided by Cannon. Both of the PSNR against the original
sequence and the bit-rate according to the file size have been verified. It is shown that the RD
results obtained by decoding the provided bitstreams match the experimental results presented in
JVT-U037 quiet well. Verification results are shown in JVT-U051_results.xls.
Software (source code) and bitstreams provided by Canon.
JVT-U050* [J. Jia, H.K. Kim, H.C. Choi, J.G. Kim] CE2: Tool 1 SP Picture for SVC
Switching
This contribution presents a design and implementation of SP picture for SVC switching
functionality. Originally performance illustration regarding both the SP picture for switching and
the SP picture for non-switching would be presented in this contribution. However, due to the
incorrect implementation of the proposed method, around 0.2 dB PSNR mismatch exists between
the decoded SP picture for switching and the decoded SP picture for non-switching, which is
reported in the primitive experiments. Therefore, no further simulation has been done. Simulation
results for illustrating the performance of the proposed method would be provided after the bug
fix.
Presentation deck reported to have been uploaded.
Not conclusive due to bug.
Simulation results not adequate – bug. JVT decision: Further work suggested after this meeting.
Remark: Relation to profiling? Lack of uptake of SP concept in current standard.
Question: Any idea of how much gain expected? Reply: About 1 dB compared with IDR picture
expected in base layer; 0.2 to 0.6 dB reported at last meeting in enhancement layer.
JVT-U038* [F. Le Leannec, P. Onno, X. Henocq] CE2: Cross-verif ETRI/Sejong JVT-U050
141
This contribution verifies document JVT-U050, ETRI / Sejong University’s response to CE2.
The verification has been performed by decoding the provided bitstreams with the provided
binaries. The simulation results of JVT-U050 are confirmed.
JVT reinforcement of internal working practices: The use of binaries is not JVT practice
(although this is not a concern at this time for this contribution).
iii. CE 3 & related docs: PR slice CABAC
JVT-U034-L [B.-K. Lee] CE3: Improved context modeling PR slices
In this contribution, a modified context modeling for PR slices is proposed. This contribution
includes new methods of separating significant coefficients in PR slices. In JD7 accumulated
coefficients are used to separate significant and refinement pass. It is claimed that the probability
of zero coefficients in significant and refinement pass is not optimal above second FGS layer. For
the presented simulation results, an IPPP coding structure is used and AR-FGS is enabled. In this
contribution, a redefined FGS coefficient partitioning is proposed. It is shown that bit-rate
savings can be up to 5.2% at highest bit-rate points of CIF sequences in third FGS layer can be
obtained.
Previous proposal number JVT-T034 (without change).
Signficant bit rate savings shown relative to current JSVM.
Compared to JVT-U082-L, worse performance with offset ½, similar performance claimed for
1/3 and ¼.
For f = 1/3, both this proposal and JVT-U082-L are reported to provide about the same amount of
benefit (JVT-U082-L proposal slightly better). For f = 1/2, JVT-U082-L more significantly
better. Overall, JVT-U082-L generally better in coding efficiency terms.
Less benefit in IPPP case than intra case.
Asserted benefit relative to JVT-U082-L is in terms of context memory. Difference in quantity
of bins is asserted to not be very significant.
Remark: Consider worst-case difference in bins.
Remark: Would be nice to have more consistent result measurement.
Remark: Both methods add some complexity.
JVT-U083-M [H. Kirchhoffer, D. Marpe, T. Wiegand] CE3: Verif JVT-U034-L PR context
model
This contribution reportedly verifies JVT-U034-L, CE3: Improved context modeling PR slices.
The verification was reportedly performed by decoding randomly-selected bitstreams with the
provided software.
JVT-U054* [Y.H. Tan] CE3: Modif CABAC for MC for M-R FGS
Simplified motion refinement was proposed in JVT-T027. By simplifying the motion estimation
process in the base layer and systematically assigning MB type of macroblock in the progressive
refinement enhancement layer, the encoding time of a fine grain scalable video with motion
142
refinement in the enhancement layers is reported to be substantially decreased. While motion
refinement is reported to increase the encoding time of a fine grain scalable video by 4 times, the
simplified scheme is reported to achieve comparable gains without significantly increasing the
complexity of the encoder. Since the MB type is systematically assigned, some information in the
enhancement layer become redundant.
Motivated by desire for encoder complexity reduction. Constrains some aspects of base layer
coding decisions. Considers elimination of syntax that is not being used in reduced-complexity
encoder operation. Small gains shown for doing so (e.g., 1% of some subset of the bit rate),
relative to imposing the same constraints on the encoding algorithm without having special
syntax to support the constrained operating mode.
The method is mainly related to encoder optimization (fast mode decision). The gain of 1% with
restricting the EL MB types relates to the case where the optimized encoder is used. The method
itself performs however slightly worse than current JSVM with motion refinement. Could later be
used in the context of developing a fast SVC encoder. Gain of up to 1.5 dB for the high end of
the third FGS layer claimed in a particular comparison. For intra-only, average BR saving of
1.9% for the second and 4.3% at the third FGS layer, for IPPP reducing to 0.61% at second and
3.2% at the third layer.
Does this harm the ability to independently parse the enhancement layer bitstream? Perhaps.
Contribution noted. No action taken.
JVT-U082-L [H. Kirchhoffer, D. Marpe, T. Wiegand] CE3: Improved CABAC for PR
slices
This contribution proposes a CABAC context modeling scheme for coding of refinement
information in PR slices. The proposed modeling approach is a layer-specific extension of
context modeling for refinement symbols as currently specified in the JD. It has been derived in
JVT-T077 by analyzing the specific properties of the quantization process in FGS coding. The
improved modeling scheme is a generic approach in the sense that it is independent of the
specific choice of dead-zone parameters or classification rules. For a typical choice of dead-zone
parameter f = 1/3 and averaged over the whole test set, BD rate savings of 1.9% and 4.3% for the
second and third FGS layer, respectively, have been reported in the case of intra-only coding.
Evaluates JVT-T077 approach (without change). Uses 16 additional context models.
Quantizer intervals are either sub-divided into three (when it is the inner interval) or into one
follow-up intervals (except 3-level base layer where the outer intervals are sub-divided into two
intervals). Proposal to observe this in the context models. This requires 3 models in the second
and nine models in the third FGS layer. In total, 16 additional models are required as compared to
current design. It would however be necessary to track the context model choices across FGS
layers. Claim is made that this could be achieved without extra memory when instead the
memory for context model is used for reconstruction of the transform coefficients as well. Gains
are highest at those points which would typically not be used (e.g. QCIF at high end of layer 3 is
up to 45 dB).
Significant improvement shown.
For f = 1/3, both this proposal and JVT-U034 are reported to provide about the same amount of
benefit (this proposal slightly better), but this approach is not tuned to a specific choice of the f
parameter, whereas the JVT-U034 approach is reportedly restricted in effectiveness to that use.
143
Remark: Substantial benefit has been shown for giving an encoder freedom to choose f
adaptively.
For alternative approach proposed in JVT-U034, a need for processing of more bins is asserted
than what are needed for this proposal.
JVT-U087-M [B.K. Lee] CE3: Verif JVT-U082-L PR slice CABAC
The purpose of this contribution was reportedly to check the validity of results provided in JVTU082-L. After extractor and decoder scripts, it was reportedly verified that there was no problem
with provided bitstreams, based on verification of a random sampling of the results. The modified
(source code) software provided by Fraunhofer HHI was reportedly used in the verification.
Most there is a bug in MMCO implementation (memory leak).
JVT decision: Adopt, under condition that results are shown to be still consistent after bug fix
and PR slices are kept in SVC. Fully verified with results up-loaded.
JVT-U084-L [D. Marpe, G. Marten, T. Wiegand] Fast CABAC renorm for H.264/MPEG4AVC
This informative contribution presents a fast standard-compliant realization of the
computationally expensive renormalization part of binary arithmetic coding in H.264/MPEG4AVC. It is reported that the proposed technique allows to replace time-consuming, bitwiseoperating input and output as well as bitwise carry-over handling in a conventional CABAC
implementation with corresponding routines operating in units of multiple bits. Experimental
results reportedly demonstrate that the proposed method enables a considerable speed-up of both
arithmetic encoding and decoding in the range of 24 to 53% average run time.
iv. CE 4 & related docs: Resampling
See also the JVT-U006 AHG report.
JVT-U065* [S. Sun, V. Bottreau] CE4: Texture upsampling results
Information document.
This contribution summarizes the coding results based on the test conditions defined in CE4.
Potential issues are discussed according to the performance analysis.
Difference from single-layer is generally larger at high fidelities (above 40 dB in particular –
where coding the remaining "noise" is not useful for prediction of the next layer), and varies
substantially from source sequence to sequence.
IPPP and QCIFCIF seem the difficult cases, esp. CAVLC (CAVLC not tested specifically in
this document, but this is the general impression).
High penalty for ESS in the long delay configurations that were tested. This was somewhat
expected.
144
See also notes on JVT-U145-L.
JVT-U024* [Y. Yan, J. He, Y. Prieto] On CE4: Dyadic spatial resampling
This document focuses on the dyadic spatial down- and up-sampling for all-intra-coding and
large-delay coding. Two issues are addressed in this proposal. First, instead unified with the
generic extended spatial scalability (ESS), the dyadic case is separately addressed to try to
improve its coding efficiency and reduce its computational complexity. Secondly, this
contribution proposes filters that are shorter than those in the current design, for further
complexity reduction. The proposal reports a comparison of the proposed upsampling and
downsampling filters to the ESS filters of JSVM_6_3 (8-tap down-sampling and 6-tap upsampling) for the dyadic case. It is reported that the proposed filters (5-tap down-sampling and 4tap up-sampling) provide less complex filtering process and also offer comparable coding
efficiency for the enhancement layers. The experimental results by use of the common test
conditions and the evaluation criterion demonstrate the proposed filters achieve overall average
PSNR gains from 0.02 dB to 0.16 dB for Configuration 1 and from -0.02 dB to 0.08 dB for
Configuration 2 in all-intra-coding and comparable performance for large-delay coding.
Proposal. Focuses on dyadic case.
Based on JSVM 6.3. Compares specific proposed filters (5 taps down, 4 taps up) to current
JSVM filter (8 down, 6 up) and JVT-T057 (8 down, 4 up) for intra and IPPP. Overall gain
reported in high-res fidelity for all-intra case.
Remark: This changes the downsampler – which will change the visual low-res quality.
Presenter replies that the low-res quality is actually improved visually.
CE document established test conditions for proposals that change the downsampler, since this
affects the rate allocation. The recommended CE method was apparently not followed in this
contribution.
Down- and upsampling filters changed. Goes back to old upsampling filters which have the
problem of luma/chroma misalignment at the lower resolution.
Remark: This changes the sampling grid alignment relative to the current design, which it seems
will cause misalignment of luma and chroma. (The contribution did not change the chroma
filtering.)
Remark: This removes the consistency between dyadic and non-dyadic cases, which is desirable
(although perhaps not necessary).
Cannot reach conclusion on this contribution for the above reasons. The above issues would
need to be addressed.
JVT decision: Add to CE.
JVT-U042* [A. Segall, J. Zhao] CE4: Texture Upsampling with 4-tap Cubic Spline
This document reports the evaluation of JVT-T057 within CE4. JVT-T057 proposes a 4-tap
cubic-splice based filter that was originally introduced in contribution JVT-S016. Utilizing the
testing conditions established within the CE, results show a degradation of less than 0.1 dB (delta
bit rate impact ranging from slight improvement to 1.5% worse) for all intra picture coding. The
degradation in coding performance for typical long-delay configurations is negligible for the
145
proposed filter. It is proposed to adopt the 4-tap spline-based filter (JVT-T057) for luma texture
upsampling in order to reduce the computational complexity.
Increase in bitrate in the range of up to 1.5 %, PSNR 0.15 dB decrease for intra, negligible for
intra with long delay. Claim that there is no visual difference.
Results for long-delay case (as opposed to all-intra) – negligible in PSNR/bit rate terms, both
dyadic and ESS. Filter proposed two meetings ago, software has been available for 4 weeks.
Proposal applies to both dyadic and non-dyadic cases.
Question: What about visual effects and specific area localized effects? Reply: Didn't notice any
objectionable effects – proponent reports that they saw no differences. Did the verifier look at
the results visually? Reply: No.
JVT decision: Adopted.
JVT-U064* [V. Bottreau] CE4: Verif Sharp inter-layer JVT-U042
The purpose of this report is to verify proposal JVT-U042 entitled “CE4: Texture Upsampling
with 4-tap Cubic Spline” from Sharp Labs of America. As a verification task, coding
performance check was carried out. For further validation, please see the proposal report. The
results presented by Sharp Labs of America in JVT-U042 are confirmed.
Was presented.
Cross check performed, software changes checked and verified. No subjective viewing
performed.
JVT-U123* [S. Regunathan, S. Srinivasan, C. Tu, S. Sun, G. Sullivan] Flexible 4-tap spat
SVC upsamp
This contribution introduces a flexible framework for upsampling filter selection based on a
family of 4-tap filters. The proposal allows one or multiple filters signaled at the sequence level.
When multiple filters are signaled at the sequence level, the filter index is then signaled in each
slice to specify which filter is to be applied for prediction between a specific base picture and the
current slice. A set of parameterized filters is proposed to allow a high degree of freedom while
keeping syntax overhead very low. Most filters proposed in JVT so far can be parameterized this
way, or at least can be closely approximated this way. The proposal meets the desire to shorten
the SVC upsampling filters from 6 taps to 4 taps in order to reduce the computational complexity.
It also allows flexibility in filter selection for potentially various downsampling options. Coding
performance is demonstrated using a specific set of filters, which outperform the current 6-tap
filters for several sequences. The same filter signaling framework can potentially be applied to
motion compensation filters as well. It is therefore proposed to adopt the flexible 4-tap filter
design for luma texture upsampling into SVC and set up a CE to investigate motion
compensation filter design for the SVC enhancement layers.
Proposes filter family approach rather than a specific fixed filter, constrains to 4-tap, mirror
symmetry, and correct interpolation of input data that falls along a straight line. Filter
characteristics specified by two small integer-valued parameters.
Two applications of filter investigated:
– Spatial SVC (dyadic and non-dyadic)
146
–
Motion compensation interpolation for fractional-sample positions
Proposed syntax allows selection among a set of parameterized filters; experiments used a
particular filter for entire sequences.
PSNR difference negligible. Slightly better results than JVT-U042; improvement asserted to be
due to linearity constraint not imposed in Sharp design.
Example filters, e.g., Catmull-Rom or the tested filter have parameters asserted to take a total of
about 32 bits to represent at sequence level.
Motion compensation experiments with one example filter using JM software. For most tested
sequences, performance difference relative to current AVC MC filter design reported to be very
small. Examples: 0.1 dB average penalty on Mobile and Calendar (up to 0.5 dB at low bit rate
end of the spectrum). Slightly better than JM on City and Crew.
Proposes to adopt this for spatial SVC luma and further investigate for motion compensation.
Cross-verification in progress by Nokia to be reported in JVT-U155-M.
Results for long-delay? Expected PSNR impact very small (smaller than intra-only). Possible
subjective benefit or benefit for pairing with encoder downsampling.
Visually? No significant difference observed.
In typical spatial SVC use (not all intra), complexity benefit not so big, although shorter filters
are obviously nicer implementation-wise.
Remarked that some other contributions bring up the multi-loop decoding concept again, which is
a scenario where upsampling filters would matter more.
Maximum penalty 0.04 dB for QCIF-CIF, 0.06 dB for CIF-4CIF. Penalty for ESS up to 0.18 dB.
Overhead for filter parameters low. Application to MC: Similar performance, except for Mobile
(up to 0.5 dB worse).
Fixed filter setting used. Benefit of adaptation not shown, but previous results (Bangkok) indicate
that adaptation can be beneficial when different downsampler is needed in an application. May
only be useful for the intra case, no evidence that adaptation is useful for long delay.
Cross verification (Qualcomm) not ready yet.
JVT concludes that there is no need to keep 6-tap upsampling in the JSVM.
JVT decision: Agreed to adopt a four-tap filter.
JVT decision: Add to CE on Resampling.
Motion compensation issue for further study.
JVT decision: Create AHG on study motion comp and de-blocking, RCDO. In general, using
lower complexity at higher spatial resolution could be beneficial (e.g. SVC with same complexity
as single layer).
147
Information from Hardware designers would be desirable on whether having different building
blocks at base and enhancement layers would be a problem. People claim that data flow is more
critical than designing new blocks.
Thomas points out problem of macroblock scan order over spatial layers in case of non-dyadic
up-sampling (except factor 1.5). Breakout chaired by Heiko to study the problem.
JVT-U147 is also related to this issue.
Consider exact 4-tap spatial SVC approach (whether to use family approach or specific filter
approach, and if not family approach, what specific filter).
JVT decision: Further study on MC aspect encouraged.
SVC enhancement layer could hypothetically have lower complexity than the base layer, to make
up for the complexity increase of supporting scalability. Interest expressed on that topic. Note
the techniques of H.241 RCDO – it was suggested to consider to study that.
Consider deblocking.
Question about hardware perspective: Response from a hardware implementer – complexity is a
big concern, if SVC is much worse than AVC in complexity terms, commercialization will be
difficult. Will that still be true 2 years from now (when the time comes to actually implement
what we are now standardizing)? Maybe not so much by then.
Memory bandwidth requirements were suggested to be more critical than operation count
aspects.
Further study of the complexity landscape encouraged.
JVT decision: Create AHG or CE to investigate
Mandate: Study complexity aspects of SVC, including
– Deblocking (including contribution JVT-U020)
– MC
– H.241 RCDO
– Interaction of scanning order and scalability reordering issues (esp. dyadic, e.g.,
supermacroblock structure)
BoG discussions were held on this (esp. scan order): Heiko coordinated.
The Filter from JVT-U042 is (except for rounding) identical with one configuration of the
adaptive filter.
JVT decision: Add to CE.
JVT-U155-M [Y. Bao] CE4: Verif JVT-U123 upsampling
The purpose of this report is to verify proposal JVT-U123 “Flexible 4-tap Filters for Texture
Upsampling in Spatial Scalability” from Microsoft. Coding performance data were generated by
using the executables and the scripts provided by the proponents. The results generated based on
Dyadic spatial scalability all-intra and non-Dyadic spatial scalability all-intra test conditions are
confirmed.
148
The results on applying the filters to motion compensation were not tested in this verification
effort.
JVT-U147-L [T. Tran, L. Liu, P. Topiwala] Down/up-sampling filter for SVC
FastVDO reports that they have proposed very low-complexity resampling filters in the dyadic
case for the past three meeting cycles. This proposal provides updated results for those
previously-proposed FIR lowpass filters that can be employed as dyadic down-sampling and upsampling filters in SVC. These short filters provide a complexity vs. performance tradeoff. For
complexity reduction, a 16-phase polyphase filter approach is not used, saving complexity
substantially (esp. in hardware) with little or no reported sacrifice in performance. These filters
reportedly have their roots from the wavelet theory, which, according to FastVDO, has long been
established to have excellent interpolation characteristics. The contribution asserts that, relative
to existing filters in the SVC design, coding efficiency does not necessarily have to be sacrificed
by employing short low-complexity integer-coefficient filters. The contribution also asserts that
some of the designs proposed by FastVDO can also be applied to the sub-sampling of
chrominance components. The filters proposed by FastVDO have not been tested in the ESS
case.
Propose as best case 5/7 (down/up) filters. This leads to different base layer signal. Results
reported for non-coded case and for case where only the base layer is encoded (intra coding).
Would be necessary to look at the overall rate (base and enhancement layers). For the case
presented, the downsampling filters probably retain sharper (but also higher alias) images,
therefore the upsampled enhancement layer without encoding would be closer to the original.
Furthermore, such filters would then require higher BL rates. In fact, base layer rates shown
(QCIF starting from 350-500 kbit/s seem to be very high.
If possible, perform subjective viewing with Tobias. Done.
(The CE conditions were not exactly followed, because QP condition for both layers were
required)
Put into CE for reporting at next meeting.
JVT-U161-M [J. Ridge] Verif JVT-U147 resampling
The 5-7 filter results for “Config 1” presented in JVT-U147 were reported to have been verified
and found to be accurate.
These results correspond to “Config 1” of the dyadic intra-only case. Verification was reported
to have been performed using provided binaries. A visual examination of the results was not
performed.
JVT-U121* [V. Bottreau] CE4: Verif LG JVT-U089 interl
The purpose of this report is to verify proposal JVT-U089. As a verification task, encoding
performance check was carried out. For further validation, please see the LG proposal report. The
results presented by LG in JVT-U089 are partly confirmed.
NOTE: JVT-U089 was withdrawn!
JVT-U095-L [J. Xu] CE4: Improv inter-layer pred
149
In JVT-T081, a new method called in-scale prediction is introduced to improve the efficiency of
inter-layer prediction. In the proposed method, the prediction of high resolution image data
consists of up-sampled low resolution reconstructed image and the high-pass information of the
inter-frame prediction. This proposal presents further improvement on the in-scale prediction
technique. Beside B-frames, the proposed method is also applied to P-frames. The motion
estimation and mode decision process is modified to make in-loop ME and MC possible and
facilitate the selection of parameters for the new prediction mode. And more experiments have
been done to show the improvement of coding performance.
Attempted upload of new version failed. Submitter provided another copy.
Significant benefit reported. Crew sequence more than 2 dB in some cases. Large gains reported
on some other cases (not as much).
Remark: Are the results only showing a benefit in unrealistic scenarios where the base layer
quality was too hight relative to the enhancement layer quality?
Remark: Relationship to smoothed reference prediction?
Remark: Causes problem with intra refresh for error resilience behavior?
Limited results provided. Have not yet implemented FGS. Limited selection of QP values for
base and enhancement layer.
Would like to see results with other base layer QPs and relationship to smoothed reference, FGS,
etc.
Alternative prediction mode “in_scale” which tries to predict the high spatial resolution by using
the lowpass component from lower spatial layer and the highpass component from previous
frame. Claim that this is beneficial because lowpass has lower correlation over time.
Results provided with lower spatial layer QP setting of 20, PSNR in range of 45 dB. High gain of
2 dB for this mode of operation which would hardly be used in real applications. Questionable if
similar gain would be possible when same quality is set at both spatial layers. No FGS coding,
results achieved by varying the QP in the higher spatial layer.
Present results on low-delay comparison where settings were QP settings were derived from the
(intra) resampling conditions. Typically indicates best gain (up to 1 dB) at lowest rate of the
higher resolution. Actually should be compared against smoothed reference prediction which is
not implemented yet in JSVM.
Claim of better error resilience. This would however only apply if not many intra_BL coded
blocks are replaced by the in scale prediction as might be the case for the optimum results
presented.
Results do not allow conclusion. Provide more results with other more realistic BL QP settings,
FGS, comparison against implemented smoothed reference prediction.
JVT decision: Continue in CE on subband technology.
JVT-U149-L [M. Mathew, B.K. Lee] CE4: Verif JVT-U095-L
This document provides verification results for JVT-U095.
150
Proponent not present.
Not presented. Says that results were verified.
JVT-U095 describes two experiments in the section “Experimental Results”.
-
Experiment 1: Improvement over a previous proposal
-
Experiment 2: Performance of Low-Delay coding
Samsung was asked to verify “Experiment 2: Performance of Low-Delay coding” configuration.
The proponents of JVT-U095 provided
encoder and decoder binaries
source code of their modified software
Configuration files for “Experiment 2: Performance of Low-Delay coding”.
The verification was done via
encoding and decoding all bit-streams using the provided binaries and configuration files
measuring the PSNR of decoded sequences
measuring the bit-rate of the generated bit-streams
The PSNR and bit-rate values of “Experiment 2: Performance of Low-Delay coding” reported in
JVT-U095 have been reproduced without any problems.
JVT-U126* [Y. Ye, Y. Bao] CE4: L-C smooth ref spat SVC
This contribution proposes a complexity reduction scheme for smoothed reference used in spatial
scalability coding (adopted from JVT-R091 from Woo-Jin Han of Samsung in Bangkok). The
current smoothed reference scheme in Joint Draft performs smoothing with (1, 2, 1) filter in both
dimensions on prediction. In this contribution, when the smoothed reference flag is turned on,
and the motion vector has fractional pixel precision, a low-complexity bilinear filter is used in the
motion compensation module to replace both the AVC fractional pixel interpolation filter and the
smoothing filter; if the motion vector has integer pixel precision, the [1, 2, 1] smoothing filtering
is carried out within motion compensation. By removing the stand-alone [1, 2, 1] smoothing
filter, and simplifying the fractional pixel interpolation filter, the system complexity is reportedly
greatly reduced. Under the CE4 testing conditions, the scheme proposed in this contribution
reportedly achieves approximately the same R-D performance. At the same time, it reportedly
significantly reduces the complexity of the existing scheme; the number of operations needed per
macroblock is reported to be reduced by 35% on average for CIF sequences and 15% for 4CIF
sequences in dyadic spatial scalability test. Smoothed reference prediction disabled for chroma.
Remark: The argument seems easier to follow in the dyadic case than in the ESS case.
Reportedly also tried [1, 4, 1] and [1, 6, 1] and confirmed that they did not improve performance
before settling on [1, 2, 1] for the proposal.
Question: Has smoothed reference been implemented for P pictures? No – let's get that done.
What happens if the old smoothed reference design works better for the P pictures?
Remark: Some aspect of encoder complexity may increase in relation to MC and filtering
decision-making.
151
Question (TW): Try adding an option for using smoothed reference prediction without using
residual prediction? Interesting question.
Remark on similarity to inter plus residual prediction case, kind of combining some old concepts
(like H.261's [1, 2, 1] switched MC filtering).
Points out that for case when both smoothed reference and sub-pel interpolation in MC are used,
two filter operations are applied. Proposes to replace by only one filter to reduce the complexity.
This would for smoothing case use subpel MC bilinear interpolation instead of 6-tap with
binomial (1 2 1) filter. Some boundary conditions omitted. Only marginal (0.01 dB range) on
luma PSNR. Also propose to not use smoothed reference pred. for chroma.
Smoothed reference for P not implemented yet.
Reduces complexity at decoder, but mode decision at encoder may become more complex?
Current design of SR prediction is always more complex at decoder when SR is on than for case
when it is off. Proposal would in particular reduce complexity in cases which are most complex
in MC (6tap horizontal and vertical) where SR on would become less complex than SR off.
Would be interesting also to look into possible gain by allowing residual switching in
combination with SR.
Implementation and verification for P frames and comparison with old method still necessary for
adoption. Later provided.
Has been implemented in P frames. Negligible gain found for long-delay configuration (P frame
far away), but proposed method performs same (both around 0.0 – 0.01 dB). This is not
surprising: Results with IPPP (or long-delay without B frames) would be necessary.
Report: Using SR gives average 0.1 dB for luma, less for chroma. No significant difference for
low-complexity scheme as compared to the original scheme.
JVT decision: Adopt. Study combination with switching residual prediction in AHG.
JVT-U122-L [S. Sun] Verif QCOM JVT-U126 smooth ref
The purpose of this report was to verify proposal JVT-U126 “Complexity Reduction for
Smoothed Reference used in Spatial Scalability” from Qualcomm. As a verification task, coding
performance checks were carried out. For further validation, please see the proposal report (it is
not clear what this statement means). Due to limited time, the results presented by Qualcomm in
JVT-U126 were reportedly only partially confirmed.
v. CE 5 & related docs: AR PR slices
JVT-U055* [S. Kamp, M. Wien] CE5: Results for JVT-U062
This contribution provides results for the local adaptation and coding of leak factor in AR-PR
slices (JVT-T062) using the testing conditions of CE5. Additionally, this document includes
results for a simplification of the original scheme.
Couples adaptation factor with mb_type. Reports bug from last meeting.
Small improvement in PSNR.
152
Large distance between 6-tap and bi-linear interpolation.
JVT-U131-M [X. Wang] Verif RWTH-Aachen JVT-U055
The purpose of this document was reportedly to verify results in JVT-U055 from RWTHAachen.
Source code was provided by the proponents of JVT-U055, for the case of “allmodes” and
“skiponly” separately. Based on the source code provided, executables were generated for
verification.
Tests were reportedly performed for the following scenarios:
– AVC interpolation for AR-PR, base-layer-qp = 30, 1 PR layer
– AVC interpolation for AR-PR, base-layer-qp = 38, 1 PR layer
– AVC interpolation for AR-PR, base-layer-qp = 38, 2 PR layer
– Bilinear interpolation for AR-PR, base-layer-qp = 30, 1 PR layer
– Bilinear interpolation for AR-PR, base-layer-qp = 38, 1 PR layer
– Bilinear interpolation for AR-PR, base-layer-qp = 38, 2 PR layer
Except Soccer sequence, identical results were reportedly obtained as those provided in JVTU055. For Soccer sequence, bitrates as well as PSNR were reportedly slightly different, which, it
was suggested, may have been due to the use of a different version of the source video test
sequence.
JVT-U074* [S. Jeong, M. Park, G. Park, K. Kim, D. Suh] CE5: Verif Aachen JVT-U055 &
T1 vs T2
The purpose of this contribution is reportedly to cross check and compare the results of Tool1
(JVT-T021) and Tool2 (JVT-T062) of CE 5 at the same system environment (Windows XP &
MS Visual C++). The performances of Tool 1 are improved up to 0.22dB for JSVM 6.1 with
CABAC and AVC filter combination, up to 0.12 dB for CABAC and bilinear filter combination.
The performances of Tool 2 are improved up to 0.22 dB for JSVM 6.1 with CABAC and AVC
filter combination, up to 0.13 dB for CABAC and bilinear filter combination. The differences of
average PSNR gains between Tool1 and Tool2 is range from -0.03dB to 0.01dB (-0.03 dB < Tool
1 – Tool 2 < +0.01 dB). Comparison results show that the main performance gains are come from
the adjustment of alpha leak factor for skip macroblocks, because two tools have almost same
performances in coding efficiency.
Stop working on this tool.
JVT-U073* [G. Park, S. Jeong, M. Park, D. Suh, K. Kim, K. Moon, J. Hong] CE5: Tool1
results JVT-U021-L
This contribution is a response to CE 5, evaluating a proposal based on JVT-T021 (Tool 1 in
JVT-T305r1). JVT-T021 proposed a leak factor overriding method in the macroblocks of SKIP
mode to improve coding efficiency of FGS coding with adaptive reference. Simulation based on
JVT-T305r1 CE5 description was reportedly carried out for all combinations of entropy methods
(CABAC/CAVLC) and interpolation filter tools (AVC/bilinear), and the results reportedly show
that the performance of the proposed method improves coding efficiency up to 0.22 dB for JSVM
6.1 with CABAC and AVC filter combination, up to 0.12 dB for CABAC and bilinear filter
combination, up to 0.29 dB for CAVLC and AVC filter combination, and up to 0.19 dB for
CAVLC and bilinear filter combination, respectively.
Remark: See notes elsewhere about inappropriate "cherry picking" of results reporting in
abstracts – average values should be included. No average results reported.
153
For one best sequence the average reported PSNR difference was 0.1 dB. Largest difference at
higher bit rates.
Adds syntax to slice header in scalable extension.
Some discussions about experiment methods and hypothetical other test conditions.
JVT-U088-L [W.-J.Han, B.-K.Lee] CE5: Verification of ETRI JVT-U073
The purpose of this report is to check the validity of results provided in JVT-U073, CE5: Tool1
results. After running encoder, extractor and decoder scripts, it is reportedly verified that there is
no problem. Additionally source-level check has reportedly been performed briefly with the
modified software provided by ETRI and KHU.
JVT-U076-L [X. Ji] CE5: Improv FGS for low-delay
This contribution reports results for CE5, which was targeted to improve the coding efficiency
for both single layer FGS coding and multiple layer FGS coding for low-delay applications. It is
claimed that the proposed cycle based FGS coding can provide higher coding efficiency than the
existing AR-FGS coding for a wide bitrate range. It is further claimed that coding efficiency
gains of up to 0.8dB can be achieved by introducing the partial-reconstructed enhancement layer
reference into the motion-compensated prediction loop of FGS layers. It is also claimed that
together with the weighting combination of different-quality enhancement layer references, more
flexible coding quality can be supported by adjusting the corresponding leaky factor to be more
suitable for varied practical application requirements.
What if you don't close the loop at the decoder. Seems more like encoder issue. More
experiments needed.
JVT decision: Continue JVT-U076 in CE.
JVT-U077-L [X. Ji] Block based FGS for low-delay
It is reported that the coding efficiency of AR-FGS is higher than that of standard FGS coding for
low-delay applications, but that this increased coding efficiency is achieved at the expense of
increased encoder and decoder complexity. In this contribution, a modified AR-FGS coding
method is proposed, and it is claimed that this coding method effectively reduces the decoder
complexity while the coding efficiency is similar to the AR-FGS in JD7. In is further claimed
that the proposed scheme is able to offer a more flexible selection between decoding complexity
& error resilience and coding efficiency.
JVT decision: Continue JVT-U076 in CE.
JVT-U096-L [J. Xu] CE5: Verif JVT-U077-L and JVT-U076-L
JVT-U078-L [L. Zhang, X. Ji, D. Zhao, W. Gao] Adapt. spatial & transform domain FGS
The contribution claims that in inter-picture coding, frequency transform is usually an efficient
method to remove the correlation among the predicted errors. However, it reportedly can not do
well if the predicted errors have low correlation. In this proposal, an adaptive prediction error
coding method in spatial and frequency domain is used for FGS coding. The initial experimental
result reportedly shows that higher coding efficiency can be achieved at low bitrates and it
154
reportedly also reduces the computation complexity since inverse transform is no longer needed
when reconstructing the predicted errors, which are encoded in spatial domain.
PR slice related.
Small gains. Not sufficient to furterh consider. contribution note.
JVT-U094-M [S. Jeong, K. Moon, J. Hong] CE5: Verif Tool 3 JVT-U076-L L-D FGS
The purpose of this report was reportedly to check the validity of the results of Tool 3 of CE 5 in
JVT-U076-L. After running encoder, extractor and decoder scripts, it was reportedly verified that
there is no problem except extraction points. However, it was reported to be difficult to directly
compare Tool 3 with other tools of CE 5.
JVT-U115* [T.C. Thang, T.M. Bae, Y.M. Ro, J.W. Kang, J.-G. Kim] AR-FGS with motion
refinement
It is claimed that AR-FGS is not working appropriately in the condition of FGS motion
refinement, specifically when residual of a FGS block is not predicted from the co-located block
in its base layer. In this contribution, a solution is proposed to handle this issue. This solution is
consistent with the basic idea of AR-FGS and does not increase the complexity of encoder and
decoder.
Proposes to disable adaptive motion refinement in AR-FGS.
Performance results? Production of results with coding efficiency improvement proposals is
encouraged. Contribution noted.
vi. CE 6 & related docs: ESS
JVT-U025* [E.Francois, V.Bottreau, J.Vieron] Modified inter-layer prediction for ESS
An alternate method to proposal JVT-T088 for dealing with mixed intra-inter base layer
macroblocks inheriting in case of non standard dyadic spatial scalability configurations is
proposed. The proposed solution consists in using intra_bl mode when a majority of the 4x4
blocks of the considered enhancement layer macroblock inherits from intra base layer
macroblocks.
Padding is proposed in those areas of EL intra MBs which overlap with BL inter MBs. Gain up to
3% bit rate reduction, average?
CE was to test JVT-T088; this proposes an alternative motivates by an interest in using a lower
complexity method than JVT-T088.
Proposes padding for regions not corresponding to intra base.
On one particular optimistic example sequence, 3% bit rate gain relative to current design (not
relative to JVT-T088 / JVT-U130) was reported, averaged for 3 bit rates, usually impact very
small (negligible PSNR, less than 1% bit rate). Overall gain? Will provide.
Average bit-rate gain: 1%.
JVT-U039-L [F. Le Leannec, P. Onno, X. Henocq] CE6: Cross-verif Thomson's JVT-U025
155
This contribution verifies document JVT-U025. The verification has been performed by decoding
the provided bitstreams with the provided binaries. The simulation results of JVT-U025 are
confirmed.
Why were binaries used rather than source code? Don't know.
Person who carried out the verification was not present. Apparently, only binary verification was
made, but source code was available to them.
JVT-U130* [X. Wang, J. Ridge] CE6: ESS Inter-layer pred
In the current JSVM, inter-layer prediction for the case of extended spatial scalability (ESS) is
done through a “virtual” base layer. For a virtual base layer MB, when it is partially intra-coded
and partially inter-coded, it is defined to be an inter MB. The practical consequence of this
approach is that reconstructed intra MBs from base layer are reportedly very often not used for
inter-layer prediction. This document proposes a change that is asserted to effectively use base
layer reconstructed intra MBs and to improve coding efficiency with little complexity overhead.
This proposal is essentially the same as JVT-T088, but with updated test results provided. The
contribution asserts that the improvement on some sequences can be rather significant, e.g. 10%
overall bit rate saving.
Note: Proponents are requested to never provide a single optimistic example of best performance
in an abstract, such as the one provided in this document as "Results show that the improvement
on some sequences can be rather significant, e.g. 10% overall bitrate saving.", without also
providing overall results across different sequences and bit rates as appropriate. We note that this
is not the only proposal with such an example of poor abstracting practice, but choose this as just
one particular example of the problem.
What are the overall results? Response: Overall using 8 sequences reportedly about 4% gain
across different bit rates and resolutions.
Propose mixture of intra prediction and residual prediction for the blocks that are affected. This
may lead to boundary effects, most probably due to the process of MV derivation. Possible
solution would be a transition with weighted averaging.
Proponents of JVT-U025 and JVT-U130 need to clarify about a) average bitrate saving and b)
complexity of both and the current design
Question: Losing some error resilience?
Potential visual artifacts discussed.
What is the real relative complexity impact? Is this method really higher complexity than current
design?
Discussion of complexity of mixed mode behaviour inside a macroblock…
Can we confine our spatial scalability needs to resampling factors of 1.5 and 2?
Revisited after some complexity consideration.
156
Remark: Note the cascading effect of intra prediction in the base layer – can we constrain
somehow the number of intra macroblocks that are required for decoding an enhancement layer
IntraBL macroblock?
Average bit-rate reduction: 3.9%
JVT decision: Adopted.
JVT-U067* [G.J. Sullivan] Position Calc for SVC Upsampling
This contribution proposes two alterations relating to the sample position calculation method for
SVC upsampling as recently adopted from JVT-S067. Both aspects are asserted to have a very
small impact on the current design, and are asserted to be minor clean-up bug fixes. The first part
proposes an alteration of the sample position computation method that is asserted to provide
approximately one to three extra bits of precision in the computations without significantly
altering the sample position computation process or its complexity (one bit of improvement for
luma positions and two or three bits for chroma). The second part discusses some considerations
relating to how the design should operate with 4:2:2 and 4:4:4 sampling structures, and proposes
to lock the luma and chroma sample position calculations together whenever the resolution of the
chroma and luma sampling grid is the same in a particular dimension.
Two basic aspects proposed for adoption:
– enhanced precision position calculation
– locking chroma and luma positions together for 4:2:2 vertical and 4:4:4 horizontal and
vertical.
Gives extra precision (on to three bits) by shuffling the arithmetic operations. Proposes to lock
luma and chroma positions together in case of same resolutions. Appears obvious improvement
on existing method.
Proposed as an obvious improvement – revisited on Tuesday for decision.
JVT decision: Adopted.
JVT decision: Create AhG on enhanced spatial scalability.
JVT-U097-L [E. Francois] CE6: Verif Nokia JVT-U130 ESS
This document reports cross-check results of proposal JVT-U130 untitled ‘CE6 report: Improved
inter-layer prediction for ESS’ from Nokia. As a verification task, coding and decoding
performance check was carried out. The results presented by Nokia in JVT-U130 are confirmed.
vii. CE 7 & related docs: Inter-layer prediction
JVT-U021-L [W. Yao, Z. G. Li, S. Rahardja] Balanced inter-layer prediction
There are two major aspects of this document, one proposes an inter-layer prediction scheme
with two base layers, and the other reportedly introduces a new concept of auxiliary layer for
combined coarse granular scalability (CGS) and spatial scalability. The former is reportedly an
extension of JVT-T053 while the latter is proposed as a supporting technology of JVT-T053.
They reportedly can be applied to obtain a good balance among all layers. The coding efficiency
of layers with higher resolution can reportedly be improved.
Considers issue of imbalance of bit allocation among different coding layers.
157
Proposes to allow two base layers for a given enhancement layer.
Introduces a concept of something called an auxiliary layer. This is just an encoder choice to add
a layer, not a proposed change of SVC design.
What is the gain of the normative aspect? How to properly measure that? Some results and
analysis seem missing. Results for top layer of enhancement structure are not provided.
Remark: Minimum number of layers to make the normative aspect provide a benefit is four, and
these must need to be mixtures of spatial and SNR scalability – this may not be the most
mainstream example use case.
Normative proposal is mode that allows combined prediction from CIF low rate and QCIF high
rate, where CIF low rate is predicted from QCIF low rate. Mixed with non-normative proposal on
“auxiliary layer”. For the combination, up to 3 dB improvement is reported for CIF low rate, but
gain for CIF high rate is not shown. Would be necessary to report results under normal CE
conditions, and also in a way that the benefit of the additional normative tool would become
obvious. With the information given, no action necessary.
JVT-U107-L [Q. Shen, Y.-K. Wang, H. Li] Adaptive inter-layer prediction
This document proposes to allow macroblock level adaptive selection of base layer for inter-layer
prediction, for coding efficiency improvement. Simulation results for coding cases with extended
spatial scalability (ESS) are provided to justify proposal. Required syntax and semantics changes,
as well as the software package including both source code and simulation scripts, are also
provided.
Relationship to document JVT-U021? Aside from ROI concept, basic idea seems the same.
Remark: A motion-constrained slice group is another way to deliver the example enhancementlayer functionality.
Example seems to be artificial, could be also resolved by using slice groups. No action necessary.
Contribution noted. No action taken.
JVT decision: Continue CE as there are multiple parties expressing an interest in such further
investigation.
JVT-U033* [K. Shimauchi] Inter-layer estimation for SVC
This contribution investigates an inter-layer estimation method using Laplacian pyramid theory
for SVC. It is asserted that estimating an enhancement layer picture from a lower layer picture is
effective for reducing inter-layer prediction error. In the estimation process, there is an issue that
the reconstructed lower layer picture includes quantization noise. Therefore, it could be
considered that the estimation method depends on quality of the reconstructed lower layer
picture. This contribution tried to incorporate the estimation method into SVC with controlling by
only QP of the lower layer. The simulation results report that the proposed method provides gain
up to 1.75 dB compared with the current SVC.
158
Nonlinear filter approach which enhances the highpass component in the up-sampled base layer
depending on QP. This could be beneficial due to the usage of decimation filters with cutoff
below half sampling frequency, provided the coding noise is not too high. For relevant BL QP
(30), no gain is observed, while the significant gain reported appears only for QP=10…20.
Defines a scaling factor s and a control threshold T, optimizes their values. s = 0 is roughly the
current design; increases of s reflect high-frequency boost.
All-intra coding experiment – reported gains are significant, but relate only in cases of very high
quality (e.g., 40-50 dB QCIF) base layer. No gain at typical base layer fidelities (e.g., QP of base
layer is 30, corresponding to 34-37 dB PSNR).
High frequency boost at high fidelity base layer, may be making up for excessive attentuation of
higher frequencies in downsampling filter?
Contribution noted; no action taken.
viii. CE 8 & related docs: Transcoding
JVT-U043* [A. Segall] CE8: SVC-to-AVC bitstream rewriting for CGS
Results from CE8: “SVC to AVC Transcoding” are provided. The CE evaluates JVT-T061 and
considers changes to the syntax and semantics of coarse grain scalable layers. These changes
enable the rewriting of an SVC bit-stream into an AVC compliant bit-stream. That is, a network
device can rewrite the SVC data into an AVC bit-stream and without needing to reconstruct the
intensity values of the sequence. Performance of the method is reported in the document.
Software is also provided.
CE not yet completed, still bug in software. Necessary to perform CGS in the transform domain
and modify intra_BL mode. Continuing CE should report separately about the effect of
transform-domain CGS, modified intra_BL mode and also disabled intra_BL mode which also
would be a solution.
Information contribution. CE work not yet completed; suggestion to continue the CE.
JVT decision: No further results, continue CE.
Changes considered for propagating transform size selection and intra mode decisions. Also
allowing transform-domain CGS. Changing Intra BL prediction mode.
Transform domain CGS possible as non-normative equivalent technique?
Remark: Consider simulcast intra operation.
Remark: Consider interference with JVT-U098.
JVT-U117-L [H. Schwarz] CE8: Verif JVT-U043 SVC-to-AVC
In JVT-U043, it is proposed to include a special mode for CGS coding in the SVC design, which
allows a direct re-writing of an SVC bit-stream to an AVC bit-stream without expensive
transcoding. In this contribution (JVT-U117), the proposed modifications are analyzed. After
carefully inspecting the provided source code, it is believed that the additional coding mode,
which is proposed in JVT-U043, enables a fast re-writing of SVC bit-streams that only use CGS
159
to AVC bit-streams. Due to a software bug, the impact on coding efficiency could only be
measured for configurations with the 8x8 transform disabled. The results show an average
increase in bit-rate of about 10%. Since the proponent of JVT-U043 did not provide a tool for the
actual bit-stream re-writing, this functionality could not be experimentally verified. Furthermore,
it was not possible to compare the coding efficiency of AVC bit-streams, which are obtained
from SVC bit-streams, with the coding efficiency of directly encoded AVC bit-streams.
Further work needed as noted.
ix. High-level syntax
JVT-U080* [B. Lee, J. Lim, M. Kim, S. Hahm, B. Kim, K. Lee, K. Park] SVC NAL unit
types for online extraction
The SVC is specified towards efficient and flexible representation of compressed bitstreams so
that it can cope with fine granular adaptation to the change of the network bandwidth or terminal
display sizes etc. Nevertheless, the contribution asserts that the current signal mechanism of SVC
NAL types does not allow for the switching of spatial scalability layers from a lower to higher
resolution for on-line extraction of SVC bitstreams. Therefore, an extension to SVC NAL types is
proposed in order to make it possible to switch between spatial scalability layers for the online
extraction of SVC bitstreams.
Remark: Contains wrong assumptions about SVC design.
JVT-U085* [A. Eleftheriadis] Clarif Nesting Temporal Levels
This contribution identifies a potential problem in the SVC syntax, with regards to temporal
nesting of temporal levels. Lack of nesting is problematic in the process of adding temporal
levels to a given bitstream, as may be performed by, e.g., a MANE. The source of the problem is
identified to be the way the pyramidal construction of temporal levels is performed, where there
is no consideration for the temporal extent of the dependency. The current JD has no mechanism
to capture this information in the VCL or SEI messages (including the Scalability Information
SEI message). The contribution proposes the introduction of a flag in the SPS that explicitly
signals if such nesting is present in the coded signal, thus greatly simplifying the operation of
MANEs and other coded-domain processing systems. The flag can also be used in profile
definitions to constrain the bitstream for specific application domains, such as low-delay, realtime communication applications.
Problem: Possible to switch down by one temporal level, but switching up may not be possible
for all cases. Only "nested" structures allow switching. Proposal to provide
temporal_level_nesting_flag. In principle, this could be derived by the decoder, but cannot be
known by the network. The flag would enable a simple mechanism to detect whether it is useful
or not to switch at this point, with the main purpose that transmission of unnecessary information
can be avoided.
JVT decision: Better to put into scalability SEI message. Reflected in revised version of
document that is Adopted.
JVT-U090-L [S.-W. Park, B.-Y. Jeon] Usage of store_base_rep_flag
This document tries to describe the usage of store_base_rep_flag in PR slices. It is reported that
store_base_rep_flag is used unnecessarily in non-key PR slices and this document proposes to
modify current syntax structure by using use_base_prediction_flag to prevent it.
Points out an inefficiency and suggests to move flag within slice header. JVT decision: Adopted.
160
JVT-U106-L [Y. Guo, Y.-K. Wang, M. M. Hannuksela, H. Li] Discardable data adaptation
This contribution proposes the support of discardable slice coding when the slice does not cover
the entire region covered by the picture. The proposal consists of two parts: some syntax changes
for all cases and a padding process for cases where upsampling is involved for spatial scalability.
Very small gains: around 0.1 dB.
Concept not very compelling.
Behaviour at slice boundaries: Intra_BL upsampling process and deblocking is handled as if the
slice boundary would be a picture boundary. JVT decision: Adopted.
Switch signalling of prediction of motion vectors, residual, Intra_BL on the macroblock layer in
slice header: mv_pred_flag, res_pred_flag, intra_bl_pred_flag. JVT decision: Adopted.
Check relationship of intra_bl_pred_flag with JVT-U098. See also section discussing JVTU160-M below.
JVT-U160-M [A. Eleftheriadis] On telescopic mode decision
This contribution addresses compression efficiency in real-time applications where telescopic
mode decision may be used. Telescopic means that enhancement layer mode decisions can be
simply derived from the base layer to avoid the computational burden of re-computing them at
the enhancement layer(s). When applied to SVC, due to a syntax limitation, telescopic mode use
results in always using residual prediction. If, in the context of today’s real-time applications, the
encoder cannot properly use it, it reportedly results in reduction in coding efficiency. The
contribution proposes the introduction of a separate flag in the slice header, dedicated to signal
use of residual prediction, separately from adaptive prediction. Experiments with the proponent’s
real-time codec on four test sequences reportedly show a savings of 12 Kbps for CIF at 30 fps. It
was further proposed that other inter prediction modes are similarly decoupled and explicitly
signaled at the slice header to turn them on and off.
No results based on JSVM (their own software and similar-design codec were used). Looks
interesting but would mean customization of syntax according to a particular type of encoder,
which is not the kind of thing ordinarily embraced here.
However, the desired functionality and syntax impact had been adopted earlier in the meeting for
other reasons. That adoption was reviewed and reconfirmed. See also discussion of JVT-U106
above.
JVT-U109-L [Y.-K. Wang, M. M. Hannuksela] On SVC high-level syntax
This proposal proposes to a couple of constraints on SVC high-level syntax (on key picture and
reference picture management) and some definitions of SVC logical entities (coded picture, layer,
etc.).
key_pics shall have nal_ref_idc > 0: sounds ok for time being – but needs further checking like
for interlace
161
Restrict MMCO to only apply to pictures with equal or a larger value of temporal_level: JVT
decision: adopt.
Express this as a constraint on how to construct temporal_level values.
Restrict RPLR such that the final list of ref pics only contains pics that have temporal_level
smaller or equal to the temporal_level of the current picture.
Express this as a constraint on how to construct temporal_level values. JVT decision: adopt.
JVT-U111-L [Y.-K. Wang, M. M. Hannuksela] SVC HRD
This document gives a first try to rectify the H.264/AVC hypothetical reference decoder (HRD)
for SVC. Reasons are listed to explain why changes to H.264/AVC HRD spec are needed. Some
specification text changes are included in the accompany document, serving as a start point. The
terms defined in JVT-U109 are applied.
Contribution noted.
In principle, HRD parameters are only needed for the inter-operability point that is to be decoded.
However, when providing an SVC bitstream, it may not be known which of the contained interoperability points will be decoded and we need a container to capture the various HRD
parameters. One question arises about the construction of HRD parameters for temporal
scalability only, as to efficiently signal cpb and dpb parameters.
Further work is strongly encouraged.
JVT-U112-M [Y. Chen, Y-. K. Wang, M. M. Hannuksela] SVC ref pic list construction
Contribution noted.
We have the intent to disallow the temporal direct mode for the base layer when an SVC
enhancement layer is present. We encourage further data on the subject. If no data are received at
the next meeting, temporal direct mode will be removed.
JVT-U116* [A. Eleftheriadis, S. Cipolli, J. Lennox] Err resil frame nums in key pics
This contribution examines the behavior of an SVC decoder in the presence of packet errors,
observing that when key pictures are lost, there is no mechanism through which the decoder can
be made aware that the reference picture list state is incorrect. Although packet loss mechanisms
can be used at the transport layer (e.g., with RTP), still there is no way to infer if a lost picture is
a key picture or not. We identify a set of bits in the new, 3-byte SVC NAL header, that are not
used under some conditions, and propose to utilize them for this purpose. Key pictures are
assigned a frame number, and non-reference pictures carry indications of what key picture they
are using.
The decoder cannot know whether the correct picture is stored e.g. in ref pic list 0. Solution:
Include numbers for key pictures, and references to these numbers in the non-key pictures.
Natural place: NAL header, where certain bits are not used under certain conditions (proposal is
to use 5 bits) Would be made conditionally.
Concerns about conditional parsing of NAL unit header, using up of reserved bits. Clarify if it
can be derived by tracking reference picture lists and MMCO commands.
162
No conclusion reached in breakout group. Frame number indication (in slice header) was already
proposed in Nice (JVT-Q091) but not adopted that time.
JVT decision: Adopt alternative solution with optional switching of one additional NAL unit
header byte (using all 8 bits for frame number index and one bit to indicate presence) for d=0,
q=0 base layer (only).
JVT-U118-L [J. Jia, H.K. Kim, H.C. Choi, J.G. Kim] Terms for SVC access unit def
In current SVC joint draft, one Access Unit (AU) includes exactly one primary coded picture
when chroma sampling is 4:0:0, 4:2:0, or 4:2:2. Enhancement coded picture belongs to coded
picture (not primary) in an AU. Considering the case that temporal scalability is used, with the
current description on AU and its structure, there would be such AU that doesn’t include any
primary coded picture, for there is no corresponding coded picture in base layer associated with
this enhancement coded picture. For this reason, a term, sub-picture, is proposed in this
contribution. With the term sub-picture, the definition of picture, primary coded picture and
redundant coded picture for AVC could be extended to SVC. A modified AU structure with the
proposed sub-picture is also described in this contribution. When chroma sampling is 4:4:4, an
access unit consists of a set of NAL units containing one or more primary coded pictures. The
corresponding definition of AU with the proposed term considering the new 4:4:4 profile is also
given in this contribution.
Considered editing issue – was discussed in a BoG – editors will take note.
x. SEI messages
JVT-U036* [P. Onno, F. Le Leannec, X. Hinocq, J. Takeda] Quality layer SEI for virtual
resolutions
This contribution proposes to extend the "quality_layers_info” SEI message to offer
rate/distortion information not only for the spatial resolution of the current layer but also for
virtual lower spatial resolutions.
For cases where lower resolution is displayed than actually decoded (after downscaling). Syntax
that defines quality layers separately for a certain number of down-sampled resolutions.
Encoding process with QP=24,27,30 base layer and same at higher res base, with 3 FGS layers
on top. Both for dyadic and extended (ratio 1.5) SS. Compare ESS and downsampled dyadic
Gain of sometimes up to 0.6 dB against ESS with 5/3 and 4/3, but significant loss for
downsampling rate 3/2.
No results given that would actually show the benefit of the proposal.
Results indicate that ESS for ratios other than 3/2 may not be most efficient. This is very
interesting, because restriction may be beneficial.
No decision to be taken here. More information needed.
JVT-U041* [A. Segall, L. Kerofsky and S. Lei] Tone Mapping SEI Message: New results
In the Klagenfurt meeting, a tone mapping SEI message was adopted by the JVT. The adoption
was conditioned on a showcase of the SEI message at the Hangzhou meeting. This document
provides the requested showcase. Software is provided as part of the document.
163
Showcase made. Software delivered. JVT decision: Adopt.
JVT-U044-L [A. Segall] Transcoding in Scalability Info SEI
It is proposed to add AVC bit-rate information to the SVC Scalability Information SEI message.
The information may be utilized by a network device to discard a portion of SVC data prior to
transcoding to an AVC bit-stream. The proposed changes are made within the context of CE8.
Parts that relate to lossless transcoding as investigated in CE8 appear beneficial, proof is
expected from ongoing CE. Part of the contribution also relates to lossy transcoding which
however cannot be normative, needs some more detailed informtion to investigate how general it
is. Would be best to adopt (when decision is made in next meeting) only parts needed for lossless
transcoding. Defer decision until CE8 results are complete.
JVT-U110* [M. M. Hannuksela, Y.-K. Wang] AVC SEI semantics in SVC context
The scalable nesting SEI message proposed in JVT-T073 was adopted in the July 2006 JVT
meeting. It carries an ordinary H.264/AVC SEI message, the semantics of which should be
amended to address the enhancement layers indicated in the scalable nesting SEI message. This
contribution proposes the semantics of some H.264/AVC messages when they are included in a
scalable nesting SEI message.
Items appear justified in general, but would need careful checking and review. Get in contact
with editors and resolve in detail.
JVT-U156-L [S. Sun, G. Sullivan] Scalable Coding Solutions Based on Various Sub
Sequence Structures
This contribution presents a few options for temporal and quality scalable video coding within
the scope of the existing AVC standard. The scalable coding features are largely based on the
concepts of sub-sequence coding and progressive refinement coding. Preliminary experiments
are reported with comparable or better performance to the current SVC JSVM software. Three
SEI messages are proposed to support the potential applications. Some of the experiment results
may indicate that the current SVC design or its reference software implementation still needs
further improvement.
Temporal scalability SEI, progressive refinement SEI, combined scalability SEI. Not capable for
spatial scalability.
This is equivalent to TS and CGS with multi-loop (while SVC is single-loop). Advantage: Would
allow to run SVC with existing AVC decoders, where however higher complexity is necessary
than with comparable SVC. Interface with network via NAL would not be as simple as with
SVC. Do we need an even simpler profile than “A” (only TS and CGS)? For this, SVC would be
less complex than this method.
Currently, the market need is not fully clear. Most probably for hardware-based devices, but
these would need higher battery power than “low cost” SVC.
JVT decision: Encourage further study on
- technical approach and related SEI messages
- market need for simple (TS+CGS) scalability using possible existing devices
- and if yes, possible creation of simple scalable profile (below A) which would be less
complex than this approch, but would need new devices.
164
xi. De-blocking filter
JVT-U020* [J. He , Y. Yan, Y. Prieto] Disabling SVC chroma deblocking
This proposal suggests the ability to turn off chroma deblocking filter in SVC so that only luma
deblocking can be enabled in order to reduce computational complexity. Two ways are proposed
for modifying the SVC bitstream to enable this capability. One is to extend the existing
semantics, and the other is to change the existing syntax. It is reported that disabling chroma
deblocking normally does not cause noticeable visual quality degradation for most of the video
applications while saving significant amount of data traffic and computation.
Claims that disabling chroma deblocking has no impact on quality but good reduction in
complexity.
JVT-U032-L [Z. Lu, J. Zheng, W. Lin, S. Rahardja] Percept. Deblock Filter for ROI SVC
The Region-of-interest (ROI) based video coding within the SVC framework can be implemented
by making use of Type 2 Flexible Macroblock Ordering (FMO), which marks independent
rectangle regions/slices inside a frame by their top-left and bottom-right coordinates. By
employing the proposed scheme, more bits can be allocated to one or more ROIs in a frame,
which then ensures achievement of high coding quality, to guarantee a high coding quality or to
fulfill some special functionality. Owning to the fact that the frame is separated into independent
regions, and the regions can be assigned with different SNR, spatial and temporal quality, it is
assumed that there can be visible blockiness around the ROI boundaries. It is argued that this
kind of blockiness cannot be automatically removed by the in-loop filters. A new perceptual
deblocking filter is proposed. The filter includes two steps: first, the complexity of the blocks
around ROI boundaries is measured; and different filtering modes are then selected accordingly
to reduce the effect of false edges. Experimental results are reported to show that coding quality
is improved by the proposed filter at low bitrate video on CGS SNR scalability conditions. It is
also argued that the current SVC reference software can not decode correctly when a type 2 slice
group is missing, so the experimental results on other quality conditions are not available yet.
This is a non-normative proposal to SVC.
Non-normative late document; deferred for potential consideration at next meeting if resubmitted.
xii. Error resilience
JVT-U023* [D.T. Nguyen, J. Ostermann] Error concealment in the NAL
This contribution presents an error concealment method applied to the Network Abstraction
Layer (NAL) for SVC. The method detects the loss of NAL units for each group of picture
(GOP) and arranges a valid set of NAL units from the available NAL units. For cases where there
are more than one possibility to arrange a valid set of NAL units, this method uses the
information about motion vectors of the preceding pictures to decide if the erroneous GOP will
be shown with a higher frame rate or a higher spatial resolution. This method works without
parsing of the NAL unit payload or using of estimation and interpolation to create the lost
pictures. Therefore, it requires very low computing time and power. This proposed error
concealment method works under the condition that the NAL units of the key pictures, which are
the prediction reference pictures for other pictures in a GOP, are not lost. The proposed method is
reported to be suitable for real-time video streaming.
Presenter not available Tue 19:00, or Wed 14:30 – later revisited.
Proposes decoder-side non-normative error concealment behaviour.
165
Consider in the future as non-normative feature candidate for inclusion in reference software.
xiii. Applications and profiles
All those documents considered in joint meeting with MPEG Requirements Wed 4-6pm.
JVT-U098* [V. Bottreau] SVC MB layer for EI slices
Focuses on high resolution with high image quality. Primarily a profile proposal for a reducedcomplexity Intra-only profile.
Disables intra prediction modes in EI slices. R-D penalty reported to be small.
It is proposed to modify the macroblock layer in scalable extension syntax for EI slices in a way
that disables AVC intra prediction modes and limits the number of allowed prediction modes to
the single case of inter-layer texture prediction. The advantage is mainly seen in a reduction of
the required encoder/decoder complexity. With that modification it is also possible to omit the
transmission of the syntax element base_mode_flag for each macroblock of EI slices. The coding
efficiency loss that results from the proposed modification has been analyzed for a wide range of
configurations for CGS, dyadic and ESS spatial scalability. The average measured rate increase is
asserted to be less than 0.6 % for CIF sequences and less than 0.8 % for 4CIF sequences.
Rather belongs into profile discussion. Proposes intra-only SVC profile which then would
disallow intra prediction modes for EI slices and not send base mode flag.
Results indicate almost same bit rate on average, reducing complexity.
Suggests that allowing intra prediction modes in enhancement layer is not necessary whenever
the base layer is available for spatial or SNR CGS prediction.
Interesting. What about coding text and graphics?
Consider interference with JVT-U043.
Remark: Profile decisions should properly involve a very large test set. Detailed analysis,
particularly for aspects that affect coding efficiency, should be provided. Our common
conditions may not be sufficient for the basis of profiling decisions. Desire to have more test
data, visual testing, … For further discussion on reflector.
JVT-U049* [Y. Gao, Y. Wu] Apps & Reqs for color bit depth SVC
This document describes the requirement for color bit depth scalability and possible applications
that can benefit from color bit depth scalable coding solution. Thomson is proposing this
requirement to make SVC standards keep up with the development of handling color information
surpassing 8-bit color in each piece of the digital imaging pipeline. Close applications include 10bit DVD authoring and digital workflows in motion picture making.
Primary interest is spatial scalability. Syntax seems to already exist in current draft.
The group has an interest in support of bit depth scalability.
Chroma format scalability is already supported in the resampling equations in the draft.
166
Remark: Isn't there a restriction in the draft to constrain the enhancement chroma format to be
equal to the base chroma format?
JVT decision: Create an AHG to work on bit depth and chroma format scalability (A. Segall,
chair). Mandate: Find/create test material, define some experiments, investigate software and text
modification needs, identify complexity issues, applications.
JVT-U070* [I. Amonou, N. Cammas, S. Kervadec, S. Pateux] On SVC performance and
profiles
This contribution proposes the definition of Baseline and High profiles for SVC. These profiles
are a direct extension of AVC profiles, taking into account the additional SVC tools. It proposes
some generic recommendations to design any SVC profiles and introduces the notion of
scalability class of a profile which indicates the kind of scalability allowed for a stream
conforming to that profile and that scalability class. These recommendations aim to ease the
design and understanding of SVC profiles. In a second part, it presents a discussion on SVC
performance on some scenarios. Experiments have been performed on some scenarios inspired
from real application needs. Results do not always show efficient performances compared with
an AVC single layer or even with an AVC simulcast. Finally, it therefore recommends
considering the re-introduction of some tools that have previously been rejected from SVC in the
draft such as multiple loop decoding for example.
JVT-U086-M [A. Eleftheriadis] Prop SVC profile for videoconf
This contribution proposes the creation of an SVC profile that is reported to be suitable for
videoconferencing applications. The objective is reportedly to identify SVC features that are
relevant for real-time, interactive communication, while at the same time ensure that complexity
is kept within reasonable levels.
JVT-U127-L [J. Ridge] Mobile profile for SVC
This document summarises Nokia’s position regarding a “mobile” profile for SVC, which could
serve as a basis for discussion at the current meeting.
JVT-U137* [B. Haskell] Simple SVC profile
In some important applications for scalable video coding, a very simple decoder implementation
is required for the lowest layer. In this contribution some parameters are suggested that such a
profile might specify.
Proponent can bring further input to next meeting.
JVT-U158-M [P. Topiwala] Requirements for HD/SD SVC
Proposal for SVC profiles:
- base layer only baseline or high
- Profile A for Videoconferencing & Mobiles with CGS/MGS restriction on spatial
scalability to 1.5 and 2, baseline
- Profile B for Internet TV, broadcast etc. with SS anything between 1 and 2, baseline or
high
- Variant on B with Intra-only
- Profile C for Mobile, baseline and FGS (incl. AR-FGS, no SS)
167
Restrictions could be put on the base layer (e.g. switching off tools)
Relationship between levels and layers to be clarified, levels to be aligned with existing AVC
levels, signalling of nested levels to be allowed
Discussion about overlap between A and C in terms of applications.
Higher bit-depth in SVC? Maybe later.
Elaborated detailed tables for each profile, included tools, levels, restrictions at base layer etc.
-
A Layered Media (Alex Eleftheriades), FT (Stéphane Pateux)
B-inter FT (Stéphane Pateux), Thomson (Jérome Vieron)
B-intra Thomson (Vincent Bottreau, Jérome Vieron)
C Nokia (Justin Ridge)
Revisited: 4 PM Thu.
Verification tests will need to be performed around each of these profiles.
Proposed profile structures diagrammed in table below.
AVC base layer
Orange (JVTU070)
(mobile)
baseline
Orange
(JVT-U070)
(TVoverIP)
restricted
baseline,
main, high
yes
LayeredMedia
(JVT-U86)
Nokia (JVTU127)
Apple (JVTU137)
baseline
restricted
baseline
restricted
baseline
yes
yes
yes, limited to
2enh layer
no
no (maybe)
no
yes
no
yes
yes, if
restricted to
full block
no
yes
no
yes
yes
yes
no
yes
yes
no
no
slice groups
limited to 8
I,P
no
no
Dyadic Spatial
scalability
ESS
CGS SNR
scalability
FGS SNR
scalability (PR
slices)
AR-FGS
CAVLC
entropy coding
CABAC
entropy coding
interlaced
ASO, FMO
yes
no
no
yes
no
slices
redundant
pictures
loop filter
smoothed ref
I,P
I,P,B
yes
yes
yes
yes, with
MGS
no
yes
yes
I,P
no
yes
no
Agreed draft SVC profiles are diagrammed in the figure and table below.
168
no
no
SVC Profiles
Spatial scalability
Spatial scalability
(dyadic, 3/2)
(arbitrary up to 2)
Coarse-grain scalability Coarse-grain scalability
SVC A
Baseline
No spatial scalability
No CABAC
Fine-grain scalability
SVC B
intra
SVC B
High
SVC C
Baseline
Existing
SVC Profiles tools table
SVC tools activation in potential SVC Profiles:
– Profile A
– Profile B
– Profile B intra only
– Profile C
Legend:
– *: activation of the tool is subjected to levels definition
– []: needs further discussions
– (): needs further studies
AVC base layer
(dependency_id equal to
0 and quality_level equal
to 0) Profile
Impacting AVC base
layer tools
SVC tools
Profile A
Profile
B
Baseline
High
Profile
B
Intra
Only
High
[c_set0_flag and
c_set1_flag
equal to 1
except slice
group map type]
slice_type
deblocking filter
constrained_intra_pred_flag in
base layer
num_slice_groups > 1
slice_group_map_type
direct_spatial_mv_pred_flag
num temporal levels
slices
smoothed ref inter pred
PR slice motion refinement
AR-PR slices
169
Profile C
Baseline
[c_set0_flag and
c_set1_flag
equal to 1
except slice
group map type]
I, P
Y
1
I, P, B
Y
1
I
n/a
1
I, P, [B], PR
Y
1
[Y*]
[2*]
n/a
[N]
I, P, [B*],
EI, EP,
[EB*]
N/[Y]
N
N
N
n/a
1
[N]
I, P, B,
EI, EP,
EB
Y
N
N
N
n/a
n/a
n/a
I, EI
N
n/a
[?]
1
I, P, [B], PR
Y
N
N
N/[Y]
Y
Y
fgs_coding_mode
interlace
CAVLC
CABAC
deblocking filter
deblocking filter (upsampling)
constrained_intra_pred_flag
arbitrary slice order
slice_group_map_type
resolution factors 2, 1.5
ESS (any factor)
ESS aligned crop window
ESS non-aligned crop window
EIDR
IROI
fragmented PR slice
CGS with varying quality levels
(MGS)
weighted prediction
use_base_representation_flag
direct_spatial_mv_pred_flag
adaptive transform block size
quant scaling matrices
num temporal levels
num dependency id
max num decoded dependency id
(using inter-layer prediction)
num quality levels
Open issues
N
N/Y*
Y
Y*
Y
Y
1
N
[2*]
Y
N
Y
N
Y
N
N
Y
N
Y
Y
Y
Y
Y
1
N
N
Y
Y
Y
Y
Y
N
N
Y
N
Y
Y
Y
n/a
Y
1
N
N
Y
Y
Y
Y
n/a
N
N
Y
N/[Y*]
Y
(1)
Y*
Y*
[N]
8
3
Y
Y
(1)
Y
Y
[N]
8
3
Y
Y
n/a
Y
Y
[N]
8
3
4
4
4
N/[Y]
N
Y
N
Y
n/a
N
N
[N]
[N]
Y
N
1
color_bit_depth, color format
Notes:
– Naming profile: may be good not to reuse AVC Profile's name, to avoid confusion.
– A: simple/?
– B: advanced/?
– dependency_id: spatial enhancement or temporal enhancement only (no SNR enhancement)
– SNR scalability only when quality level increases.
Questions:
– Level definition: need to clarify nested levels?
– How to define the cost for decoding a MB of an upper layer? Cost of decoding of a MB of an
upper
– layer = cost of decoding the MB + function of (inter-layer prediction, crop window, RF,
deblocking
– filter, number of MBs used for inter-layer prediction)
xiv. Other
JVT-U119* [Y. Bandoh, S. Takamura, K. Kamikura, Y. Yashima] Sep luma/chroma comp.
in SVC
It is claimed that it is useful to separate luma component and chroma components in order to
reduce bit-rate
with subjective quality of reconstruct images maintaining. SVC supports separation of luma
component and chroma components partly. However, it is claimed that the syntax holds the
possibility of drift error in the decoding process. In this contribution, it is proposed to modify the
170
syntax concerning the separation of luma component and chroma component, in order to avoid
the drift error. We also investigates the extension of the separation, which enables the separation
of luma components and chroma components in all enhancement layers.
Separation of luma and chroma, to be signaled by a flag; purpose is error resilience.
Two proposals:
Removal of DeltaQP may lead to wrong QP value: move DeltaQP to the beginning of the first
cycle.
(non-normative) get rid of chroma faster than in our current stream.
Relevant outcome is documented in JVT-U125.
JVT-U133-M [S.-T. Hsiang] Intra subband/wavelet framework
Base layer picture: Wavelet down-sampled was sharper with much more aliasing.
Further presentation Wed a.m. Subjective viewing of Bus and Foreman QCIF base layer intra
coded with observations consistent with above remarks.
Contribution does not assert that the wavelet-based scheme is superior overall as a design.
Understood that some artifacts appear in this scheme.
Consider the intra-only case as a special case? Such filters do not seem to work for interframe
prediction.
Remark: We could use higher cutoff downsampling filters within the current JSVM design
scheme. Replying remark that reported gains seem to go away with downsampling filters of
JSVM in use.
JVT decision: Further study in CE.
l. JVT SVC non-normative modifications
i. Encoder / extractor optimization
JVT-U081* [J. Lim, P. Chen, B. Lee, M. Kim, S. Hahm, B. Kim, K. Lee, K. Park] Optimal
SVC bitstream extraction
This contribution proposes an informative method for selecting scalability levels. Depending on
the given network bandwidth, the spatial, temporal and quality scalability levels can be
controlled. This selection problem given the constraint is interpreted as an optimization problem.
An approach motivated by the proponent as optimization-theoretic for selecting the optimal
scalability levels as a non-normative method.
Works in combination with JVT-U080 in more modes. Does not require JVT-U080 for
functioning of bitstream extraction method.
171
Suggested to adopt as non-normative recommendation. Remark: Subjective? How to quantify
performance of method (how to weight subjective difference between frame rate, quality, etc.)?
How could we create a good method to assess performance of such a scheme. We may need
assistance, e.g., from MPEG Test group (suggestion to use ITU MOS scoring methods) to
construct viewing experiments to find out in what way the method may provide better results
than the extractor of the JSVM. The extractor can already be run in a variety of ways. The
contribution raises issues that may be useful for further investigation. Further study encouraged.
ii. JVT SVC informative contributions
JVT-U124-L [S. Kamp, M. Wien] Low-delay leaky base layer
This information document discusses results for quality scalability using leaky base layer
prediction for low-delay IPPP coding with PR slices in SVC. The temporal prediction reference
for the base layer is generated by calculating the weighted average of the quality base layer and
quality enhancement layer reference frames. This is reported to provide performance gains at the
enhancement layer rate point while introducing drift into the base layer if the enhancement layer
is truncated. Approaches using global weighting and locally adaptive weighting have been
investigated. Although the presented method requires modifications to the SVC decoder, the
resulting base layer bitstream is still AVC compliant.
JVT-U139-M [P. Amon, T. Rathgen, D. Singer] SVC file format
This is an informative contribution consisting of a pre-publication draft for an invited journal
publication within a special issue on Scalable Video Coding. This draft is preliminary and
therefore potentially subject to change prior to its actual publication.
The paper is on the subject of support for SVC in the ISO media file format specification.
JVT-U140-M [M. Wien, R. Cazoulat, A. Graffunder, A. Hutter, P. Amon] R-T SVC
streaming syst
This was an informative contribution consisting of a pre-publication draft for an invited journal
publication within a special issue on Scalable Video Coding. This draft is preliminary and
therefore potentially subject to change prior to its actual publication.
The paper presents the integration of SVC into a generic platform for multimedia adaptation. The
platform provides a full MPEG-21 chain including server, adaptation nodes, and clients. An
efficient adaptation framework using SVC and MPEG-21 Digital Item Adaptation (DIA) is
integrated and it is shown that SVC can seamlessly be adapted using DIA. For protection of
packet losses in an error prone environment an unequal erasure protection scheme for SVC is
provided. The platform includes a real-time SVC encoder capable of encoding CIF video with a
QCIF base layer and fine grain scalable quality refinement at 12.5 fps on off-the-shelf high-end
PCs. The reported quality degradation due to the optimization of the encoding algorithm is below
0.6 dB for the tested sequences.
JVT-U141-M [M. Wien, H. Schwarz, T. Oelbaum] SVC performance analysis
This was an informative contribution consisting of a pre-publication draft for an invited journal
publication within a special issue on Scalable Video Coding. This draft is preliminary and
therefore potentially subject to change prior to its actual publication.
172
This paper provides a performance analysis of the emerging scalable extension of ITU-T H.264 |
MPEG-4 AVC. A short overview presenting the main functionalities of SVC is given and main
issues in encoder control and bit stream extraction are outlined. Some aspects of rate-distortion
optimization in the context of SVC are discussed and strategies for derivation of optimized
configurations relative to the investigated scalability scenarios are presented. Based on these
methods rate-distortion results for SVC especially for spatial, quality and combined spa-tial and
quality scalability are presented and compared to rate-distortion optimized H.264 | AVC single
layer coding. For reference a comparison to rate-distortion optimized MPEG-4 Visual (Advanced
Simple Profile) coding results is provided. The results reportedly show that the gap between
single layer coding and scalable video coding can be very small and that SVC clearly
outperforms previous single layer video coding technology such as MPEG-4 ASP.
JVT-U144-L [I. Amonou, N. Cammas, S. Kervadec, S. Pateux] R-D extract quality layers
SVC
This was an informative contribution consisting of a pre-publication draft for an invited journal
publication within a special issue on Scalable Video Coding. This draft is preliminary and
therefore potentially subject to change prior to its actual publication.
The subject of the contribution was the topic of quality layers and their rate-distortion
performance optimzation.
JVT-U145-L [H. Schwarz, D. Marpe, T. Wiegand] SVC overview
This is an informative contribution consisting of a pre-publication draft for an invited journal
publication within a special issue on Scalable Video Coding. This draft is preliminary and
therefore potentially subject to change prior to its actual publication.
The paper presents an overview of the draft SVC design currently under standardization in the
JVT.
Partial presentation of spatial scalability aspects. For resampling ratios greater than 2, the base
layer is reportedly not used effectively. Also don't want a base layer that is "too good", as noted
above.
JVT-U146-L [E. Francois, J. Vieron, V. Bottreau] Interlaced coding in SVC
This is an informative contribution consisting of a pre-publication draft for an invited journal
publication within a special issue on Scalable Video Coding. This draft is preliminary and
therefore potentially subject to change prior to its actual publication.
This paper presents the basic concepts for supporting interlaced coding in SVC. The
generalizations of AVC interlaced tools and of SVC FGS SNR scalability are first described.
Then main issues related to interlaced video scalable encoding are identified and the new
mechanisms introduced in the SVC specification for raising these issues are presented. The paper
also discusses related applications side and identifies several use cases illustrating the interest of
interlaced support in SVC.
JVT-U150-L [J. Xu] 3D wavelet SVC coding scheme
This is an informative contribution consisting of a pre-publication draft for an invited journal
publication within a special issue on Scalable Video Coding. This draft is preliminary and
therefore potentially subject to change prior to its actual publication.
173
This paper first overviews the Barbell lifting coding scheme, which is adopted as the common
software by MPEG ad-hoc group on further exploration in wavelet video coding. The core
techniques used in the scheme, such as Barbell lifting, layered motion coding, 3D entropy coding
and base layer embedding, are discussed in detail. At the same time, this paper analyzes and
compares the proposed Barbell lifting coding scheme with the oncoming H.264/MPEG-4 SVC
(scalable video coding) standard because the temporal prediction technique used in
H.264/MPEG-4 SVC is also developed from motion compensated temporal lifting. The
commonalities and differences between these two schemes are exhibited for audience to better
understand modern scalable video coding technologies. There are still several challenges on
scalable video coding, e.g. coding performance of spatial scalability and accurate motion
compensated lifting. Two new techniques are also presented in this paper although they are not
integrated into the common software yet. Finally, experimental results demonstrate the
performance of the proposed Barbell lifting scheme and comparisons with H.264/MPEG-4 SVC
and MC-EZBC that is another famous 3D wavelet-based coding scheme.
JVT-U151-M [Y.-K. Wang, M.M. Hannuksela, S. Pateux, A. Eleftheriadis] SVC System &
Transport Interface
This is an informative contribution consisting of a pre-publication draft for an invited journal
publication within a special issue on Scalable Video Coding. This draft is preliminary and
therefore potentially subject to change prior to its actual publication.
Scalability in video coding and transmission has been a desire for many years, to meet the
requirements of heterogeneous receiving devices connected to varying bandwidths with a single
bitstream. Although earlier trials of scalable coding in standards like H.263 and MPEG-4 Visual
have not been commercially successful, the Joint Video Team has recently devoted most of its
effort for the development of a new scalable video coding standard, known as SVC, which will
become an extension to H.264/AVC. While it is certainly important to develop coding tools for
high coding efficiency, the design of the features interfacing system and transport is also of vital
importance for SVC applications. Indeed, this interface is the mechanism through which system
designs can take advantage of the scalability features of the coded video signal. This paper gives
an overview of such interfacing features as they are currently specified in the SVC specification,
including bitstream structure, extended network abstraction layer (NAL) unit header, suffix NAL
unit, scalability information related supplemental enhancement information (SEI) messages, nonrequired picture SEI message, quality layer SEI message, reference picture marking process, and
efficient layer switching support.
JVT-U152-M [S. Wenger, Y.-K. Wang, T. Schierl] SVC in IP networks
This is an informative contribution consisting of a pre-publication draft for an invited journal
publication within a special issue on Scalable Video Coding. This draft is preliminary and
therefore potentially subject to change prior to its actual publication.
The transport of scalable media, and in particular of scalable video conforming to the
forthcoming Scalable Video Coding (SVC) technology, presents challenges not only in the video
compression technology, but also in the transport layer and signaling. In this paper, we discuss
the current status of standardization of the support for scalable media, and SVC in particular,
over IP based networks. Both the transport of SVC over the Real-time Transport Protocol (RTP),
and the signaling support – namely the additional mechanisms in the Session Description
Protocol (SDP) – are covered. As it turns out, the support of SVC over RTP is not quite as
straightforward as that of non-scalable video bitstreams. Specifically, the signaling architecture
174
requires an almost complete overhaul, and new protocol mechanisms need to be introduced into
the packetization.
m. JVT 4:4:4 and
modifications
professional
applications
coding
normative
Draft JVT-T204 was produced from the last meeting. Revision JVT-U022 and JVT-U136
submitted to revise it. JVT-U120 proposes variants of High 10 and High 4:2:2. JVT-U066
proposes a "simple intra" profile. JVT-U143-M discusses level parameters. Two liaison letters
from SMPTE (JVT-U018 and JVT-U019) on High 10 and Intra profiles.
Consider MinCR for new (and prior profiles), constraints on first picture in bitstream. Do not
apply to High 10, High 4:2:2 and High 4:4:4 predictive; do not apply to intra-only profiles. Do
apply to High profile. (Note in JVT-U210.)
Agreed to have some constraint in new profiles on maximum slice size (only for Intra?).
Limit: 1/4 of max picture size supported in levels 3.1 and higher imposed only for coded picture
sizes larger than 720x576. Applies to High 10 Intra, High 4:2:2 Intra, High 4:4:4 Intra, and High
4:4:4 Predictive. JVT decision: Agreed.
Try to apply this to High 4:2:2 and High 10 as a corrigendum change? JVT decision: Agreed.
(Note in JVT-U210.)
There is some interest in further potential refinement of the "decoder friendliness" restriction,
possibly as non-normative content. – Work on this in AHG.
In all-Intra profiles, all pictures shall be IDR pictures. JVT decision: Agreed.
Professional Profiles proposed are summarized as follows:
- 4:2:0 Intra 8b as subset of High
- 4:2:0 Intra 10b as subset of High 10
- 4:2:2 Intra 10b as subset of High 4:2:2
- 4:4:4 Intra 14b as superset of the other intra
- 4:4:4 predictive 14b as superset of the high profiles
- 4:4:4 Intra 12b without CABAC (not compatible with previous ones)
The latter breaks the “onion” structure – is it really needed? In which application is CABAC a
problem? Remark: Could be difficult on Laptops (evidence to be brought about this point, defer
decision)
Consensus on option to make de-blocking filter not mandatory – agreed.
Currently, do not define the non-CABAC profile and the 4:2:0 8-bit Intra profile. Agreed.
However, provide a syntax "hook" for the 4:2:0 8-bit profile to potentially be defined in the
future.
Also can use a level constraint trick to allow indication of a lower bit rate conformance point esp.
for predictive profiles. This can be done now or in the future since it is only a constraint trick.
Two constraint set flags would suffice to accomplish these. See additional notes elsewhere.
175
The new profile structure is shown in the figure below.
“Professional” Profiles (update)
High 4:4:4 Intra
(14b)
High 4:4:4 Predictive
(14b)
High 4:2:2 Intra
(10b)
High 4:2:2
(Predictive 10b)
High 10 Intra
(4:2:0 10b)
High 10
(4:2:0 Predictive 10b)
Existing
Notes:
Arrows denote capability subset hierarchy.
Four profiles not shown: Baseline, Extended, Main, High.
Exact syntax for profile and level indicators and bit rate scale factors for new profiles is left to
editors and JVT management team at this point – for review and confirmation at next meeting.
JVT decision: Agreed.
JVT-U022* [H. Yu, G. Sullivan] Proposed 4:4:4 draft changes
This contribution summarizes the problems that have been found in JVT-T204 “Draft Text of
H.264/AVC Amendment 2 to 2005 Edition”, and provides an updated version of the draft text
with the proposed changes.
JVT decision: Adopted. Further editorial improvement is needed.
Topics identified as open issues in the contribution:
– The title of the amendment or the names of the two new profiles may need modification to
better match each other. Documented elsewhere in report..
– It appears that decoder conformance to the High 4:4:4 Intra profile requires the decoding of
bitstreams of other profiles that use inter-picture prediction. This seems ill-advised. JVT
decision: Agreed.
– Do not require decoding of bitstreams of prior profiles? Left to editors and JVT
management and review of outcome at next meeting. JVT decision: Agreed.
– Make some new all-Intra profiles with lower bit depths, chroma formats, and bit
rates? Documented elsewhere in report.
– Many of the constraints expressed in clause 8 appear to be misplaced, as clause 7 is a more
appropriate place for specifying such syntax constraints. Many of them also appear
redundant, and undesirably so, as this may confuse the reader into wondering whether
something extra or different is being specified. Editorial – editors to fix (among any other
remaining editorial problems). JVT decision: Agreed.
176
Also:
– Require (and infer) max_dec_frame_buffering = 0 for Intra-only profile JVT decision:
Agreed.
– Should we require all pictures to be IDR pictures? (If we do, then all frame_num will be
equal to 0 and all PicOrderCnt will be equal to 0 and all coded video sequences will contain
only a single picture). JVT decision: Agreed all IDR..
Remark: Constrain slice size for parallel-decoding friendliness (although parallelism across
pictures is already feasible)? Any other such decoder friendliness constraints? Documented
elsewhere in report.
JVT-U136-L [S. Sekiguchi] Prop changes to 4:4:4 draft
This contribution is to provide proposed changes to the 4:4:4 FPDAM issued at the last meeting
as the result of our study regarding open issues.
JVT decision: Accepted as input to editing process.
JVT-U099-L [S. Sekiguchi, Y. Yamada, K. Asai] Advanced 4:4:4 profiles
This contribution describes a position and specific proposals on advanced 4:4:4 profiles.
– Supports spirit of JVT-U120, adding Intra-only variants of High 10 and High 4:2:2
– Establish a conformance point without deblocking
– Request to minimize number of profiles and use onion-shell structure to the extent feasible.
On deblocking, write spec such that deblocking control information is advisory (like SEI or
VUI), but conformance is measured prior to deblocking. Whatever post-processing is done after
that is discretionary. The specified method becomes an advisory example only. JVT decision:
Agreed.
Considering the potential need for extensibility, e.g., to define a future subset profile, should we
define a constraint set flag that indicates conformance to a 4:4:4 profile and require decoders to
decode whenever that flag is 1 even if profile_idc is something else (as we have for Baseline,
Main, and Extended profiles)? Note that the same flag could be used for constraining the existing
profiles. documented elsewhere in report.
To be reviewed in joint meeting with MPEG Requirements Wed 16-18
For naming of new profiles, "High 4:4:4 Intra", "High 4:4:4 Predictive", etc.
n. JVT 4:4:4 coding non-normative modifications
o. JVT CE9: Error resilience
This topic was postponed until at least Sunday due to travel difficulties for a key participant.
JVT-U057* [S. Rane, P. Baccichet, B. Girod] CE9: On error prot redundant slices
This document details the results of a core experiment originally instituted for the proposal JVTS025 concerning the use of redundant slices in conjunction with Reed-Solomon codes. According
to the recommendations of JVT-T309, this CE evaluates the performance of this lossy error
177
protection scheme using LA-RDO as an anchor scheme. A Systematic Lossy Error Protection
(SLEP) scheme is considered, that adaptively selects the bit rate of the redundant slices, and the
channel coding rate to ensure error protection at the given packet loss percentage. When
compared to LA-RDO, SLEP provides an average PSNR improvement of 0.6 dB to 3 dB across
all the sequences considered, and results in a significant reduction in instantaneous PSNR
fluctuations caused by packet loss.
Shows good improvements compared to LA-RDO but not compared to FEC. FEC is just
applying parity to the transmitted pictures. Can be done with IETF RFC 2733.
Results are not suitable for low-delay applications as the parity is computed over multiple
pictures.
Comment that LA-RDO may contain bugs and may not represent the actual performance.
Comment: Do we need this technique?
Question on complexity of decoding the redundant slices: create all packets at lower bit-rate
except the missing one and do the inverse parity operation.
Test against intra refresh: Yes.
What needs to be specified for inter-operability:
– Mapping of the NAL unit payloads to NAL unit payloads with smaller size and generation of
bitstream (with constraints on the resulting NAL units payloads wrt conformance)
– Generation of parity information
– Bitstream syntax for parity information
– Inverse parity operation
– Specify replacement of missing NAL unit with reconstructed NAL unit
Approximated size of text description: 30 pages. No text currently available.
No one in the group indicated that in the foreseeable near-term future they would use this
technique. Currently there not support for adopting this.
JVT-U113-M [Y. Guo, Y.-K. Wang, H. Li] CE9: Verif JVT-U057 redund slices
This document reports verification results for JVT-U057 (Progress Report on CE9: Systematic
Lossy Error Protection using H.264/AVC Redundant Slices). The consistency between the
algorithm and the source code provided by the proponents was reported to have been checked
and confirmed. The simulation results of the verification were reported to have been performed
by encoding the original sequences using the binary and configuration files provided by the
proponents. A subset of the tests reported in JVT-U057 were verified and confirmed.
Verified consistency of description and ran sub-set of results.
JVT-U075* [D.Y. Suh, G.H. Park, J. Oh, M. Park] CE9: JVT-S028 extension redundant pic
(withdrawn)
This contribution is a progress report on CE6 JVT-T028 which proposed a method to recover lost
(primary) coded pictures in the client. The method is an extension of the redundant coded picture
adopted in H.264/AVC and H.264/SVC. While the previous redundant picture (or RP) is used to
178
protect picture in 1:1 redundancy, this contribution enables to recover one lost picture out of
multiple pictures in n:1 redundancy by using the same amount of redundancy. One redundant
coded picture is generated by performing XOR operation on multiple coded slices of selected
layer.
<<withdrawn>>
JVT-U114-M [C. Zhu, Y.-K. Wang, H. Li] Adaptive redundant picture coding
Information about adaptive transmission of redundant pictures.
Indicates large gains compared to LA-RDO.
But JVT-U114 and JVT-U057 are not comparable. Experimental conditions were significantly
different. Contribution noted.
p. JVT SEI message issues
JVT-U035* [S. Wittmann, T. Wedi] Post-filter hint SEI
More results with the SEI message containing post-filter hints (JVT-T039) are presented in this
contribution. The idea is to transmit filter coefficients of a filter designed on encoder-side or
cross-correlations between the encoded and the original signal to the decoder where this
additional information is used to design a post-filter. One exemplary post-filter can be a Wiener
filter that minimizes the mean-square error between an input signal and a corrupted signal (e.g.
by coding errors). Coding results are reported for sequences with 4:2:0, 4:2:2 and 4:4:4 color
sampling. Furthermore bit-rate reductions at specific PSNR points are listed. Bit-rate reductions
of 8.5% are reported in average for the tested sequences.
Tested with 2-D 5x5 nonseparable filter. 0.2 to 0.5 dB gain shown with deblocking filter on.
720p sequences only – Bigships, City, Crew, Harbour.
Average 0.4 dB or 8.5% bit rate reduction reported averaged over all points measured and all
color components.
For smaller sequences, overhead of filter representation gets rather high.
Previous meeting JVT-T039 showed separable filter. Also idea in JVT-S030.
Is the filter position-dependent? No. Did you try one that was? No. Suggestion: Perhaps it
would be possible for the decoder to derive position-dependent processing even though the
information provided was not position dependent.
Syntax supports both separable and non-separable. Also supports sending correlation
information.
Sharp verbally reported that they liked the idea and had partially confirmed its results.
Have they tried smaller kernels? Yes, at some point, but coding efficiency gain was not as high.
Syntax supports any size filter. Some reasonable limit should be imposed.
Showcase? Decoder source code and encoder binary provided in contribution. Difficult to show
visually here due to high resolution display and viewing requirements.
179
Visual? (Sharp looked only at PSNR.) Asserted that overall image looks better.
Perhaps the 4:4:4 amendment would be too soon, given the lack of time to finalize the text at
least for ITU-T approval.
Asserted to be among the highest gains reported at this meeting.
How much text? 2 pages.
Showcase? Yes. Was presented. Uploaded in .zip container.
Interesting. Mature for adoption now in time-frame of 4:4:4 amendment? Can postpone to next
amendment if necessary. JVT decision: Adopt in 4:4:4 amendment with minor TBD adjustment
of syntax to ensure extensibility.
JVT-U058* [Q. Chen, Z. Chen] Modif scene info SEI message
This document proposes to add a new scene type, “flash”, in scene_transition_type in Scene
information SEI message for the frequently appeared case. No syntax change is needed in this
proposal except the semantic definition.
Proposes adding an SEI message in which a "flash" indication is inserted into an otherwiseunchanged scene info SEI message.
Seems too minor to add another whole, mostly-duplicate SEI message just for this – although it
probably would have been a good idea if we had thought to include this or to reserve additional
possible values in the first place (prior to standardization of the current SEI message).
We encourage investigation of the possibility of creating a more capable and flexible future SEI
message that might include this one small item within its scope of capabilities.
JVT-U059* [Z. Chen, Q. Chen, X.D. Gu] SEI for functional app
Organization, fast indexing and retrieval of desired media data from huge amounts of storage
media are becoming more and more important due to the fast increasing of digital multimedia
content. However, the existing H.264/AVC video coding standard does not provide such a
function for fast video/image indexing and retrieval applications, and this is alleged to limit
further usage of H.264/AVC to some extent. This proposal aims to solve this perceived problem.
SEI Message for image/video retrieval is proposed to be inserted into H.264/AVC bit stream. It is
reported that with a small amount of SEI overhead, fast image/video retrieval can be achieved
without decoding the whole video bit stream. Some potentially benefited applications are Internet
image/video retrieval, personal media content retrieval, and huge amounts of media retrieval in
TV station.
Proposes a hierarchical structure containing three types of SEI messages (or equivalents). One
describes colour characteristics. Another describes "motion activity" degree according to an
arbitrary scale to be determined at the discretion of the encoder. Another is a "semantic
metadata" message containing arbitrary text strings in ASCII.
Some potential overlap with MPEG-7, MAFs, etc. Idea seems important and potentially fruitful,
but needs further study. JVT decision: Create "video annotation" AHG chaired by T.Wiegand to
conduct such study.
180
q. JVT Multi-view coding
AHG reports (JVT-U015, JVT-U016, JVT-U017) presented by A. Vetro and Y. Su.
Still some gaps between the JMVM text and the software, but nothing critical at this phase.
Some issues with JMVM (also detailed in input contribution):
-
reference picture management: No differentiation between anchor and non-anchor
pictures
reference picture list construction not well defined
view_id and anchor_pic_flag are somewhat redundant
marking process for anchor pictures
HRD: Parallel output of pictures
Decide whether the doc on encoder opt. should be kept separate or be integrated in JMVM. At
least have a pointer. Make sure all information is available to JVT – submit any needed info as
early input to next meeting if necessary.
i. CE 10 & related docs: view interpolation
JVT-U063* [S. Yea, A. Vetro] CE10: View synthesis prediction
This document provides an update on the previous CE 10 report of Klagenfurt and describes the
current status of CE 10 view synthesis prediction for multiview coding.
The opinion was expressed that the initial provided abstract (slightly worse than shown above)
was not adequate. A new abstract was reportedly prepared and provided in a revised copy of the
document. However, the revised upload could not subsequently be located. The authors are
requested to endeavour to follow the JVT working practices more diligently in the future.
Finding that by RD opt. more 16x16 blocks were selected than expected (though not giving a
more accurate disparity field). Current results with warping of one frame (P slice), not fully
implemented. New ideas on adaptive depth search (non uniform) and depth to disparity
conversion. Correction vector coding seems to be suboptimum, giving too large rate overhead.
Current gains marginal, roughly 0.15 dB for sequences that were tested until now. Proponents
True depth estimates usually diverge from the RD optimized disparity vectors. Warping
prediction is used, but unclear what coding gain would justify the additional complexity.
Proponents expect more gain from B slice encoding, because then more frequently residual
coding would be zero. Also study visual quality impact, not only PSNR.
JVT decision: Put into CE.
JVT-U093* [H. Kimata, S. Shimizu, M. Tanimoto, T. Fujii] CE10: MVC view interpolation
pred
This document presents the overview of the proposed method for View Interpolation Prediction
for MVC, and it shows the summary of experimental results of CE10, which was the core
experiment on the technologies of view interpolation. Based on the test conditions of CE10,
highest results were asserted to be about 10% in bitrate savings. For additional experiments
181
regarding alternative inter-view prediction structures, highest results were asserted to be 22% at
PSNR 34.5.
Camera parameters to give the “zero” offset of disparity. Evaluation: Average effiiency for all
views, and efficiency for anchor pictures. Additionally to the structure IBPBP.. of CE, other
structures such as IbBbP are used. For Rena, max. 10% saving, for Akko&Kayo max. 8%. With
IbBbP, the proposed method gives 22% max for Rena, 10% max. for Akko&Kayo. For anchor
pictures, gains are even higher.
Interesting gain, but need to be verified by subjective test.
Same QP was used for B and b in the IbBbP structure. This should be corrected, because the
performance is suboptimum.
Need all sequences, average savings (4 points to compute Bjontegaard measures).
Search range for anchor was possibly too small, this should be corrected.
Prepare more information how PSNR behaves over view and time.
Put into CE.
JVT-U102* [Y. Ho, C. Lee, S. Yoon, K. Oh, B. Choi] View Interpolation for MVC
This document presents a method of view interpolation for multi-view video coding (MVC), and
it reports experimental results. The document reports improvements in the quality of the
synthesized image using several steps. The first step is the initial disparity estimation using
region dividing which does not need the maximum disparity. Upon initial disparities, the
proponents estimate find disparities using that variable block-based estimation and pixel-level
estimation having adaptive search range. In addition, the disparity error correction process has
included reducing the disparity errors. The experimental results reportedly show the quality of
synthesized image to have been improved about 1-3dB.
The opinion was expressed that the abstract provided with the contribution (somewhat worse than
what is shown above) was not adequate for JVT purposes. The authors are asked to endeavour to
work harder to follow the proper JVT working practices in the future.
Try to improve view interpolation by imposing ordering constraint and applying region subdivision to DE. Variable block based estimation, adaptive search range, error correction and
median filtering, modified cost term for homogeneity of disparity field. Synthesize image by
linear interpolation. No coding results given. Claim that it performs better than the scheme from
JVT-U093 showing that PSNR values for the interpolation result are better.
Added to CE 10.
JVT-U138-L [T. Senoh, T. Aoki, H. Yasuda, T. Kogure] CE10: Inter-camera prediction
A result of inter camera picture prediction experiment is reported here. At lower bitrates such as
QP=31 or 32, disparity vector prediction results outperformed the method without it about 0.1
dB.
NOTE: Not a proposal – only a report.
182
Use IbBbP structure. Use camera parameters for prediction of disparity vectors. No temporal
predction used. Generate common disparity vector for all three (bBb) pictures. This performs
better for the two b, and worse for the B picture. Rate is lower, but on average, RD performance
seems to be approx. same as without the method.
Summary:
Continue CE with the modifications to get missing information as said above (Anthony Vetro to
coordinate new formulation).
Suggested to perform subjective testing jointly with Tobias during the week. Activity put into CE
plan.
ii. CE 11 & related docs: illumination compensation
JVT-U027* [D. Sim, S.N. Park] CE11: MB-based illumination comp.
This document describes a macroblock-based illumination compensation for MVC. In this
proposal, not only offset but also weight value is employed to compensate the illumination
change.
Illumination invariant ME only applied to 16x16 inter mode. Average of block is subtracted and
weight applied prior to ME. Transmit weight resolution in slice header and weight factor in MB
header. Prediction error (for average) also sent in MB header. Prediction of current macroblock
average performed similar as 16x16 intra prediction in AVC.
Results: Race, Exit, Uli same as JMVM. Race1 0.4 dB gain, Flamenco2 0.2, Rena 0.4,
Akko&Kayo 0.4 gain.
Proponent suspects that mainly the average prediction is effective, and that weighting only costs
rate.
Comparison done against JSVM without weighting.
Currently, no gains were found for the cases of non-16x16 inter macroblocks.
Outcome noted elsewhere in report.
JVT-U053* [Y.-L. Lee, J.-H. Hur, S.H. Cho, N.H. Hur, J.W. Kim] CE11 Kwangwoon
University Illum Comp
This document presents the cross check result for the core experiment 11 on multiview video
coding of Kwangwoon University. For the verification test, source code was compiled and ran for
all the provided bitstreams. The appendix of this document encloses the experimental results
produced by the Kwangwoon decoder. The excel file zipped with this input document gives the
experimental results. It is reported that all the results of RD- curves are the same compared to
those provided by Kwangwoon University.
Results were verified, source code compiled but not closely checked.
JVT-U031-L [J.-H. Yang] CE11: Illum. comp. consistent pred.
This contribution proposes an illumination compensation ( IC ) scheme for MVC. The IC
information consists of the IC flag and the IC offset, which reflects the relation between a coding
block and its reference block. Thus, the bi-predictive block has two pairs of IC information. The
183
IC offset of a block is predicted with those of the neighboring blocks by the proposed prediction
scheme. The IC flag and, if the IC flag is true, the residual of the IC offset are coded by the
entropy coder. In case of the bi-predictive block, the averaged IC offset is encoded, and two IC
offsets for each reference block are reconstructed with the help of the means of reference blocks
in the decoder side. The simulation results report an SNR improvement in the range of 0.1 ~
0.5dB, depending on the test data.
IC offset handled similar to MV: Prediction of IC information from neighboring MBs. Assume
that constistency of IC information is given when reference frame of two adjacent MBs is the
same. “Weak” consistence in case of B prediction when one of the reference frames is the same.
In direct mode, no IC information is sent.
Max PSNR gains (usually quite consistent over whole range of rates) Ballroom 0.1; Exit 0.15;
Uli 0.15; Race1 0.5; Flamenco2 0.3; Breakdancers <0.2; Rena 0.5; Akio&Kayo 0.4.
JVT-U072* [K. Sohn, Y. Kim, J. Seo, J. Yoon, J. Kim] CE11: Verif LG/SNU JVT-U031-L
Illum comp
This document reports the cross-check results of JVT-U031-L “CE11: Illumination compensation
consistent prediction” by LG/SNU. The source code, configuration files and coded bitstreams
were provided. The verification has been performed by decoding the bitstreams provided by
LG/SNU. The simulation results of JVT-U031-L are confirmed.
Not presented in detail. Noted.
JVT-U052* [Y.-L. Lee, J.-H. Hur, S.H. Cho, N.H. Hur, J.W. Kim, Y. Su, P. Yin, C. Gomila,
J.-H. Kim, P.-L. Lai, A. Ortega] CE11: Illumination compensation
No abstract in document.
Compared to AVC, the proposed method employs predictive coding for the DC component of
Inter prediction residues. The predictor for illumination change is formed from neighboring
blocks because illumination differences tend to be correlated spatially. The proposed scheme is
enabled for a number of MB coding modes. The proposed method is reported to show 0.1~0.7dB
PSNR gains for various MVC test sequences in comparison to JSVM 6.5 without weighted
prediction. Compared to RD-optimized weighted prediction, the proposed method is reported to
provide up to 0.6 dB gain.
Revised contribution, which only performs operation on 16x16 blocks. Perform Prediction of IC
offset values from surrounding MBs. For skipped MB and for direct mode, IC is derived without
encoding. Contribution includes syntax proposal.
Comp. against JMVM w/o WP (max gains). Ballroom 0.1, Exit 0.1, Uli 0.1, Race1 0.6,
Flamingo2 0.3 dB, Breakdancers 0.2, Rena 0.5, Akko&Kayo 0.4
Comp against JMVM w opt. WP (max gains): Ballroom 0, Exit 0.1, Uli , Race1 0.6, Flamingo2
0, Breakdancer 0.1, Rena 0.4, Akko&Kayo 0.4.
Average over all rate points and sequences around 0.2 dB.
Weighted prediction at slice level tested with no WP, scale only, offset only.
184
Complexity increase at encoder estimated around 14-15%.
Outcome noted elsewhere.
JVT-U028* [D. Sim, S.N. Park] CE11 Sejong/ETRI's illum. comp. JVT-U052
This document presents cross check results for the Sejong Univ./ETRI/Thomson proposal JVTU052 for the core experiment 11 on illumination compensation for MVC.
Establish BoG to identify the common basis of all proposals, which most likely are the method of
IC offset prediction and encoding of difference (including flag to turn on and off) and usage of
direct mode.
Outcome noted elsewhere; CE continuing.
JVT-U053* [Y.-L. Lee, J.-H. Hur, S.H. Cho, N.H. Hur, J.W. Kim] CE11 Kwangwoon
University Illum Comp
This document presents the cross check result for the core experiment 11 on multiview video
coding of Kwangwoon University. For the verification test, source code was compiled and ran for
all the provided bitstreams. The appendix of this document encloses the experimental results
produced by the Kwangwoon decoder. The excel file zipped with this input document gives the
experimental results. It is reported that all the results of RD- curves are the same compared to
those provided by Kwangwoon University.
Verification noted.
JVT-U072* [K. Sohn, Y. Kim, J. Seo, J. Yoon, J. Kim] CE11: Verif LG/SNU JVT-U031-L
Illum comp
This document reports the cross-check results of JVT-U031-L “CE11: Illumination compensation
consistent prediction” by LG/SNU. The source code, configuration files and coded bitstreams
were provided. The verification has been performed by decoding the bitstreams provided by
LG/SNU. The simulation results of JVT-U031-L are confirmed.
Unified proposal:
- prediction of DC offset
- direct mode
JVT decision: Adopt unified proposal for JMVM
- this may be the first building block of a package of tools which is sought to have
significant improvement of compression performance in multiview applications
- therefore, it is needed in the developments of ongoing CEs
- does not imply automatic future transferral into WD by next meeting.
Potentially check benefits of this tool for motion-compensated single-view coding? (note: look at
JVT-C066 and JVT-D122) – proponent indicated no gain found.
185
iii. High-level syntax
JVT-U026* [P. Pandit, Y. Su, P. Yin] Comments on High-level Syntax for MVC
In the current MVC specification, frame_num and POC between the different views is decoupled,
thus allowing pictures with the same frame_num and POC to be present in the DPB. These
pictures are differentiated using the view_id associated with it. In order to manage the decoded
picture buffer (DPB), the current implementation uses AVC compatible MMCO commands.
These MMCO commands only operate on the pictures with the same view_id as the one that is
used to carry these MMCO commands. This increases the DPB requirements for a MVC system.
In order to allow for a smaller DPB size (thus using less memory) the way MMCO commands
are currently defined require a change. This contribution proposes changes to the existing
MMCO syntax to efficiently manage the DPB. Additionally, the default initialization of the
reference pictures and subsequent reordering of these reference pictures (using new RPLR
syntax) is also presented.
Contribution noted. Further discusson needed.
Comment on view scalability: view_id may not be needed.
Comment on base view compatibility.
considered in break-out report; results noted elsewhere.
JVT-U046-L [W.S. Shim, H.S. Song, Y.H. Mun, J.B. Choi] High-level syntax for flexible I
frame position
In this document, high level syntax for I-frame position method is proposed aiming to reduce the
imbalance of image quality in each view and efficient syntax to represent view dependency
change.
Signals dependencies, anchor pictures and I frame positions. Actually something for an SEI
message. Benefits not shown to motivate group to further consider this.
JVT-U048* [S. Lin, P. Zeng, J. Zhou, Q. Xie, C. Hu, L. Xiong] MVC high level syntax:
Camera Parameters
This contribution proposes camera parameters be coded in MVC in order to improve the potential
coding efficiency by exploiting the view dependency.
In general, we believe that camera parameters are useful for display and potentially for efficient
coding. Until we are been shown a use for these parameters we are unable to specify a
transmission method. The uncertainty in specifying these includes aspects such as which
parameters to transmit at which accuracy to enhance the display process and maybe the
compression process or both together.
JVT decision: For further study.
JVT-U060* [H. Nakamura, M. Ueda] MVC H-L syntax parallel proc
This contribution proposes to add a new high-level syntax element for MVC. The parallel
processing is one of the essential functions for MVC decoder in order to decode multi-view video
bitstreams in real-time. In the views coded with using disparity compensation, MVC decoder
needs to delay the decoding timing of each view compared with the decoding timing of the
reference views. The proposed syntax element help in facilitation of finding a decoding timing
186
for each view aiming at enabling efficient decoder implementations on parallel processing
platforms.
View_dependency_count to determine the decoding time (delay) necessary in parallel processing
of views. Seems useful, this should be an SEI message.
Information is already present.
Create/extend SEI message covering the aspect addressed in JVT-U060, indicating maximum
number of views, and number of reorder pictures for a decoder to limit processing requirements
and to facilitate parallel processing. JVT decision: Agreed – left to JMVM editor to reflect in
JMVM.
JVT-U062* [A. Vetro, S. Yea] On MVC DPB management
This document describes several issues related to the text for MVC reference picture
management. In particular, the requirements for reference picture lists construction and the
reference picture marking process are reviewed and suggested changes to the text are described.
Clarifications and improvements of existing syntax.
Would be useful to get information from industry to potentially re-use existing hardware for
MVC, to decide how far the syntax of an MVC profile should be allowed to deviate from existing
syntax. Current assumption: May be problem below slice header.
Notes elsewhere
JVT-U103-L [Y.-K. Wang, Y. Chen, M. M. Hannuksela] Comments to JMVM 1.0
This document presents some comments on JMVM 1.0, related to view_subset_id, the inter-view
reference picture, view subset and random access.
Reduction to 8 bits: no.
Change u(10) to ue(v): no. (for the time being, at least)
Remove view_subset_id: yes.
Remove constraint on anchors: no
JVT requests proposals on a concept for HRD, Levels depending on number of views,
frame_num, POC and DPB handling.
temporal_level inserted in NALU header extension (unless those 3 bits can be used much more
efficiently): yes.
JVT decision: Adopt
JVT-U104-L [Y.-K. Wang, Y. Chen, M. M. Hannuksela] Time-first coding for MVC
This doc proposes a different decoding order (referred to as time-first coding) than the one
specified in JMVM 1.0 (referred to as view-first coding). In view-first coding, for each group of
pictures (GOP), pictures of any view are contiguous in decoding order. In time-first coding,
pictures of any temporal location are contiguous in decoding order. Also an analysis is given
showing that, with view scalability, time-first coding requires smaller decoded picture buffer
(DPB) size than view-first coding.
Access unit definition: Can there be more than one sample per time instant
187
JVT-U105-L [Y. Chen, Y.-K. Wang, M. M. Hannuksela] MVC reference picture
management
In this proposal, methods for reference picture marking, including both sliding window and
adaptive memory control mechanisms are proposed to efficiently manage the decoded reference
pictures. Methods on reference picture list construction including both reference picture list
initialization and reordering are also proposed. The proposed methods are primarily targeted for
time-first coding. However, some tools can also be used for view-first coding.
Contribution noted.
Results of breakout reported by Anthony Vetro. Adopt the HL syntax with updates produced by
BoG to JD. Uploaded as d0 of output JD draft. This documents some above issues remarked as
"documented elsewhere"
iv. Other technical inputs on MVC
JVT-U040-L [S.-H. Lee, S.-H. Lee, N.-I. Cho, J.-H. Yang] MVC: Disparity vector
prediction
For disparity vectors, spatially co-located values are often not available for prediction.
Assumption that disparity vectors are highly correlated over time axis. Use disparity vector from
co-located macroblock in reference frames. Flamenco2 0.25 dB gain is maximum, usually around
0.0-0.1 dB. Expect more gain if illumination compensation is implemented, would like to further
investigate.
Additional complexity due to necessary memory of disparity vectors from reference frames.
Contribution noted.
JVT-U047* [H. Yan, J. Huo, Y. Chang, S. Lin, P. Zeng, L. Xiong] Regional Disparity
Est/Comp for MVC
In this document, regional disparity estimation/compensation technique together with region
partitioning method for multi-view video coding (MVC) is proposed.
Depth object is an image region where pixels have approximately the same disparity. Global
disparity compensation does not help because of different depth of objects. Partitioning is made
on the basis of MBs. Regional disparity value is used as initial search position. Differential
encoding performed. Comparison made for JSVM with certain search range against proposed
method with region prediction values already known.
Is purely about encoder optimization, but actual complexity reduction not known.
Contribution noted.
JVT-U068* [K. Ugur, J. Lainema, M.M. Hannuksela, H. Liu] On parallel
encoding/decoding of MVC
Inter-view dependencies between pictures in Multi-view Video Coding (MVC) may impose
serious parallelism issues to the video system, because two pictures at different views need to be
decoded sequentially. This is especially problematic for 3D-TV use-cases with displays
supporting head-motion parallax, where many views are displayed simultaneously. Because
pictures in different views can not be decoded in parallel, the only way to display simultaneous
views is by having a decoder running N times faster than a regular 2D video decoder (N being the
188
number of views simultaneously displayed, which might be over 100). Similar problem also exist
for real-time Free Viewpoint Video (FVV) use-cases where the MVC encoder has to compress
the inter-dependent views real-time. This contribution proposes a coding structure that enables
parallel encoder/decoder implementation for different views, even though there are dependencies
between views. This is achieved by coding views with some constraints, so that any macroblock
in a certain view is allowed to depend only on reconstruction values of a subset of macroblocks
in other views.
Parallel encoding important, but interview dependencies impose a burden on parallel
implementation, requires shared memory etc.. Impose constraints about macroblocks that may
not be allowed to be accessed. Two parameters are made available at each MB: pds_bloc_size
and pds_initial_delay. Operation of de-blocking filter is changed as well: “sliding deblocking”
does not filter block edges towards not available macroblocks. For interpolation, padding is used.
Difficult to test, as reference dynamically changes.
Penalty of scheme is around 0.5 dB for one row delay and around 1 dB for one MB delay. In
addition, visual artifacts due to additional block edges. Better use simulcast which may be less
complex and can use existing hardware.
Further work not really encouraged.
JVT-U091-L [H.-S. Koo, Y.-J. Jeon, B.-Y. Jeon] MVC motion from neighbor view
This document proposes a motion skip mode for MVC which is generated from the idea that
there is a similarity of motion between the neighboring two views. In motion skip mode, the
motion information such as mb_type, motion vector, and reference indices is inferred from the
corresponding macroblock in the neighboring view at the same temporal instance. Thus the
motion skip mode is very similar to base mode in SVC skipping motion information in the stream.
Since the disparity between two views exists, the global disparity between two views is
calculated and applied to find the corresponding macroblock. Preliminary experimental results
show that maximum gain is up to 0.6 dB.
Idea to infer the motion information from neighboring views (similar idea as SVC where the
motion information is inferred from the next lower layer). Introduce motion skip flag when the
motion vector can be derived. Use global disparity to decide the position of the related
macroblock. Preliminary results only for Ballroom and Exit. Gain around 0.6 dB for Ballroom,
minor for Exit. Suggested establishing a CE.
Comment: May be wrong to use the global disparity. That would only work for background area,
where however the motion would be zero typically. Model seems to be reasonable, but the
question is where to get information about local disparity which is not available when MC is
used. Outcome noted elsewhere.
JVT-U100* [Y. Ho, K. Oh, C. Lee, P. Park, B. Choi] Global Disparity Comp for MVC
This document proposes global disparity compensation for MVC. After explaining the global
disparity and its compensation experimental results are presented asserting the the effectiveness
of the proposed method.
Global disparity is capable to compensate for large offset that is typically present in disparity (in
particular for cameras with parallel optical axes). Helps in cases where search range is low. With
sufficiently large search range, no gain is achieved.
189
Would be beneficial to implement in JMVM a method that allows searching with a smaller
search range starting from global disparity offset. Does not seem to be necessary to actually
transmit the offset value.
JMVM software: Check if it is possible to integrate approach for disparity estimation with
reduced search range around global offset. Outcome noted elsewhere.
JVT-U101* [Y. Ho, K. Oh, C. Lee, P. Park] Reference Frame for MVC
This document describes the reconstruction of reference frames for MVC.
The opinion was expressed that the abstract provided with the contribution (somewhat worse than
what is shown above) was not of adequate quality for JVT purposes. The authors are asked to
endeavour to work harder to follow proper JVT working practices in the future.
Global DC based on camera parameters used for rectification which increases the correlation
between images in case of non-parallel cameras. Race1 and Uli Coding gain is below 0.1 dB.
Unclear which type of filter is used.
Further work encouraged.
JVT-U134-L [H. Kimata, S. Shimizu] On direct mode for MVC anchors
This document proposes a simplified coding method of direct mode for anchor pictures of MVC
to reduce memory usage for disparity information.
Proposes a new method for derivation of disparity vectors from collocated picture. Coding loss
negligible.
Seems useful to reduce the complexity (exact figures by how much he complexity is reduced
would still need to be given. Seems not to be appropriate for case where existing AVC decoder
shall be used for MVC. Further study recommended.
JVT decision: Establish CE from JVT-U040 and JVT-U091. Leader: H.-S. Koo.
v. Reference software, common conditions, encoder optimization
JVT-U061* [A. Vetro, S. Yea, P. Pandit, Y. Su] MVC ref software implementation plan
This document outlines several issues with the current reference software and a proposed plan to
resolve them.
Detailed presentation not needed.
JVT-U069* [K. Ugur, J. Lainema] On common conditions for MVC
This contribution proposes simplified common conditions for MVC.
Proposes to remove need to test some temporal prediction structures in testing of some proposals.
No action.
JVT-U071* [K. Sohn, Y. Kim, J. Seo, J. Yoon, J. Kim] Encoder optimization of MVC
This document reports updated results for encoder optimization of MVC. The previous prediction
structure described in JVT-T102 is modified to obtain improved results. The proposed encoder
obtained some PSNR gain (about 0.2 dB) for the test sequences.
190
Contribution noted but not reviewed in detail due to lack of time. Information on topic can be
resubmitted in future
r. JVT proposals of additional profiles and levels
These documents reviewed in joint meeting with MPEG Requirements Wed 4-6pm.
JVT-U018* [SMPTE] LS: Constraints on High 10 profile (WG 11 input document M13841)
Response by SMPTE to liaison statement N8278 from MPEG, “Constraints on High 10 Profile”.
SMPTE suggests that two new intra-only H.264/AVC profiles be added. The first is an intra only
version of High 10; the second is an intra only version of High 4:2:2.
Proposes two new profiles (clusters "A", "B") with characteristics as follows:
– Intra-only
– One like High 10
– One like High 4:2:2
Proposal is essentially the same as JVT-U120 proposal (no differences identified).
Decision documented elsewhere in this report.
Liaison reply sent by MPEG parent body as documented below.
JVT-U019* [SMPTE] LS: New profile for production (WG 11 input document M13842)
Response by SMPTE to liaison statement N8278 from MPEG, “Constraints on High 10 Profile”.
SMPTE agrees that we should collaborate in identifying new applications for AVC. SMTPE
continues to believe that there is an opportunity for a new Profile intended for high-quality
production applications that is designed to minimize computational complexity.
Proposes a new profile (cluster "C") with characteristics as follows:
– Application focus is production (very high quality, very high bit rate)
– Minimization of computational complexity.
Discussion: Presumably at least 10 bit and at least 4:2:2 if interlace support is needed. Probably
4:4:4 and lower formats; probably up to 12 bits; probably intra-only (based on prior LS content).
Liaison reply sent by MPEG parent body as documented below.
JVT-U143-M [T. Suzuki] Level definitions for prof apps
Suggests that current bit rates and CPB sizes for High 10 and High 4:2:2 were motivated by a
need to support all-Intra coding.
Suggests if we define new intra-only profiles corresponding to current High 10 and High 4:2:2,
then to define a corrigendum to lower the maximum bit rates and CPB sizes of High 10 and High
4:2:2. (If we don't do that, suggestion is to keep the current definitions as they are.) JVT
decision: Open to further study.
191
Suggests to lower the bit rate and CPB sizes for the drafted "High 4:4:4 Inter" profile under
development. JVT decision: Open to further study.
Remark: Intra and inter coding efficiencies approach each other at high bit rates.
Proposes to create a new profile (or "conformance point" – cluster "D"):
– Intra only
– Tools of High profile (8 bit only, 4:2:0 only)
– Maybe a higher bit rate
Proposes to use constraint_setX_flag to minimize the number of profiles while enabling the
definition of more "conformance points". JVT decision: Keep this suggestion in mind.
JVT-U066* [P. Symes, H. Yu] Simple Intra profile for prof apps
This contribution presents a proposal for a new Profile for H.264/AVC intended for large picture
/ high quality / high bit rate applications that minimizes computational complexity.
Proposes a new profile (cluster "C") with characteristics as follows:
– Intra-only (no deblocking) [as with draft High 4:4:4 Intra, if specify without deblocking]
– All color formats supported (4:4:4, 4:2:2, 4:2:0, Monochrome) [as with draft High 4:4:4
Intra]
– Up to 12 bits [less than with draft High 4:4:4 Intra]
– No CABAC [primary apparent difference with draft High 4:4:4 Intra]
– Otherwise roughly like draft High 4:4:4 Intra profile
Why not 13 & 14 bits?
Applications include very high bit rates and resolutions. Various application details described.
Motivation: Implementation on general-purpose CPUs (e.g., laptops). Remark: Is that a realistic
expectation? – very high bit rates, bit depth, picture size, …
Part of document describes applications with picture sizes up to 4096x2160 (our level 5.1
currently supports up to 26.7 fps), frame rates up to 300 fps, 12-bit 4:4:4 lossless, bit rates up to 5
Gb/s. Indicates that compatibility with existing profiles may not be required.
Outcome noted elsewhere
JVT-U120* [T. Wedi, H. Ohtaka, J. Wus, S. Sekiguchi] Intra-only profile for prof apps
This document summarizes a proposal for the creation of new intra-only profiles for professional
applications. In particular, additional Intra-only High 10 and Intra-only High 4:2:2 profiles are
proposed that use all of the tools within the existing High 10 and the High 4:2:2 profiles with the
exception of the Inter coding tools. Furthermore, it is proposed that these new Intra-only profiles
are defined using an onion shell representation together with the High 4:4:4 Intra profile.
Proposes two new profiles (clusters "A" and "B") with characteristics as follows:
– Intra-only
– Tools and other constraints (bit rate, etc.) otherwise corresponding to High 10 and High
4:2:2.
192
Outcome noted elsewhere
s. JVT errata and clarification issues for AVC
Output document JVT-U210 to be produced incorporating issues noted herein and others
identified by the editor of the output document, Gary Sullivan.
t. JVT JM encoder optimization
JVT-U029-M [A. Leontaris, A.M. Tourapis, K. Suehring] ME & MC Enhancements to JM
ref soft
Some expression of support for including.
Testing/debugging of the latest software encouraged.
Thanks were expressed for the good hard work; the ref software coordinator and the relevant
AHG was given discretion for final handling of provided software.
JVT-U030-L [A.M. Tourapis, K. Suehring, G.J. Sullivan, A. Leontaris] Revision of JM ref
software manual
Revision proposed for reference software manual. JVT decision: Adopted with thanks.
JVT-U079* [K.B. Kim, M.-C. Hong] Search range for fast ME
DSR (Dynamic Search Range) decision has been adopted for fast motion estimation in previous
JVT meeting, and the modification methods have been presented in JVT meeting. In this
contribution, we propose a modified DSR algorithm and VSS (Variable Step Search) motion
estimation algorithm. The experimental results assert that with the new search range decision,
71% reduction of encoding time can be obtained with marginal sacrifice of PSNR (less than
average 0.04 dB) than FS (Full Search) motion estimation (5% more encoding time reduction
than the previous DSR), and that with the combination of proposed DSR and VSS, 87%
reduction of encoding time can be obtained with 0.07 dB PSNR loss than FS motion estimation
(20% more encoding time reduction than the previous DSR).
Proponent could not be present at final session when presentation opportunity arose. Insufficient
time requires deferring consideration to further study in the future.
u. JVT internal operating rules
JVT decision: The following clarifications/adjustments of JVT operating rules have been
approved.
The JVT decided that participants shall to refrain from long (=more than 4 Minutes) presentations
of their proposal, if the results of their coding efficiency experiments have provided less than 2%
bit-rate on average (or equivalently 0.1 dB gain on average).
Also see additional notes elsewhere regarding inappropriate "cherry picking" of results for
summary reporting in abstracts and presentations.
193
Regarding late contributions: Due to our difficulties with a large quantity of late-submitted
contributions at this and other recent meetings, the JVT has agreed that for its next meeting, no
late-uploaded (non-AHG-report, non-liaison) contribution will be presented without having a
minimum of 4 JVT participants (working for organizations other than that of the primary
contribution author) recorded by name as supporting the allowance of such a presentation, in
addition to a consensus of the general JVT membership to allow the presentation. Such support
to allow a presentation is to be understood to not necessarily imply support of the adoption of the
content of the late contribution, but only as a positive expression that the document should be
allowed to be presented. Additionally, the provider of a presented late contribution shall send an
email apology to the JVT email reflector. This rule does not apply to material requested by the
JVT at the meeting (e.g., reports of JVT-authorized side activities).
All submissions must be made in JVT-Uxxx.zip format with the word docs, excel sheets and
other information being in the zip container. The document must contain an abstract and be
accompanied with an e-mail notification containing title, authors and abstract (identical to the one
in the doc) which is no longer than 200 words and is written in 3rd person in a manner that does
not express endorsement of the content of the document.
On filenames inside of .zip containers – use a filename so that if you take the files out of the zip
container, you'll still know what contribution they came from. Every file in the .zip container for
document JVT-Uxxx should start with JVT-Uxxx. Example: JVT-Uxxx.doc (main document),
JVT-Uxxx_presentation.pdf, JVT-Uxxx_results1.xls, etc. PDF is preferred over PPT for
presentations when the PPT filesize is large and there is no need for the slide deck to be editable
by others.
When providing additional or revised files, do not include copies of files that were already
included in the prior .zip archive for the same contribution and do not re-use the same filenames
without adding revision numbers (r1, r2, etc.) – this saves us needing to worry about whether the
files we get with the same filenames are the same or different.
Independent verification (necessary for adoption of a proposal) is provided either through
a) independent implementation by 1 or more company different than the proponent based on
the textual description (after adoption, both decoder source code versions must be made
publicly available and one encoder version)
b) providing source code to all CE participants prior to the meeting (CEs can only be joined
at the meeting, when the CE is created. CEs are created at each meeting and last until the
next meeting.)
Simply running binary executables provided by a proponent is not ordinarily considered
independent verification. Source code should be provided and used, and the verifying party
should invest a proper degree effort to ensure that the “verification” they perform is a meaningful
and professional study with significant depth rather than just a perfunctory procedural formality.
For every SEI message and every syntax element that are currently in the SVC draft, a showcase
has to be provided in order to retain it in the JSVM/WD. If such a showcase is not provided at the
next meeting for an SEI message or parts of it, the SEI message or the respective parts will be
removed from the JSVM/WD. The source code and executables for the showcase must be made
available.
A first CE description must be available at the last day of the meeting. Changes of the CE
description are only allowed until 1 month prior to the next meeting. These changes must be of
evolutionary characteristic relative to the input documents on which the CE is based and must be
agreed by those who contributed the respective input document(s) or be added as an option.
194
v. List of JVT adoptions
Person listed in bracket is responsible for provision of text and software integration.
i. Normative SVC adoptions into JSVM
JVT-U125* [Y. Bao] CE1: Results PR slice improve: CAF
JVT-U129-L [J. Ridge] Component separation FGS: Byte alignment, SEI message (JVTU129r1-L)
JVT-U082-L [D. Marpe] CE3: Improved CABAC for PR slices
JVT-U042* [A. Segall] CE4: Texture Upsampling with 4-tap Cubic Spline
JVT-U126* [Y. Bao] CE4: L-C smooth ref spat SVC
JVT-U130* [X. Wang] CE6: ESS Inter-layer pred
JVT-U067* [G.J. Sullivan] Position Calc for SVC Upsampling
JVT-U085* [A. Eleftheriadis] Clarif Nesting Temporal Levels: Add to scalability SEI
message (JVT-U085r1)
JVT-U090-L [S.-W. Park, B.-Y. Jeon] Usage of store_base_rep_flag
JVT-U106-L [Y.-K. Wang] Discardable data adaptation: Behaviour at slice boundaries:
Make it switchable: Intra_BL upsampling process and deblocking is handled as if the slice
boundary would be a picture boundary.
JVT-U109-L [Y.-K. Wang] On SVC high-level syntax:
Restrict MMCO to only apply to pictures with equal or a larger value of temporal_level
Express this as a constraint on how to construct temporal_level values.
Restrict RPLR such that the final list of ref pics only contains pics that have temporal_level
smaller or equal to the temporal_level of the current picture. Express this as a constraint on how
to construct temporal_level values.
JVT-U116* [A. Eleftheriadis] Err resil frame nums in key pics: Adopt as extra Byte at NUL
header extension for d=0, q=0 base layer and switchable with a bit.
Discussion to remove a conditionally-adopted SEI message: JVT-T073 for association of an SEI
message with a scalable layer. Seems to be a fundamental part of high-level syntax design
operation of SVC. As an exception, it is agreed that the showcase requirement will be waived for
this.
ii. Non-Normative SVC adoptions
None unless noted elsewhere in this report.
195
iii. SVC software adoptions
None unless noted elsewhere in this report.
iv. Normative 4:4:4 and professional profile adoptions
See above section on 4:4:4 proposal dispositions.
JVT-U035* [S. Wittmann, T. Wedi] Post-filter hint SEI
JVT decision: Adopt in 4:4:4 amendment with minor TBD adjustment of syntax to ensure
extensibility.
None others unless noted elsewhere in this report.
v. Normative MVC adoptions
JVT decision: The MVC BoG activity reported adoption of JVT-U052, JVT-U060, JVT-U062 ref
pic lis const & ref pic marking, JVT-U103 (syntax changes as noted elsewhere), JVT-U104, JVTU105/JVT-U026 sliding window indep. for each view.
vi. Other normative adoptions
None unless noted elsewhere in this report.
vii. Other non-normative adoptions
See section on JM encoder optimization contributions.
None others unless noted elsewhere in this report.
w. List of JVT AHGs established
JVT Project Management and Errata Reporting (jvt-experts@lists.rwth-aachen.de),
Chairs: Gary Sullivan, Jens Rainer Ohm, Ajay Luthra, and Thomas Wiegand
Continue mandates from previous meetings.
JM Text and Reference Software (jvt-experts@lists.rwth-aachen.de), Chairs: Thomas
Wiegand, Karsten Sühring, Alexis Tourapis, and Keng Pang Lim
Continue mandates from previous meetings.
Bitstreams and Conformance (jvt-bitstream@lists.rwth-aachen.de), Chair: Teruhiko
Suzuki
Continue mandates from previous meetings.
196
Study of 4:4:4 Functionality (jvt-experts@lists.rwth-aachen.de), Chairs: Teruhiko Suzuki
[Notes need update – at least he name is obsolete]
Mandates:
- To define test conditions for the investigation of 4:4:4 video coding tools.
- To investigate the complexity of 4:4:4 video coding tools.
- To maintain the specification and software for 4:4:4 coding
- To study profile definition
JSVM software and new functionality integration (jvt-svc@lists.rwth-aachen.de), Chair: J.
Vieron, M. Wien, H.Schwarz
Mandates:
- Coordinate JSVM software integration
- Coordinate bug-fixing process for the JSVM software
- Maintain JSVM software manual
JSVM and JD Text Editing (jvt-svc@lists.rwth-aachen.de), Chairs: Julien Reichel, Heiko
Schwarz, Mathias Wien
Continue mandates from previous meetings.
Spatial Scalability, Resampling and Inter-layer Prediction (jvt-experts@lists.rwthaachen.de), Chairs: Shijun Sun, A. Segall
Mandates:
- To consider alternative inter-layer residual prediction methods to improve coding
efficiency.
- To consider adaptive filter design for the luma upsampling.
- To consider practical (or shorter) downsampling filter design for both dyadic and nondyadic cases.
- To conduct experiments to evaluate the coding performance and (when necessary) visual
quality comparing to the current JSVM.
SVC High-Level Syntax and Error Resilience (jvt-experts@lists.rwth-aachen.de), Chairs:
Ye Kui Wang, S. Pateux, P. Amon, T. Schierl
Mandates:
- To optimize high-level syntax for NAL unit header, SPS, PPS and slice header
- To study whether the AVC HRD is suitable for SVC
- To study the adaptation of AVC SEI messages for SVC use
- To study enhancements to scalability information SEI message
- To consider SVC restrictions
- To refine the error resilience test conditions if needed
- To study error resilience in scalable video applications
- To build error resilient simulation environment
SVC Interlaced Coding (jvt-svc@lists.rwth-aachen.de), Chairs: Jerome Vieron
Mandates:
- To refine test conditions for validation and evaluation of interlace tools
- To complete the implementation of interlace tools in JSVM software
- To investigate solutions for improving inter-layer prediction for interlace material
- To evaluate SVC interlaced coding tools for different use cases
SVC Quantization, CAVLC and CABAC (jvt-svc@lists.rwth-aachen.de), Chairs: Justin
Ridge, Detlev Marpe, Gary Sullivan
Mandates:
197
-
To reduce complexity and cleanup of quantization, CABAC and CAVLC methods in
SVC.
SVC Complexity Reduction (jvt-svc@lists.rwth-aachen.de), Chairs: H. Schwarz, Y. Bao
Continue mandates from previous meetings.
MVC High-level syntax and buffer management (jvt-mvc@lists.rwth-aachen.de), Chairs:
A. Vetro, Y. Su
Mandates:
- To discuss high-level syntax for MVC including NAL unit type, NAL unit header
extension, SPS extensions, slice layer and integration with SVC syntax.
- To discuss reference picture management to enable simultaneous picture output of
different views and to facilitate parallel processing.
- To propose refined syntax and decoding processes for JMVM.
JMVM and JD text editing (jvt-mvc@lists.rwth-aachen.de), Chairs: Hideaki Kimata,
Aljoscha Smolic, Yeping Su, Anthony Vetro
Mandates:
- To collect comments on draft, perform necessary editing and upload final document by
the deadline.
- To maintain JMVM and JD document and collect comments on the text until the next
meeting.
JMVM software and new functionality integration (jvt-mvc@lists.rwth-aachen.de), Chairs:
P. Pandit, A. Vetro
Mandates:
- To implement high-level syntax and reference picture management process described in
JMVM into the reference software.
- To implement coding tools described in JMVM into the reference software.
- To upload the software for verification and testing according to the software integration
plan.
AhG on residual prediction modification, Chair: Yiliang Bao
Mandate:
- To investigate adding a switch for residual prediction in case of smooth reference
prediction.
AhG on enhanced spatial scalability, Chair: Jerome Vieron
Mandates:
- To consider alternative inter-layer motion prediction methods to improve coding
efficiency.
- To consider alternative inter-layer texture prediction methods to improve coding
efficiency.
- To consider alternative inter-layer prediction methods to reduce the complexity of the
current design.
- To evaluate requirements for ESS regarding the SVC profile definition.
AHG on bit depth and chroma format scalability (Yongying Gao, Andrew Segall, Thomas
Wiegand).
Mandates:
- Identify applications
- Work out suggestions for detailed needs
198
-
Find/create test material
define experiments
investigate software and text modification needs
identify complexity issues
AHG on video annotation (Jens-Rainer Ohm, Thomas Wiegand)
Mandates:
- Identify applications
- Work out suggestions for needs
- Find/create test material
- Define experiments
x. JVT software integration planning
Due to a lack of remaining meeting time, the scheduling of software integration was deferred to
be a post-meeting activity.
y. JVT Conformance bitstream planning
Volunteers for 4:4:4 and all-Intra profile conformance bitstreams: Mitsubishi (High 4:4:4 Intraonly and High 4:4:4 Predictive), Panasonic.
The following companies each announce to provide at least 10 conformance bitstreams for SVC:
HHI, Sharp, Thomson, RWTH (maybe), Nokia (potentially), Orange, Microsoft, Qualcomm,
Layered Media.
z. Resolutions conveyed by JVT to MPEG parent body
The JVT approved the following resolutions for conveyance to its MPEG (WG11) parent body.
JVT Meeting 21 WG11 Resolution 1: The WG11 video subgroup and the JVT recommend
approval of the following documents.
No.
Title
Available
14496-10 Advanced Video Coding
8449
Defect Report on ISO/IEC 14496-10:2005 (Version 2)
07/01/10
8450
Disposition of Comments on ISO/IEC 14496-10:2005/FPDAM1
06/10/27
8451
Text of ISO/IEC 14496-10:2005/FDAM 1 Support for Colour
06/11/10
Spaces and Aspect Ratios
8452
Study Text of ISO/IEC 14496-10:2005/FPDAM2 Advanced 4:4:4 06/11/14
Profiles
8453
Joint 4:4:4 Video Model (JFVM) 5
06/11/14
8454
JFVM 5 Software
06/11/14
8455
Study Text of ISO/IEC 14496-10:2005/FPDAM3 Scalable Video 06/11/10
Coding
8456
Joint Scalable Video Model (JSVM) 8
06/12/08
8457
JSVM 8 Software
07/01/05
8458
Working Draft 1 of ISO/IEC 14496-10:2005/Amd.4 Multiview
06/11/10
Video Coding
8459
Joint Multiview Video Model (JMVM) 2
06/11/10
199
8460
JMVM 2 Software
06/11/17
JVT Meeting 21 WG11 Resolution 2: The JVT and the WG11 video subgroup thank the
WG11 National Bodies of Germany, Japan, Netherlands, Ukraine and US for their ballot
comments on ISO/IEC 14496-10:2005/FPDAM2.
JVT Meeting 21 WG11 Resolution 3: The JVT and the WG11 video subgroup request the
WG11 National Bodies to kindly consider the Study Document N8452 in their upcoming
ballot votes on ISO/IEC 14496-10:2005/FPDAM2.
JVT Meeting 21 WG11 Resolution 4: The JVT and the WG11 video subgroup request
National Bodies to kindly consider the Study Document N8455 and JSVM text N8456 in
their upcoming ballot votes on ISO/IEC 14496-10:2005/FPDAM3.
JVT Meeting 21 WG11 Resolution 5: The JVT and the WG11 Video and Test subgroups
recommend approval of the following document.
No.
Title
Available
14496-10 Advanced Video Coding
8553 Draft SVC Verification Test Plan Version 2
06/11/10
200
JVT Meeting 21 WG11 Resolution 6: The JVT provides the following list of JVT ad hoc
groups appointed to progress work in the interim period until the next JVT meeting.
Title and Email Reflector
Chairs
Mtg
JVT Project Management and Errata Reporting
Gary Sullivan,
N
(jvt-experts@lists.rwth-aachen.de)
Jens Rainer Ohm, Ajay Luthra,
and Thomas Wiegand
JM Text and Reference Software
Thomas Wiegand,
N
(jvt-experts@lists.rwth-aachen.de)
Karsten Sühring,
Alexis Tourapis, and
Keng Pang Lim
Bitstreams and Conformance
Teruhiko Suzuki
N
(jvt-bitstream@lists.rwth-aachen.de)
Professional applications
Teruhiko Suzuki
N
(jvt-experts@lists.rwth-aachen.de)
JSVM software and new functionality integration Jerome Vieron, Mathias Wien,
N
(jvt-svc@lists.rwth-aachen.de)
Heiko Schwarz
JSVM and JD Text Editing
Julien Reichel, Heiko Schwarz,
N
(jvt-svc@lists.rwth-aachen.de)
Mathias Wien,
SVC Spatial Scalability, Resampling and InterShijun Sun, Andrew Segall
N
layer Prediction
(jvt-experts@lists.rwth-aachen.de)
SVC High-Level Syntax and Error Resilience
Ye-Kui Wang, Stéphane Pateux, N
(jvt-experts@lists.rwth-aachen.de)
Peter Amon, Thomas Schierl
SVC Interlaced Coding
Jerome Vieron
N
(jvt-svc@lists.rwth-aachen.de)
SVC Quantization, CAVLC and CABAC
Justin Ridge, Detlev Marpe,
N
(jvt-svc@lists.rwth-aachen.de)
Gary Sullivan
SVC Complexity Reduction
Heiko Schwarz, Yiliang Bao
N
(jvt-svc@lists.rwth-aachen.de)
SVC residual prediction modification
Yiliang Bao
N
(jvt-svc@lists.rwth-aachen.de)
SVC enhanced spatial scalability
Jerome Vieron
N
(jvt-svc@lists.rwth-aachen.de)
SVC bit depth and chroma format scalability
Yongying Gao, Andrew Segall,
N
(jvt-svc@lists.rwth-aachen.de)
Thomas Wiegand
MVC High-level syntax and buffer management
Anthony Vetro, Yeping Su
N
(jvt-mvc@lists.rwth-aachen.de)
JMVM and JD text editing
Hideaki Kimata, Aljoscha
N
(jvt-mvc@lists.rwth-aachen.de)
Smolic, Yeping Su, Anthony
Vetro
JMVM software and new functionality
Purvin Pandit, Anthony Vetro
N
integration
(jvt-mvc@lists.rwth-aachen.de)
AHG on video annotation
Jens-Rainer Ohm, Thomas
N
(jvt-experts@lists.rwth-aachen.de)
Wiegand
JVT Meeting 21 WG11 Resolution 7: The JVT chairmen propose to hold a JVT meeting
during 13-19 January 2007 under WG 11 auspices in Marrakech, Morocco. Further
meetings are proposed to be held during April 2007 under WG 11 auspices in San José, US,
201
during the first week of July under the auspices of the meeting of ITU-T SG 16 in Geneva,
CH, and during October 2007 under WG 11 auspices in Shenzhen, CN.
Addendum: The JVT chairmen note the following related liaison outputs from the WG11
parent body.
No.
Title
Available
Liaison Statements
Liaison Statement to ITU-R SG6 WP 6J concerning colour space
06/10/27
8529
amendments
Liaison Statement to SMPTE on 4:2:2 and 4:2:0 Intra-only profiles 06/10/27
8532
of AVC
8533 Liaison Statement to SMPTE on 4:4:4 Intra-only profile of AVC
06/10/27
8537 Liaison Statement to ITU-T SG 9 concerning FTV and MVC
06/10/27
aa. JVT Attendance
Persons registered to attend the JVT meeting, as recorded by a sign-in sheet circulated during the
meeting, were the following (195 listed participants):
1.
Gary Sullivan (Microsoft Corp.)
2.
Jens-Rainer Ohm (RWTH Aachen Univ.)
3.
Yun He (Tsinghua Univ.)
4.
Gang Zhu (Tsinghua Univ.)
5.
Ping Yang (Tsinghua Univ.)
6.
Xiaozhong Xu (Tsinghua Univ.)
7.
Zhijie Yang (Broadcom)
8.
Yung-Lyul Lee (Sejong Univ.)
9.
Jae-Ho Hur (Sejong Univ.)
10. Sung Chang Lim (Sejong Univ.)
11. Dongkyun Kim (Sejong Univ.)
12. Dae-Yeon Kim (Sejong Univ.)
13. Jae-Gon Kim (ETRI)
14. Jie Jia (Sejong Univ.)
15. Jung Won Kang (ETRI)
16. Xianglin Wang (Nokia)
17. Ying Chen (Tampere Univ. Tech.)
18. Lulin Chen (Omneon Video Networks USA)
19. Truong Cong Thang (ICU)
20. Jun Zhang (Huawei Tech.)
21. Yeping Su (Thomson USA)
22. Kemal Ugur (Nokia)
23. Jesus Sampedro (Polycom)
24. Hiroya Nakamura (JVC)
25. Takashi Itoh (Fujitsu Labs)
26. Yukihiro Bandoh (NTT)
27. Hideaki Kimata (NTT)
28. Chang-Won Seo (Sejong Univ.)
29. Sang-mi Kim (Sejong Univ.)
30. Steffen Wittmann (Panasonic)
31. Akiyuki Tanizawa (Toshiba)
32. Takeshi Chujoh (Toshiba)
33. Masato Shima (Texas Instruments Japan)
34. Kyung-Jun Lee (Kyung Hee Univ.)
202
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
Jaeyull Oh (Kyung Hee Univ.)
Toshiyaki Fujii (Nagoya Univ.)
Shawmin Lei (Sharp Labs USA)
Andrew Segall (Sharp Labs USA)
Zhongkang Lu (Inst. for Infocomm. Research)
Arnaud Bourge (Philips / NXP)
Patrice Onno (Canon France)
Nathalie Cammas (Orange – France Telecom.)
Han-Suh Koo (LG Electronics)
Sang-Heon Lee (Seoul Natl. Univ.)
Seung-Wook Park (LG Electronics)
Won Seon Song (Soongsil Univ.)
Kwon Yul Choi (Soongsil Univ.)
Yeong Gyoo Jeon (Soongsil Univ.)
Seishi Takamura (NTT)
Onur G. Guleryuz (Docomo USA Labs)
Xiangyang Ji (CAS ICT)
Kyoung Hwan Kim (Soongsil Univ.)
Ki Beom Kim (Soongsil Univ.)
Peter Amon (Siemens AG)
Yong Yan (Freescale)
Jones He (Freescale)
Jan De Cock (Ghent Univ.)
Davy De Schryver (Ghent Univ.)
Saar De Zutter (Ghent Univ.)
Jong-Ki Han (Sejong Univ.)
Gwang-Hoon Park (Kyung Hee Univ.)
Lianhuan Xiong (Huawei)
Seyoon Jeong (ETRI)
Min-woo Park (Kyung Hee Univ.)
Seong-seon Baek (Kyung Hee Univ.)
Won-Jun Choi (Kyung Hee Univ.)
Yong-Hun Lee (Kyung Hee Univ.)
Dae-Yeon Kim (Kyung Hee Univ.) [apparently not the same person as entry 12]
Bae-Keun Lee (Samsung Electronics)
Woo-Sung Shim (Samsung Electronics)
Song Rae Lee (Samsung Electronics)
Yan Ye (Qualcomm)
Yiliang Bao (Qualcomm)
Yingyong Qi (Qualcomm)
Bumshik Lee (ICU)
Jeongyeon Lim (ICU)
Junyan Huo (Xidian Univ.)
Haitao Yang (Xidian Univ.)
Xiaozhen Zheng (Huawei)
Sixin Lin (Huawei)
Pengxin Zeng (Huawei)
Byeong Moon Jeon (LG Electronics)
Jizheng Xu (Microsoft)
Thomas Wiegand (Fraunhofer HHI)
Heiko Schwarz (Fraunhofer HHI)
Mathias Wien (RWTH Aachen Univ.)
Steffen Kamp (RWTH Aachen Univ.)
203
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
98.
99.
100.
101.
102.
103.
104.
105.
106.
107.
108.
109.
110.
111.
112.
113.
114.
115.
116.
117.
118.
119.
120.
121.
122.
123.
124.
125.
126.
127.
128.
129.
130.
131.
132.
133.
134.
135.
136.
137.
138.
139.
Masayuki Tanimoto (Nagoya Univ.)
Vincent Bottreau (Thomson R&D France)
Tomoyuki Yamamoto (Sharp)
Kwan Jung Oh (GIST)
Yo-Sung Ho (GIST)
Satoru Sakazume (JVC)
Kazuhiro Shimauchi (JVC)
Jeong-Hyu Yang (LG Electronics)
Takahiro Kimoto (NEC)
Shankar Regunathan (Microsoft)
Kenneth Andersson (Ericsson)
Jerome Vieron (Thomson R&D France)
Lu Yu (Zhejiang Univ.)
Ye-Kui Wang (Nokia)
Mike Nilsson (BT)
Teruhiko Suzuki (Sony)
Yongjoon Jeon (LG Electronics)
Stephane Pateux (Orange – France Telecom)
Shijun Sun (Microsoft)
[end of list as of Friday 20 October]
Justin Ridge (Nokia)
Donggyu Sim (Kwangwoon Univ.)
Seanae Park (Kwangwoon Univ.)
Junghak Nam (Kwangwoon Univ.)
Je Woo Kim (Korea Electronics Tech. Inst. - KETI)
Byeongho Choi (KETI)
Yong-Hwan Kim (KETI)
Jungyoup Yang (SKKU)
Yongying Gao (Thomson)
Quqing Chen (Thomson)
Zhibo Chen (Thomson)
Zhengguo Li (I2R)
Yih Han Tan (I2R)
Wei Yao (I2R)
Marta Karczewicz (Qualcomm)
Thiow Keng Tan (NTT DoCoMo)
Sunil Lee (KAIST)
Dalwon Jang (KAIST)
Chang Yoo (KAIST)
Kyuheon Kim (Kyunghee Univ.)
Zhibo Ni (Zhejiang Univ.)
Dandan Ding (Zhejiang Univ.)
Leszek Cieplinski (Mitsubishi Electric)
Faisal Ishtiaq (Motorola)
Shih-Ta Hsiang (Motorola)
Peng Yin (Thomson)
Lihua Zhu (Thomson)
Chong Soon Lim (Panasonic)
Sebastien Branguolo (SSM)
Weimin Zeng (Micronas USA)
Tomokazu Murakami (Hitachi)
Shun-ichi Sekiguchi (Mitsubishi)
Dae-Sung Cho (Samsung AIT)
204
140.
141.
142.
143.
144.
145.
146.
147.
148.
149.
150.
151.
152.
153.
154.
155.
156.
157.
158.
159.
160.
161.
162.
163.
164.
165.
166.
167.
168.
169.
170.
171.
172.
173.
174.
175.
176.
177.
178.
179.
180.
181.
182.
183.
184.
185.
186.
187.
188.
Thomas Wedi (Panasonic)
John Wus (Panasonic)
Jaewoo Jung (Samsung AIT)
Yoshihisa Yamada (Mitsubishi)
Tokumichi Murakami (Mitsubishi)
[end of list as of Saturday 21 October]
Haoping Yu (Thomson)
Per Fröjdh (Ericsson)
Huifang Sun (Mitsubishi)
Tokuyo Kogure (Univ. Tokyo)
Kohtaro Asai (Mitsubishi)
Munchurl Kim (Info & Comm. Univ. KR)
Peter List (Deutsche Telekom)
Fons Bruls (Philips)
Satoshi Hasuo (Oki)
Lowell Winger (LSI Logic)
Thomas Rathgen (Ilmenau Univ.)
Shantanu Rane (Stanford Univ.)
Yi-Shin Tung (Setabox Tech. Corp.)
Minhua Zhou (Texas Inst.)
Anthony Vetro (Mitsubishi Electric)
Tobias Oelbaum (Tech. Univ. Munich)
Sehoon Yea (MERL)
Barry Haskell (Apple Computer)
Hideki Ohtaka (Matsushita Electric)
Michael Horowitz (CoVi Tech.)
Wen Hsiao Peng (Samsung AIT)
Jungdong Seo (Yonsei Univ.)
Jan Lievens (Vrije Univ. Brussels)
Sei Naito (KDDI)
[end of list as of Sunday 22 October]
[no meetings Monday 23 October]
Peter Symes (Thomson)
Alex Eleftheriadis (Layered Media)
Yuwen Wu (Thomson)
Song-Heon Lee (Seoul Natl. Univ.)
Kang-Jae Chung (LG Electronics)
Yi-Jen Chiu (Intel)
Alexandros Tourapis (Dolby Labs)
Xhixiong Wu (Oki)
Doug Young Suh (KHU)
Gisle Bjøntegaard (Tandberg)
Livio Lima (Univ. Brescia)
Silxiou Simbotelecan (VUB)
Min-Cheol Hong (Soongsil Univ.)
Hae-Chul Choi (ETRI)
Hae Kwang Kim (Sejong Univ.)
[end of list as of Tuesday 24 October]
Arkady Kopansky (Sarnoff)
Arild Fuldseth (Tandberg)
Young-Hoon Cho (Dongguk Univ.)
Matthias Narroschke (Univ. Hannover)
Sung Min Kim (Dongguk Univ.)
205
189.
190.
191.
192.
193.
194.
195.
Pankaj Topiwala (FastVDO)
Ping Wu (Tandberg TV)
Xin Jin (Huazhong Univ. of Sci. & Tech.)
Herbert Thoma (Fraunhover IIS)
Marina Bosi (MPEG LA, LLC)
Jing Wang (Huawei)
Joern Ostermann (Univ. Hannover)
[end of list]
206
Annex I – Audio report
Source: Schuyler Quackenbush, Chair, Audio Subgroup
1
2
Opening of the meeting ......................................................................................................... 208
Administrative matters .......................................................................................................... 208
2.1 Approval of previous meeting report 208
2.2 Approval of agenda and allocation of contributions 208
2.3 Task Groups 208
2.4 Communications from the Chair
208
2.5 Joint meetings 208
2.6 Received National Body Comments and Liaison matters 208
3 Record of AhG meetings ....................................................................................................... 209
4 Audio plenary, joint meeting and task group activities ......................................................... 209
4.1 Review of AHG reports
209
4.2 Received national body comments and liaison matters
209
4.3 Joint Meetings 209
4.3.1 Systems, MDS on Archival MAF and support for large files ..................................... 209
4.3.2 Requirements on SAOC .............................................................................................. 210
4.4 Task Group discussions
210
4.4.1 MPEG-4 Audio............................................................................................................ 210
4.4.2 MPEG Surround and MPEG Surround next steps ...................................................... 213
4.4.3 Exploration of Speech and Audio ............................................................................... 215
4.4.4 Symbolic Music Representation - Pierfrancesco Bellini ............................................ 215
4.5 Audio closing plenary discussions 215
5 Meeting deliverables ............................................................................................................. 216
5.1 Recommendations for final plenary 216
5.2 Establishment of Ad-hoc Groups
216
5.3 Approval of output documents
216
5.4 Responses to Liaison and NB comments
216
5.5 Press statement
216
6 Future activities ..................................................................................................................... 216
6.1 Schedule of future meetings 216
6.2 Agenda for next meeting
216
6.3 All other business
216
6.4 Closing of the meeting
216
Annex A Participants ............................................................................................................... 217
Annex B Audio Contributions and Schedule .......................................................................... 219
Annex C Task Groups ............................................................................................................. 223
Annex D Output Documents ................................................................................................... 224
Annex E Agenda for the 79th MPEG Audio Meeting ............................................................ 226
207
1
Opening of the meeting
The MPEG Audio Subgroup meeting was held during the 78th
meeting of WG11, October
23-27, Hangzhou, China. The list of participants is given in Annex A.
2. Administrative matters
a.
Approval of previous meeting report
The 77th
approved.
Audio Subgroup meeting report was registered as a contribution, and was
b.
Approval of agenda and allocation of contributions
The agenda and schedule for the meeting was discussed, edited and approved. It shows the
documents contributed to this meeting and presented to the Audio Subgroup, either in the task
groups or in Audio plenary. The Chair brought relevant documents from Requirements, Systems
and MDS to the attention of the group. It was revised in the course of the week to reflect the
progress of the meeting, and the final version is shown in Annex B.
c.
Task Groups
Task groups were convened for the duration of the MPEG meeting, as shown in Annex C.
Results of task group activities are reported below.
d.
Communications from the Chair
The Chair summarised the issues raised at the Sunday evening Chair’s meeting, proposed task groups for the week, and proposed agenda items for
discussion in Audio plenary.
e.
Joint meetings
The joint meetings with Audio over the course of the week are listed here and are reported on
below.
Groups
What
Where Day Time
Audio, MDS
Audio Archival MAF, 13881 Music Audio Wed 1400Player, 13913, Reference Software
1500
Audio, Req
Review requirements
Audio Thu 14001430
f.
Received National Body Comments and Liaison matters
The NB Comments and Liaison documents for the meeting that require a response are as shown
below.
No.
Title
Synopsis
Response
by
USNB Contribution:
Speech and Audio
13866
SRQ
Coding Exploration
Support
208
13417
(to 77th
MPEG
meeting)
Liaison Statement from
ITU-T SG 16 [SC 29 N
7473]
13808
Liaison Statement from
ITU-R SG 6/WP 6Q [SC
29 N 7794]
Outgoing
Audio Subgroup will send a
revised spreadsheet.
Attached please find a modified
Version of the ITU-T SG16 Q.23
MediaCoding Summary Database.
The major changes are new
SRQ
columns for some MPEG-4 Audio
Object types and profiles and
capturing the eaac+ column from
3GPP. Modified or added entries
are highlighted in yellow.
ITU-R WP 6Q currently is
working towards an extension of
Recommendation ITU-R
BS.1387-1 to address the
SRQ
measurement of multi-channel
audio signals. Final call for
proposals and time schedule
included
SCTE new audio technology from
WG11 may be useful and cite
SRQ
other bodies deploying technology
as illustration
3. Record of AhG meetings
There were no AhG meetings prior to this MPEG meeting.
4. Audio plenary, joint meeting and task group activities
a.
Review of AHG reports
There were no requests to review any of the AHG reports.
b.
Received national body comments and liaison matters
One national body comment and one liaison documents were reviewed and the drafting of the
responses was delegated.
c.
i.
Joint Meetings
Systems, MDS on Archival MAF and support for large files
Wed 14001500
Joint with MDS at Audio
Audio Archival MAF, 13881
Music Player, 13913, Reference Software
Harald Fuchs, FhG, reviewed the status of the Second Edition of the Music Player MAF. The
new document was presented, with a quick review of the changes made in response to ballot
comments.
There was considerable discussion on conformance, and it was agreed that conformance data
consists of
 Complete files containing media
 Constituent compressed media from those files
 Optionally, the fully decoded media for audio decoders (i.e. a reference decoded
waveform).
209
MAF conformance consists of extraction of specified constituent data elements, such as
compressed audio data or XML data. It should be indicated that conformance of this data is via
referencing other MPEG conformance specifications.
We reviewed reference software, and a timetable was presented to get final reference code for the
unprotected and default protection modes available by next MPEG meeting, and full reference
code by the end of the year.
Noboru Harada, reviewed the status of the Audio Archival MAF. He raised a number of open
issues associated with the Audio Archival MAF. The group agrees that the work should be
partitioned between Audio, MDS or File Format experts, and Audio experts worry about audio
issues.
It was noted that the MAF should not create specification, but rather reference MPEG
specifications.
Chris Barlas, RightsCom, noted the “Open document format” ISO/IEC 26300 is a zip archive
format that could be referenced by this MAF.
ii.
Requirements on SAOC
Joint with Requirements at Audio
On Requirements for SAOC
Juergen Herre, FhG, presented a draft output document, “SAOC use cases, draft requirements
and architecture.”
This document contains the following use cases
 Interactive re-mix
 Interactive gaming
 Teleconferencing
Thu 1400-1500
He reviewed each use case and its associated requirements and also presented the architecture for
realizing an SOAC decoder with full “re-use” of the already standardized MPEG Surround
decoder.
The Requirements Chair gave the following comments:
 The document is very complete.
 The document should be precise in its use of terms such as “should” and “shall.”
 The document should present the entire set of requirements in a single section.
He strongly encouraged the Audio Subgroup to make an open Call for Proposals for technology
meeting these requirements. This could be issued as a Preliminary Call from this meeting and a
Final Call from the 79th meeting.
The Audio Subgroup will discuss this proposal and decide how to proceed.
d.
i.
Task Group discussions
MPEG-4 Audio
JungHoe Kim, Samsung, presented
JungHoe Kim
Proposed updates on SLS reference software with
13900
Eunmi Oh
ER BSAC
This notes that
 14496-3 specifies BSAC as a core coder for SLS, but this is not implemented in the
reference software.
 Currently, the SLS reference software only supports mono and stereo.
 Currently, the SLS reference software core encoder does not window switch, although the
decoder does support window switching.
Samsung has offered to contribute BSAC reference software to support of BSAC in SLS.
This will be discussed later in the week to resolve the latter two bullet items. The Chair suggests
that a workplan be drafted to clarify coordinating this work.
210
On Thursday the status of the SLS reference software was clarified. It was decided that a
workplan will be drafted to progress this work.
JungHoe Kim, Samsung, presented
JungHoe Kim
Proposed study on 14496-4:2004/FPDAM 14,
13901
KangEun Lee
BSAC Conformance
Eunmi Oh
This defines new bitstreams for BSAC conformance. It is anticipated that this “Proposed Study
on” can become the FPDAM text pending a careful review of the ballot comments. The Chair
recommended that JungHoe Kim create a separate “Status of” document for the Amd 14
conformance effort.
JungHoe Kim, Samsung, presented
JungHoe Kim
Proposed changes for BSAC Extensions combined
13902
Eunmi Oh
with MPEG Surround
This contribution defines a method to embed MPEG Surround data in a BSAC bitstream. This
could be included in ISO/IEC 14496-3:2005/FPDAM 5, BSAC Extensions. However, it was
noted that this mode of carriage of MPEG Surround data does not support scalability of the
multichannel output, i.e. the MPEG Surround data is the first component of the bitstream that is
lost in scaling.
Scalability is achieved only if one uses multiple elementary streams, one for BSAC and another
for MPEG Surround.
The Audio Subgroup agreed to incorporate the proposed changes in to a Study on ISO/IEC
14496-3:2005/FPDAM 5, BSAC Extensions.
Noboru Harada, NTT, presented
Noboru Harada
Proposed text to MPEG-4 audio extensions for 6413880
Takehiro Moriya
bit address space file format support
Yutaka Kamamoto
This contributions proposed a new box for the MPEG-4 file format that supports 64-bit reference
addresses to point to original audio file header, trailer and “aux” items. This would require using
the MPEG-4 File Format registration authority as a means to reference the definition
specification, which is envisioned to be in MPEG-4 subpart 1.
In addition, it is suggested that the ALS specification be extended to provide a mode that
removes the redundancy between the ALS functionality and the MPEG-4 File Format
functionality.
The Audio Subgroup agreed to make the proposal a WD on a new amendment to MPEG-4 Audio.
Ralph Sperschneider, FhG, presented
Ralph Sperschneider Conformance issues regarding AAC utilizing the
13917
Michael Matejko
LTP tool
This contribution raises some issues concerning the LTP tool, specifically that it appears to not be
deterministic to the extent that it can be guaranteed to deliver PCM words with at least N (e.g.
15) bits that match the conformance reference waveform. It was the consensus of the Audio
Subgroup to that the Chair will send an email to experts at Nokia to ask for a proposed solution to
this conformance problem, to be delivered prior to the 80th MPEG meeting.
Juergen Herre, FhG, presented
Markus Schnell
Ralph Sperschneider
Markus Schmidt
13958
Juergen Herre
Proposal for an Enhanced Low Delay Coding Mode
Ralf Geiger
Gerald Schuller
Manfred Lutzky
211
This contribution reviewed Low Delay AAC, and noted that it has recently enjoyed considerable
success in the marketplace, in part due to the fact that it delivers a wideband signal and does so
without any signal model, making it robust to e.g. speech doubletalk or speech with music or
noise as a background signal. It notes that it could be more successful if it delivered greater
compression efficiency. The contribution proposes to achieve this by combining Low Delay
AAC with the SBR tool, such that the combination achieves 1/3 reduction in bitrate with only
moderate increase in system throughput delay (from 30 ms to 42 ms when the input is sampled at
48 kHz). The contribution notes that G.722.1 Annex C has 40 ms algorithmic delay, and G.729.1
has 48 ms algorithmic delay.
Adding SBR as an additional coding tool delivers significant coding efficiency. However, adding
SBR as a post-processor typically incurs a significant additional delay. To mitigate the latter
problem, it is proposed that the filterbank of AAC LD be changed in a way that the combined
tool has minimum delay. The specific changes proposed are:
 SBR look-ahead be prohibited from crossing frame boundaries. This has no impact on
SBR syntax structure and algorithm, and only a minor impact on syntax and semantics.
 AAC LD use a different window function for its MDCT/IMDCT. This window has two
new characteristics
o The window has a zero interval at the leading edge, which leads to lower
throughput delay
o The window has a very long “tail” at the trailing edge, which leads to improved
frequency selectivity.
The contribution also presented
 Subjective test results for the proposal
 Complexity of the proposal as compared to MPEG-4 AAC LD and MPEG-2 AAC LC
profile.
FT has volunteered to do a cross-check on the subjective results presented in this contribution,
and may be able to deliver this during the week.
Hari Garudadri, Qualcomm, asked about system delay when using the proposal over IP channels
in which jitter buffers are required.
The contribution listed 7 companies that currently use AAC LD technology in their products, or
are interested in the proposed technology for future products.
It was desired that there be some additional discussion “off-line” on the proposal, with the
possible actions being:
 Should this form the basis of an amendment to MPEG Audio?
 What point in the standardization does the amendment launch?
In additional discussion on Thursday, it was the consensus of the Audio Subgroup that this
technology be used to launch a new amendment to MPEG-4 Audio, and that this amendment will
start at the CD phase.
It is noted that France Telecom experts object to the creation of a PDAM on enhanced AAC Low
Delay at this meeting, based on the following statements:
 France Telecom's opinion is that evidence of the merits of the technology shall be
formally assessed before launching any new amendment. As far as compression
efficiency is addressed, expected performance shall be carefully quantified with that
respect.
 This assessment shall include items specific to the envisioned application: in that context,
communication applications being addressed, France Telecom experts urges that a
substantial number of speech items to be used in that assessment.
 In any case, France Telecom experts think that the normal procedure is to launch any new
amendment activity as Working Draft, before entering the PDAM state.
212
It is noted that there were several experts that were interested in additional evidence of
performance when using only speech items. It is the expectation of the Audio Subgroup that this
evidence will be available at the next meeting.
ii.
MPEG Surround and MPEG Surround next steps
1. MPEG Surround
Pierrick Philippe, France Telecom, presented
Report on the pre-selection process for MPEG
Pierrick Philippe
14009
Surround verification tests
David Virette
This contribution presented the results of work done during the AhG period relating to selection
of material for the MPEG Surround Verification Test. The work done is summarized here
 Use items from NBC test
 Collect new items from Univ. of Dusseldorf, Philips and France Telecom
 Identify new HRTF from Philips
All items were limited to less than 20 seconds in length and converted to extensible WAV format.
These were then encoded using AAC and HE-AAC. The coded items were rated by expert
listeners, with only one coder considered per listening session. Defects were identified as being
from a list of 10 possible categories.
The contribution noted several guidelines in selection of material, for example, that the items
used to develop MPEG Surround not be part of the Verification Test items.
Kristofer Kjörling , Coding Technologies, presented
Kristofer Kjörling
Jonas Rödén
Further revision of the verification test proposal for
13923
Heiko Purnhagen
MPEG Surround
Werner Oomen
Johannes Hilpert
This contribution proposes a number of changes to the verification test plan. These proposals will
be incorporated by the task group into a draft workplan for subsequent review by the group.
Werner Oomen, Philips, gave a short presentation on a candidate HRTF for use in the verification
test. This HRTF was made in a room with ITU 5.0 loudspeaker setup. The impulse response was
captured to 4096 samples using in-ear microphones and an individual’s head.
Post-processing consisted of
 Early arrival was separated from reverberant part
 Reverberant part of center channel was used to replace the reverberant part of all other
HRTFs.
 Some manual equalization was applied
 Impulse responses were truncated to 2048 taps
Heiko Purnhagen, Coding Technologies, presented
Heiko Purnhagen
13925
Update on conformance testing for MPEG Surround
Andreas Schneider
There was discussion as to how to organize the entire 23003 specification. It was agreed that
MPEG Surround specification, conformance and reference software shall be contained in three
separate MS Word files. There was no strong preference as to whether the MPEG Surround
specification, conformance and reference software all be in part 1, or should they be three
separate parts.
The contribution proposed
 Restrictions on bitstreams
 Definition of bitstreams
 Conformance procedure.
213
and will be used to produce a PDAM document at this meeting. The Chair suggested that, when
appropriate, MPEG-4 file format be used for the conformance data.
Heiko Purnhagen, Coding Technologies, presented
13924
Heiko Purnhagen
Update on reference software for MPEG Surround
This contribution raises the same issues as to “subpart” or part of standard. It proposes that this
be a stand-alone software repository.
Heiko Purnhagen, Coding Technologies, presented
13926
Heiko Purnhagen
Update on transport of MPEG Surround
This contribution proposes some minor bug-fixes for the transport of MPEG Surround in MPEG2 AAC and MPEG-4 BSAC extensions.
JungHoe Kim, Samsung, presented
JungHoe Kim
Proposed residual coding with ER BSAC for MPEG
13904
Eunmi Oh
Surround
This contribution noted that if MPEG Surround using residual coding is combined with an
MPEG-4 ER BSAC coded downmix, then the decoding system requires an ER BSAC decoder,
an MPEG Surround decoder and an MPEG AAC decoder (for residual decoding). If residuals are
coded with BSAC, then a savings in memory storage can be realized.
Werner Oomen, Philips, noted that the MPEG Surround residual coder is actually a simplified
version of MPEG AAC. He further noted that the proposal raises issues of interoperability, in that
MPEG Surround information is applicable to any base coder. Jonas Rödén, Coding Technologies,
noted that if one adds the capability to use BSAC for residual coding, then one would still have to
implement an AAC residual decoder, so that this savings in memory storage actually is not
realized.
It was the consensus of the Audio Subgroup to not adopt this proposal.
2. SAOC
Seungkwon Beack, ETRI, presented
Seungkwon Beack
Jeongil Seo
13899
Taejin Lee
Further information of a new application for SAOC
Inseon Jang
Dae-young Jang
This contribution presented a number of potential applications for SAOC, and presented
associated requirements for the set of use cases. ETRI has a number of demonstrations of these
applications.
Juergen Herre, FhG, presented
Juergen Herre
13935
Werner Oomen
Thoughts on an SAOC Architecture
Kristofer Kjoerling
This contribution reviewed the MPEG Surround architecture and the proposed SAOC
architecture. It proposed three broad categories of applications:
 Backward compatible interactive re-mix
 Gaming and rich media
 Teleconferencing
In addition, it presented a table in which requirements for each application are indicated.
A second section of the contribution presented a possible architecture for using elements of
MPEG Surround to accomplish the goals of the example applications. Major issues are:
 MPEG Surround is a sophisticated rendering engine.
 It is most computationally efficient if MPEG Surround’s outputs are not “objects” but
rather the target loudspeaker output signals.
214

SAOC bitstreams are agnostic to final loudspeaker presentation.
Hence SAOC can be viewed as a “spatial information transcoder” from SAOC object-based
format to MPEG Surround loudspeaker (i.e. rendering) based format. This transcoder maps N
objects to M rendered output channels. The transcoder affects only the parameters, and does not
have to touch the downmix signal. Note that the number and position of the output channels
(i.e.loudspeakers) and the position of the objects in the rendered acoustic space are “playback
parameters,” that are set interactively by the user to reflect the local decoder configuration.
The architecture requires the normative definition of
 SAOC bitstream
 SAOC-to-MPEG-Surround transcoding engine
 Rendering matrix for the engine interface
An informative Annex would give an example of to derive the normative rendering matrix from
the position of the output channels (i.e.loudspeakers) and the position of the objects in the
rendered acoustic space. The contribution notes that the object positions could be delivered via
LASeR or BIFS so that the SAOC/MPEG Surround engine could support interactive control in a
standardized method
3. Scene Control
Marc Emerit, France Telecom, presented
A survey of audio middleware parameters for Audio
Scene Control reusing MPEG Surround
This contribution presented a survey of the spatialization parameters used by rendering engines
that are widely used in the marketplace. It concluded with a recommended minimum set of
parameters and functionality to support the envisioned functionality.
The group agreed that this needs additional discussion.
4. Additional Discussion
14008
Marc Emerit
Juergen Herre, FhG, presented a draft output document that captures
 Use Cases with goals and draft requirements for each case
 Architecture for SAOC
The Architecture section showed an architecture block diagram that illustrates the relationship
between MPEG Surround specification and the proposed new work.
Issues that must be clarified as a means to understand next steps are
 Requirements for new work
 Scope of new work
 Process for conducting new work
These will be discussed in the joint meeting with Requirements.
iii.
Exploration of Speech and Audio
The Chair presented a draft workplan that specifies work to be done during the next AhG period
to support this exploration effort. There was considerable discussion, but in the end, a workplan
was produced that had the consensus of the Audio Subgroup.
iv.
Symbolic Music Representation - Pierfrancesco Bellini
The SMR breakout accomplished the following at the 78th MPEG meeting:
 Responded to the NB comments
 Prepared the text for the new MPEG-4 part 23
 Prepared the WD on SMR reference software (submitted to Systems)
 Discussed and finalized with systems the integration of SMR in BIFS
e.
Audio closing plenary discussions
There was some additional discussion and editing of some documents prior to approval, including
215


Workplan for Speech and Audio Exploration
Draft Call for Proposals on Spatial Audio Object coding
5. Meeting deliverables
a.
Recommendations for final plenary
The Audio recommendations were presented and approved.
b.
Establishment of Ad-hoc Groups
The following ad-hoc groups were established by the Audio subgroup:
No.
Title
8643
AHG on Audio Standards Maintenance
8644
AHG on Exploration of Speech and Audio Coding
8645
AHG on MPEG Surround Verification Test and SAOC CfP
c.
Mtg
No
Yes
Yes
Approval of output documents
All output documents, shown in Annex D, were presented in Audio plenary and were approved.
d.
Responses to Liaison and NB comments
The responses to Liaison and NB comments were prepared and approved.
e.
Press statement
The Audio part of the press statement was prepared and approved.
6. Future activities
a.
Schedule of future meetings
Ad Hoc group meetings are indicated in Section 5.b. Unless otherwise indicated, Ad Hoc group
meetings will be held at the location of the next MPEG meeting on the weekend preceding that
meeting.
b.
Agenda for next meeting
The agenda for the next MPEG meeting is shown in 0.
c.
All other business
There was none.
d.
Closing of the meeting
The 78th Audio Subgroup meeting was adjourned Friday at 12:30 (this could be a record)!
216
Annex A Participants
First Name
Seungkwon
Johannes
Shuixian
Last Name
Beack
Boehm
Chen
Country
KR
DE
CN
Sang Bae
Chon
KR
Zhengzhong Du
CN
Marc
Bernhard
Noboru
Oliver
Jürgen
Huan
Haibin
Yang-Won
Junghoe
Minsoo
Emerit
Feiten
Harada
Hellmuth
Herre
Hou
Huang
Jung
Kim
Kim
FR
DE
JP
DE
DE
CN
SG
KR
KR
KR
Kristofer
Te
Tilman
Kjörling
Li
Liebchen
S
SG
DE
Hongfei
Ma
CN
Han Gil
Takehiro
Oliver
Toshiyuki
Eunmi
Henney
Moon
Moriya
Niemeyer
Nomura
Oh
Oh
KR
JP
DE
JP
KR
KR
Werner
Oomen
NL
Pierrick
Philippe
FR
Heiko
Fang
Schuyler
Susanto
Purnhagen
Qin
Quackenbush
Rahardja
SE
CN
USA
SG
Jonas
Rödén
SE
Jianye
Rong
CN
Andreas
Schneider
DE
Affiliation
ETRI
Thomson
Wuhan Univ.
Seoul National
Univ.
Huawei
Technologies
France Telecom
R&D
Deutsche Telekom
NTT
Fraunhofer IIS
Fraunhofer IIS
Tsinghua Univ.
I2R
LG Electronics
Samsung AIT
Pixtree
Coding
Technologies
I2R
LG Electronics
Huawei
Technologies
Samsung
Electronics
NTT
Thomson
NEC
Samsung
LG Electronics
Philips Applied
Technologies
France Telecom
R&D
Coding
Technologies
SVA
ARL
I2R
Coding
Technologies
Huawei
Technologies
Coding
Technologies
217
Jeongil
Ralph
Anisse
Seo
KR
Sperschneider DE
Taleb
SE
Tinghong
Wang
CN
Wei
Xiao
CN
Lijing
Xu
CN
Jun
Shuhua
Zhang
Zhang
CN
CN
ETRI
Fraunhofer IIS
Ericsson AB
Huawei
Technologies
Huawei
Technologies
Huawei
Technologies
Huawei
Technologies
Tsinghua Univ.
218
Annex B Audio Contributions and Schedule
Number
Author
Title or Activity
X
Monday
0900-1200
MPEG Plenary
1200-1400
Lunch
1400-1800
Audio Plenary
Welcome
Approval of agenda and allocation of contributions
Communications from the Chair
Sunday Chairs meeting
Conformance and Software Assest
Joint meetings
Review of task groups and mandates
13922
S. Quackenbush
78th MPEG Audio Tasks
X
All
Audio
X
All
Audio
All
Audio
MPEG-4
Audio
MPEG-4
Audio
Approval of previous meeting report
13921
S. Quackenbush
77th MPEG Audio Report
Review of AhG reports
13752
R. Sperschneider
AHG on Audio Standards Maintenance
X
13753
S. Quackenbush
AHG on Exploration of Audio Spatialization and
Speech and Audio Coding
X
Summary of Voting
13814
SC 29 Secretariat
Summary of Voting on ISO/IEC 138184:2004/Amd.2:2005/DCOR 1 Additional audio
conformance test sequences
13815
SC 29 Secretariat
Summary of Voting on ISO/IEC 13818-7:2006/PDAM
1 Transport of MPEG Surround in AAC
13816
SC 29 Secretariat
Summary of Voting on ISO/IEC 144964:2004/Amd.11:2006/DCOR 2 Parametric stereo
conformance
13817
SC 29 Secretariat
Summary of Voting on ISO/IEC 14496-4:2004/PDAM
14 BSAC conformance
13818
SC 29 Secretariat
Summary of Voting on ISO/IEC 14496-4:2004/PDAM
18 MPEG-1/2 audio in MPEG-4 conformance
13819
SC 29 Secretariat
Summary of Voting on ISO/IEC 14496-4:2004/PDAM
19 Audio Lossless Coding (ALS) conformance
13820
SC 29 Secretariat
Summary of Voting on ISO/IEC 14496-4:2004/PDAM
20 Scalable to Lossless Coding (SLS) conformance
13830
SC 29 Secretariat
Summary of Voting on ISO/IEC CD 23000-2 [2nd
Edition] MPEG music player application format
NB Comments and Liaison Statements
13866
Andy Tescher for
USNB
1600-1800
USNB Contribution: Speech and Audio Coding
Exploration Support
MPEG-4
13900
JungHoe Kim
Eunmi Oh
Proposed updates on SLS reference software with ER
BSAC
13901
JungHoe Kim
KangEun Lee
Eunmi Oh
Proposed study on 14496-4:2004/FPDAM 14, BSAC
Conformance
X
X
13902
13880
JungHoe Kim
Eunmi Oh
Proposed changes for BSAC Extensions combined with
MPEG Surround
X
MPEG-4
Audio
MPEG4
Audio
Report on the pre-selection process
MPEGfor MPEG Surround verification tests X
D
Audio
Noboru Harada
Takehiro
Proposed text to MPEG-4 audio
Moriya
extensions for 64-bit address space
Yutaka
file format support
Kamamoto
X
Verification Test
14009
Pierrick
Philippe
David Virette
Tuesday
0900-1000
MPEG-4
13917
Ralph
Sperschneider
Michael
Matejko
X
13923
Kristofer
Kjörling
Jonas Rödén
Heiko
Further revision of the verification
Purnhagen
test proposal for MPEG Surround
Werner Oomen
Johannes
Hilpert
13958
Markus Schnell
Ralph
Sperschneider
Markus
Schmidt
Proposal for an Enhanced Low
Juergen Herre Delay Coding Mode
Ralf Geiger
Gerald Schuller
Manfred
Lutzky
Conformance issues regarding AAC
utilizing the LTP tool
1300-1400
Lunch
1400-1500
Workplan for
MPEG Surround Verif.
Test
1500-1700
MPEG-D
MPEG4
Audio
X
MPEGD
Audio
X
MPEG4
Audio
13924
Heiko
Purnhagen
Update on reference software for
MPEG Surround
X
MPEGD
Audio
13925
Heiko
Purnhagen
Andreas
Update on conformance testing for
MPEG Surround
x
MPEGD
Audio
220
Schneider
13926
Heiko
Purnhagen
Update on transport of MPEG
Surround
X
MPEGD
Audio
13904
JungHoe Kim
Eunmi Oh
Proposed residual coding with ER
BSAC for MPEG Surround
X
MPEGD
Audio
X
MPEGD
Audio
X
MPEGD
Audio
MPEGD
Audio
1700-1800
SAOC
13899
Seungkwon
Beack
Jeongil Seo
Taejin Lee
Inseon Jang
Dae-young
Jang
13935
Juergen Herre
Werner Oomen
Thoughts on an SAOC Architecture
Kristofer
Kjoerling
Further information of a new
application for SAOC
1800-1900
Liaison Meeting
1900-
Chairs Meeting
Wednesday
0900-1100
MPEG Plenary
1100-1200
Continue discussion on contribution X
13935
Scene Control
14008
Marc Emerit
A survey of audio middleware
parameters for Audio Scene Control
reusing MPEG Surround
1200-1300
Speech and Audio Exploration
Workplan
1300-1400
Lunch
1400-1500
Joint with MDS at Audio
Audio Archival MAF, 13881
Music Player, 13913, Reference
Software
1500-1600
Speech and Audio Exploration
(continued)
Workplan
1730-
Social
X
X
Thursday
0900-1300
Review NB and Liaison response
221
X
Speech and Audio Exploration
Workplan
X
MPEG Surround next steps
Applications
Requirements
Architecture
Preliminary CfP
X
1300-1400
Lunch
1400-1500
Joint with Req at Audio on SAOC
X
MPEG Surround Verification Test X
Workplan
AAC ELD
FT cross-check
Disposition
1800-
X
Chairs meeting
Friday
0900-1300
Audio plenary
Recommendations for final plenary
X
Establishment of new Ad-hoc groups X
X
AhG Mandates
1000
Get document numbers
Approve Responses to NB comments
Approve Liaison statements
1030
Press statement
Approval of output documents
Review of Audio presentation to
MPEG plenary
Agenda for next meeting
A.O.B.
Closing of the Audio meeting
1300-1400
Lunch
1400-
MPEG Plenary
222
Annex C Task Groups
78th Audio Task Groups
1. MPEG-4 Audio
2. MPEG Surround
3. Exploration of Scalable Speech and Audio
4. Symbolic Music Representation
Mandates for all groups:
1. Review contributions
2. Prepare DoC and Text for milestone documents.
3. Prepare any other documents
Major tasks for the week:
1. MPEG Surround
a. Verification test
b. Next steps
2. Exploration of Speech and Audio
223
Annex D Output Documents
No.
8607
No.
8609
No.
8610
8611
No.
8612
8613
8614
8615
8616
No.
8617
8618
8619
8620
8621
8622
8623
8624
8625
8626
8627
8628
8629
No.
8630
No.
8631
8632
No.
8633
8634
Title
11172-5 Reference Software
ISO/IEC 11172-5:199x/DCOR 1
Title
13818-4 Conformance testing
ISO/IEC 13818-4:2004/AMD 2:2005/Cor. 1
Title
13818-7 Adavnced Audio Coding
DoC on ISO/IEC 13818-7:2006/PDAM 1
ISO/IEC 13818-7:2006/FPDAM 1, Transport of MPEG Surround data
in AAC
Title
14496-3 Audio
Study on ISO/IEC 14496-3:2005/PDAM 5, BSAC Extensions
DoC on ISO/IEC 14496-3:2006/PDAM 6, Symbolic Music
Representation
WD on Support for 64-bit address space in ancillary data
Request for Amendment, AAC-ELD
ISO/IEC 14496-3:2005/PDAM 9, AAC-ELD
Title
14496-4 Conformance testing
ISO/IEC 14496-4:2004/AMD11/Cor. 2 Parametric Stereo Conformance
DoC on ISO/IEC 14496-4:2004/PDAM 14, BSAC Extension
Conformance
ISO/IEC 14496-4:2004/FPDAM 14, BSAC Extension Conformance
DoC on ISO/IEC 14496-4:2004/PDAM 18, MPEG-1 and -2 on MPEG4 Conformance
ISO/IEC 14496-4:2004/FPDAM 18, MPEG-1 and -2 on MPEG-4
Conformance
DoC on ISO/IEC 14496-4:2004/PDAM 19, ALS Conformance
ISO/IEC 14496-4:2004/FPDAM 19, ALS Conformance
DoC on ISO/IEC 14496-4:2004/PDAM 20, SLS Conformance
ISO/IEC 14496-4:2004/FPDAM 20, SLS Conformance
Status of BSAC Extension conformance
Status of ALS Conformance
Status of SLS Conformance
Status of MPEG-4 Audio Conformance
Title
14496-5 Reference Software
Workplan for updates on SLS reference software
Title
14496-23 Symbolic Music Representation
Request for Subdivision, Symbolic Music Representation
ISO/IEC 14496-23:200x/FCD, Symbolic Music Representation
Title
23003-1 MPEG Surround
Request for Amendment, MPEG Surround conformance testing
ISO/IEC 23003-1:2006/PDAM 1, MPEG Surround conformance testing
224
TBP Available
No
06/10/27
TBP Available
No
06/10/27
TBP Available
No
No
06/10/27
06/10/27
TBP Available
No
No
06/11/10
06/10/27
No
No
No
TBP
06/11/10
06/10/27
06/10/27
Available
No
No
06/10/27
06/10/27
No
No
06/10/27
06/10/27
No
06/10/27
No
No
No
No
No
No
No
No
TBP
06/10/27
06/10/27
06/10/27
06/10/27
06/10/27
06/10/27
06/10/27
06/10/27
Available
No
06/10/27
TBP Available
No
06/10/27
No
06/10/27
TBP Available
No
No
06/10/27
06/12/15
Request for Amendment MPEG Surround reference software
ISO/IEC 23003-1:2006/PDAM 2, MPEG Surround reference software
Workplan for MPEG Surround verification test
SAOC use cases, draft requirements and architecture
Draft Call for Proposals on Spatial Audio Object Coding
Title
Scalable audio and speech coding
8640 Workplan for Exploration of Speech and Audio Coding
No.
Title
MPEG Promotion
8641 Audio Bifs version 3
8642 Audio Conformance and Reference Software Assets
8635
8636
8637
8638
8639
No.
225
No
No
No
No
Yes
TBP
06/10/27
06/12/15
06/10/27
06/10/27
06/10/27
Available
No
06/10/27
TBP Available
Yes
Yes
06/10/27
06/10/27
Annex E Agenda for the 79th MPEG Audio Meeting
Agenda Item
1. Opening of the meeting
2. Administrative matters
2.1. Approval of agenda and allocation of contributions
2.2. Communications from the Chair
2.3. Joint meetings
2.4. Review of task groups and mandates
2.5. Approval of previous meeting report
2.6. Review of AhG reports
2.7. Received national body comments and liaison matters
3. Plenary issues
4. Task group activities
4.1. MPEG Maintenance, including MPEG-1, MPEG-2 and MPEG-4 issues
4.2. Spatial Audio Coding Extensions
4.3. Speech and Audio Exploration
4.4. Symbolic Music Representation
5. Discussion of unallocated contributions
6. Meeting deliverables
6.1. Recommendations for final plenary
6.2. Establishment of new Ad-hoc groups
6.3. Approval of output documents
6.4. Responses to NB comments
6.5. Responses to Liaison statements
6.6. Press statement
7. Future activities
8. Agenda for next meeting
9. A.O.B
10. Closing of the meeting
226
Annex J – 3DG report
Source:
1
Mahnjin Han (Samsung AIT)
Opening of the Meeting
e.
Approval of the agenda
f.
Goals for the week
The goals of this week are:
 Review on-going AFX explorations
 Issue FDAM of Morphing and Texture conformance and reference software
 Issue FPDAM of GFX conformance
 Issue PDAM of Geometry and Shadow conformance and reference software
 Review new contributions regarding AFX
 Discussion on Future of MPEG 3D Graphics
The output documents related to 3D Graphics Compression are:
Title
14496-4 MPEG-4 Conformance
DoC on ISO/IEC 14496-4:2004/ FPDAM12 (Morphing &
Textures)
Text of ISO/IEC 14496-4:2004/ FDAM12 (Morphing &
Textures)
DoC on ISO/IEC 14496-4:2004/ PDAM16 (MPEG-J GFX)
Text of ISO/IEC 14496-4:2004/ FPDAM16 (MPEG-J GFX)
Request for ISO/IEC 14496-4:2004/ AMD21 (Geometry &
Shadow)
Text of ISO/IEC 14496-4:2004/ PDAM21 (Geometry &
Shadow)
Editor
Jeong-Hwan Ahn
Jeong-Hwan Ahn
Vishy Swaminathan,
Mark Callow
Vishy Swaminathan,
Mark Callow
Jeong-Hwan Ahn
Jeong-Hwan Ahn
Title
14496-5 MPEG-4 Reference Software
DoC on ISO/IEC 14496-5:2001/ FPDAM9 (Morphing &
Textures)
Text of ISO/IEC 14496-5:2001/ FDAM9 (Morphing &
Textures)
Request for ISO/IEC 14496-5:2001/AMD13 (Geometry &
Shadow)
Text of ISO/IEC 14496-5:2001/ PDAM13 (Geometry &
Shadow)
Editor
Title
14496-16 MPEG-4 Animation Framework eXtension (AFX)
3D Graphics Core Experiments Description
Editor
227
Francisco Morán
Francisco Morán
Patrick Gioia
Patrick Gioia
Marius Preda
3D Graphics Compression FAQ 16.0
g.
Patrick Gioia
Standards from 3DG
In red, status reached at this meeting. In yellow, status reached at next meeting. Projects that
reached International Standard status have been removed.
Std
Pt
Edit.
Project Description
4
4
2004
4
4
2004
4
4
2004
4
5
2001
4
5
2001
4
5
2001
4 16
200x
Amd.12 Conformance on
Morphing and
Textures
Amd.16 MPEG-J GFX
conformance
Amd.21 Geometry and
Shadow conformance
Amd.9 Reference software
on Morphing and
Textures
Amd.11 MPEG-J GFX
reference software
Amd.13 Geometry and
Shadow reference
software
Amd.1 Geometry and
Shadow
h.
CfP
WD
CD
FCD
FDIS
PDAM FPDAM FDAM
DCOR
COR
05/04 05/10
06/04
06/10
06/04
06/10
07/04
06/07
06/10
07/04
07/10
05/04
05/10
06/04
06/10
06/01
06/04
06/07
07/01
06/07
06/10
07/04
07/10
05/04
06/04
06/07
07/01
Room allocation
3DG : Yuanqi (1st floor of Office Building)
i.
Allocation of contributions
N°
D1
Title
Schedule
D1
D1
09:00~12:30
D1
12:30~14:00
D1
14:00~16:00
Monday
MPEG Plenary
Lunch Break
3DG Plenary
M13745
Roll call, Agenda, Goals, FAQ,
etc.
Mahnjin Han
Report of AHG on 3DGC
documents, experiments and
software maintenance
Marius Preda
Jeong-Hwan
Ahn
Francisco
Morán Burgos
Vishy
228
Activity
MPEG General
3DG General
N°
Title
Schedule
Activity
Swaminathan
Conformance, ref s/w status,
voting results (M13764, M14765)
3DGC collaboration
16:30~17:30
D1
17:30~18:00
3DG Implementations
Implementation of JPEG 2000
M13864 elementary stream support in
MPEG-4 reference software
M13962
www.3DoD.org : an MPEG-4 3D
Graphics Database
Implementation
Marcos Avilés
Francisco Morán
Marius Preda
Marius Preda
Son Tran
Duc Tran
Ivica Arsov
Françoise
Preteux
D1
18:00~18:30
Systems
Demos
3DG Demos in Systems demo
session
D2
D2
D2
12:00~14:00
D2
14:00~14:30
Tuesday
Lunch Break
3DG Reference Software
Reference s/w
Comments on the inclusion of
M13911 3DMC-Extension in Part 11 Scene
description and application engine
3DG Exploration Experiment
Results of Exploration
Experiments (EE1: Static and
M13839
Animated 3D Object
Compression)
Results of evaluation experiment
EE1 on static and animated 3D
M13888
mesh coding : skinning-based
dynamic mesh compression
Results of evaluation experiment
EE1 on static and animated 3D
M14028 mesh coding : skinning-based
compression versus MPEG-4
AFX-IC
Proposal for Large 3D
Environments Profile
D2
16:30~17:00
New
D2
17:00~17:30
New
Khaled Mamou,
Titus Zaharia,
Marius Preda,
Françoise
Prêteux
Titus Zaharia,
Marius Preda,
Khaled Mamou,
Françoise
Prêteux
Patrick Gioia
3DG New – Technical
M13961 Proposal for geometry related
EE1
Shinjun Lee,
Jeong-Hwan
Ahn
3DG New – Profile
M13960
D2
14:30~16:00
Romain
229
N°
Title
space partitioning streams
Schedule
Cavagna,
Patrick Gioia
D2
17:30~18:00
3DG New – Informative
Error-resilient profile for
M13883 MeshGrid: robust encoding of the
reference-grid
D3
Activity
New
Dan Cernea
Adrian
Munteanu
Maryse Stoufs
Alin Alecu
Jan Cornelis
Peter Schelkens
D3
D3
09:00~11:00
D3
11:00~11:30
Wednesday
MPEG Plenary
3DG Plenary
MPEG General
3DG General
Work status review
D3
11:30~13:00
3DG Exploration Experiment
EE1
EE1 discussion
D3
13:00~14:00
D3
14:00~15:00
Lunch Break
3DG New
New
M13883 discussion
D3
15:00~16:00
3DG Discussion
3DG General
Reconfigurable Graphics Coding
discussion
D4
D4
D4
10:30~11:30
Thursday
3DG Experiment Discussion
EE1
EE1 discussion
D4
11:30~12:00
3DG + Requirements in 3DG
EE1
CfP for EE1
D4
12:00~14:00
D4
14:00~18:00
Lunch Break
3DG documentations
D5
Output documents review
Friday
D5
D5
09:00~12:00
3DG Plenary
3DG General
3DG General
AhGs and resolutions
D5
12:00~14:00
D5 14:00~
Lunch Break
MPEG Plenary
230
MPEG General
j.
Attendance list
Name
Mahnjin Han
Jeong-Hwan Ahn
Francisco Morán
Marius Preda
Françoise Prêteux
Titus Zaharia
Patrick Gioia
Country
Korea
Korea
Spain
France
France
France
France
Eun-Young Chang
Euee S. Jang
Sunyoung Lee
Sinwook Lee
Jae Bum Jun
Dan Cernea
Itaru Kaneko
Korea
Korea
Korea
Korea
Korea
Belgium
Japan
Demin Wang
Canada
Company
Samsung AIT
Samsung AIT
UPM
INT
INT
INT
France Telecom
R&D
ETRI
Hanyang Univ.
Hanyang Univ.
Hanyang Univ.
Hanyang Univ.
VUB
Tokyo
Polytechnic
University
CRC
231
2
k.
General issues
General Discussion
i.
Experiments
Last meeting resolution
In the core experiments and exploration experiments, each participant must have an input
contribution to the next meeting. Otherwise, they will be removed from the participant list of that
experiment at the next meeting.
Resolution
For each new specification development activity, 5 National Bodies should commit resources to that
activity. Contributions should be made at each meeting from those NBs until that activity is
finalized.
3DGC will no longer have Exploration Experiments.
3DGC will only have Core Experiments for any official experiments.
The condition for the CE is to have at least 2 active participants (companies or universities having
support from companies on that experiment) dedicating resources to do the work and making
contributions at each meeting.
If a participant does not make any contribution at a meeting, then that participant will not be
considered as active.
The activity in the CE does not necessarily imply adoption into the standard.
l.
Liaisons
i.
TC184 SC4
ISO/TC 184/SC 4 is a committee that produced ISO 10303, also known as STEP (Standard for the
Exchange of Product data). The chair of SC4 has asked MPEG Liaison chair on the possibility of
using MPEG 3D Graphics tools for visualizing their PLM/CAD/CAM data. MPEG has provided
informative letter to TC184 SC4, explaining the 3D Graphics technology in MPEG.
ii.
Khronos
WG11 and Khronos have many related area of work and it would be beneficial to both groups for
exchanging information and raise awareness of the different specifications developed. Through such
communication, it may even grow to establishing new standardization activity that can benefit from
the expertise from both groups. For example, there is a need in the market for the compression of
COLLADA, the authoring format standardized by Khronos, and MPEG has the expertise in
compression. However, in order to have such collaboration, few management issues still remain to
be resolved. As a starter, Khronos has provided liaison questionnaire and MPEG has produced
response to it, together with the request of the Liaison establishment.
3
m.
AFX (14496-16) activities
Exploration Experiments
i.
E1. Static and Animated 3D Object Compression
Last meeting resolution
Reopen EE1 and add MCGV/TDCT as additional tools to be compared.
232
The test shall be performed according to the test condition described in EE1
The 23 test sequences are added to the EE1 as additional test data set.
Industry support is needed.
1.
M13839 – Results of Exploration Experiments (EE1: Static and
Animated 3D Object Compression)
6 models have been used for the experiment of CoordinateInterpolator Compression.
The test set contains articulated motion, non-articulated motion and deformation.
The presentation contains additional results that compare DA measurement with Hausdorff distance
program.
2.
M13888 – Results of evaluation experiment EE1 on static and
animated 3D mesh coding : skinning-based dynamic mesh compression
This contribution is presented together with M14028.
Please see the section below for more detail
3.
M14028 – Results of evaluation experiment EE1 on static and
animated 3D mesh coding : skinning-based compression versus MPEG-4 AFX-IC
This contribution shows the results in M13888 and additionally shows the AFX IC results.
Skinning method shows out performance on most of the data set used in this contribution.
However, due to some misunderstanding, only the new test data set added at the last meeting were
used and not the ones used originally for the EE1 experiments.
Discussion on EE1
There were long hours of discussion on whether or not the exploration experiment conditions are
met so as to be promoted to the core experiment stage. The proponent of the skinning-based
compression technology claimed that enough evidence has been provided to show that current
standard can be outperformed by other technology available today. However, some other parties
claimed that the experiment has to be completely performed in order to know the exact situation.
Also, it was suggested that in order to gather more interest from the industry on this experiment, we
need more exposure. Therefore, making of a call for proposal has been discussed with the
Requirements group. However, it has been realized that what should be requested by the group for
this experiment is very detailed and the size of the problem too small to make the call for proposal.
Instead, we had other options, such as call for evidence or core experiment.
The group came to the conclusion that having exploration experiment stage is very confusing and so
it should be merged to core experiment activity. General discussion section of this report (clause
2.1.1) contains the resolution on the conditions of future core experiments.
Resolution for EE1
This work shall continue as a core experiment with additional mandate to extend the experiment by
adopting animated 3D object compression framework idea where harmonization of various
compressed data has to be proven.
DA shall be used for future experiments unless there is a contribution showing Hausdorff program
having different tendencies compared to DA in low bitrate. However, proponents are welcomed to
give additional test results with Hausdorff program.
n.
Reference Software & Conformance
Last meeting resolution:
Produce a working draft of the reference software and conformance of 14496-16:200x/Amd.1.
233
i.
Morphing & Textures
The voting comment from France on the FPDAM of ref s/w and conformance is as follows:
“The FNB disaproves this project but will change its position into an approval once the remaining
errors in the bitstreams and/or software are resolved.”
Resolution
Errors are not found in the current reference software nor in the conformance.
Both shall be promoted to FDAM.
ii.
Geometry & Shadow
The Shadow software is being integrated to IM1 and the conformance document is ready
Multiresolution footprint-based representation conformance part is ready, together with the
bitstreams.
3DMC Extension conformance and reference software document is ready.
Conformance and reference software for support of JPEG2000 should also be added, and two weeks
of editing period is needed for this.
Resolution
The conformance and reference software documents shall be promoted to PDAM.
iii.
M13911 - Comments on the inclusion of 3DMC-Extension in Part 11
Scene description and application engine
While integration of 3DMC extension with the reference software (MPEG-4 IM1), it was realized
that there is no way to support the 3DMC extension tool because IndexedFaceSet node is currently
being compressed by 3DMC.
The proposal is to define additional value for the type field in BitWrapper node.
Currently, the 3DMC uses type 0. The proposal is to use the value 1 for 3DMC extension.
The proposal also shows how it can be implemented in IM1 by introducing
JWAFX3DMCExtensionDecoder class.
AFXExtDescriptor has the AFX object code which makes the type value redundant when an
elementary stream is attached.
However, in case the buffer is being used, the type value is the only way to distinguish the type of
data the buffer contains.
Resolution
Having a new type value for the BitWrapper of IndexedFaceSet is approved.
In addition, the AFX object code should be defined.
In the case where the type value and AFX object code conflicts, then AFX object code prevails.
o.
Promotions
Last meeting resolution
The 3DGC subgroup would like to thank INT and IMEC for providing the software and will look
forward to seeing even more improved demonstration in the future.
i.
M13962 - www.3DoD.org : an MPEG-4 3D Graphics Database
This contribution introduces an on-line 3D database, developed by INT, that uses MPEG-4
technology for compression of objects and animations.
234
It contains thousands of models and provides not only the uploading and browsing functionalities,
but also provides on-line visualization using integrated MPEG-4 3D Graphics Player.
This site is open to the public and anyone can upload their own content after registering.
Resolution
The 3DGC subgroup thanks INT for providing MPEG-4 3D database with online visualization on
the website and encourages everyone to try them and give feedback.
p.
Additional AFX related issues
Last meeting resolution
Continue the work described in three bullets below.
- using predictive mode for the predictive based approach
- improve the sw implementation (better use of masking)
- visualize the compressed files in the MPEG-4 player
Contact the SC24 KNB to ask for their purpose of the request to SC24 and inform them about BBA
which may satisfy their needs.
i.
M13960 - Proposal for Large 3D Environments Profile
This contribution proposes the making of a profile for 3D navigation applications based on
geographic data.
Demonstration during the meeting has shown MPEG-4 compliant set of services that makes use of
geographic and urban environment. It showed example service which is similar to Google Earth, but
with enhanced 3D data and multimedia services.
The contributor informed the group that it is ready to be commercially exploited by several
companies and organizations in France.
Resolution
There are participants within the group that are interested in the proposed profile (UPM, INT).
However, the group encourages other companies to take interest and give support to form critical
mass for this profile.
This work will continue until the next meeting by the interested parties so that the description of the
profile is completed.
ii.
M13961 - Proposal for geometry related space partitioning streams
This contribution is a preliminary proposal to gather interest for sending space partitioning
information with the geometric data.
This would be important for managing large scene in client/server or P2P mode.
It is reported that reconstruction time for Rennes without space partitioning takes 20.256 seconds,
while with space partitioning, it only takes 5.72.
However, this topic also related to Systems because it includes management of large scene.
Resolution
Open a core experiment to perform test on what we can gain from various tools.
Also, the proposal should be refined until the next meeting to have a more detailed proposition
including the aspect related to Systems.
More industry interest is needed for this activity.
235
iii.
M13883 - Error-resilient profile for MeshGrid: robust encoding of the
reference-grid
This is a technical, but informative, contribution on adding error resilient functionality to the
MeshGrid stream.
It shows how to make error resilient streams in case of error prone environment or lack of
bandwidth.
The proposal is to add protection data at the end of the bitstream.
The important part is to decide which packet is more important than others.
However, this contribution only deals with reference grid.
Resolution
The 3DG subgroup thanks VUB for the informative technical contribution related to MeshGrid.
The work should continue to support error resiliency for the connectivity wireframe.
iv.
Configurable Media Coding
This part of the report is about a general discussion within the 3DG group on configurable media
coding.
The chair of RVC AhG was invited to explain about the current status and the general idea behind
RVC
The RVC activity has been initiated by video, where many coding tools are competing each other.
In RVC, a toolbox is being built that includes coding tools.
The stream will provide information about decoder in addition to the actual bitstream data.
Having a collection of tools in the toolbox, one can easily maintain the tools and add new tools.
Discussion
3D graphics has a special characteristic that allows fitting in well within the configurable media
coding framework.
The procedure for 3D graphics is still an open question.
New methodology that follows the concept of RVC is also welcomed.
It opens more possibilities to combine tools. It can also be used as standard maintenance purpose.
Resolution
The group agrees that the reconfigurability can be a good functionality for 3D graphics.
This issue will be raised in the 3DGC reflector for more discussion.
The result of the discussion will be reported at the next meeting.
4
q.
GFX (14496-21) activities
Reference Software & Conformance
Last meeting resolution
The reference software is promoted to FPDAM.
The conformance stays at the PDAM level. However, the study document for DoC and the
Conformance document have been produced.
Discussion
Conformance has been delayed one meeting and is scheduled for FPDAM at this meeting.
The document is ready, but the bitstreams are not completed yet.
236
Resolution
The conformance for GFX will be promoted to FPDAM with 4 weeks editing period in order to
complete the bitstreams.
5
Resolutions of 3DG
r.
Output documents
i.
The 3DG subgroup recommends to approve the following documents
No.
8489
8490
8491
8492
8493
8494
No.
8495
8496
8497
8498
No.
8499
8506
s.
TBP Available
Title
14496-5 Reference Software
DoC on ISO/IEC 14496-5:2001/ FPDAM9 (Morphing & Textures)
Text of ISO/IEC 14496-5:2001/ FDAM9 (Morphing & Textures)
Request for ISO/IEC 14496-5:2001/AMD13 (Geometry & Shadow)
Text of ISO/IEC 14496-5:2001/ PDAM13 (Geometry & Shadow)
TBP Available
Title
14496-16 Animation Framework eXtension (AFX)
3D Graphics Core Experiments Description
3D Graphics Compression FAQ 16.0
TBP Available
No
No
No
No
No
No
No
No
No
No
No
Yes
06/10/27
06/10/27
06/10/27
06/11/24
06/10/27
06/11/10
06/10/27
06/10/27
06/10/27
06/11/10
06/10/27
06/11/17
Resolutions



t.
Title
14496-4 Conformance testing
DoC on ISO/IEC 14496-4:2004/ FPDAM12 (Morphing & Textures)
Text of ISO/IEC 14496-4:2004/ FDAM12 (Morphing & Textures)
DoC on ISO/IEC 14496-4:2004/ PDAM16 (MPEG-J GFX)
Text of ISO/IEC 14496-4:2004/ FPDAM16 (MPEG-J GFX)
Request for ISO/IEC 14496-4:2004/ AMD21 (Geometry & Shadow)
Text of ISO/IEC 14496-4:2004/ PDAM21 (Geometry & Shadow)
The 3DG subgroup recommends appointing Jeong-Hwan Ahn (Samsung AIT) as the editor
of ISO/IEC 14496-4:2004/ AMD21 and thanks him for taking the responsibility of that
project.
The 3DG subgroup recommends appointing Patrick Gioia (France Telecom R&D) as the
editor of ISO/IEC 14496-5:2001/ AMD13 and thanks him for taking the responsibility of
that project.
The 3DG subgroup thanks INT for providing MPEG-4 3D database with online
visualization on the website at http://www.3DoD.org, and encourages everyone to try them
and give feedback to Dr. Marius Preda.
Establishment of 3DG Ad-Hoc Groups
N8507
Mandate:
AHG on 3DG documents, experiments and software maintenance
1. Maintain and edit 3DG documents
2. Coordinate 3DG CE activity
3. Coordinate 3DG related conformance and reference software
237
Chairman: Marius Preda (INT)
Co-chairs Jeong-Hwan Ahn (Samsung AIT)
Francisco Morán Burgos (UPM)
Vishy Swaminathan (SUN Microsystems)
Duration:
Meetings
Reflector:
Subscribe:
6
Until 79th Meeting
Sunday before 79th meeting
mpeg-3dgc AT gti. ssr. upm. es
http://www.gti.ssr.upm.es/mailman/listinfo/mpeg-3dgc
Closing of the Meeting
See you in Marrakech
238
Annex K – Test report
Source: Tobias Oelbaum
Opening of the Meeting
Goals for the week
The goals of this week are:
 Refine and extend the draft verification test plan for SVC
 Discuss the possibility of testing or taking part in a test that evaluates the quality of the
DIRAC video codec
 Support of the JVT group if visual results have to be produced for proposed algorithms
 Issue a request for new test sequences
Test Activities
Scalable Video Coding Verification Test
A verification test for the Scalable Video Coding activity will be needed to finish the work on this
project. For this reason the Draft Verification Test Plan was refined.
Input from the discussion about profiles in SVC lead to the design of three possible application
scenarios: TV broadcast, wireless HD camera and mobile video communication. Test conditions
were drafted along these application scenarios, however so far no detailed test conditions regarding
bit rates, sequences or extraction paths have been made. Proposals concerning the methods for the
visual evaluation have been made.
An AHG for designing and conducting the SVC verification test has been set up. This AHG is
chaired by Tobias Oelbaum and Mathias Wien. There are associate chairs from the companies
requesting a profile for SVC.
Scalable Video Coding
A viewing session was conducted to support the JVT subgroup in the discussion about the
advantages of different down sampling filters. Results of this viewing session were used in the
discussion about contribution JVT-U147.
Dual Track Approach
In a joint meeting with requirements the possibility of participating in a test that evaluates the
performance of DIRAC (DIRAC is a wavelet video codec developed by the BBC which is claimed
to be royalty free and to deliver a good quality) was discussed. It was decided to issue a resolution
showing the interest of WG11 of participation in such a test.
Video Test Sequences
In document m13874 new test sequences that are available inside MPEG are described.
A request for new video test sequences was issued.
239
Test Resolutions
Resolutions
WG11 announces its availability to participate in designing tests to assess the performance of
existing video codecs that are available at “royalty free” conditions. Subject to availability of
internal resources of WG11 would also like to be involved in the actual performance evaluation of
such codecs.
Output Documents


8553 Draft SVC Verification Test Plan
8554 Request for Video Test Sequences
240
Annex L – ISG report
Source:
1.
Marco Mattavelli (EPFL)
Overview
The main work items of the Implementation Studies Subgroup in Hangzhou are:
1. MPEG-4 Part 9 Reference HW description:
 The editing of the Study of the Third Edition of the TR concerning the extension of
features and documentation for the “virtual socket” integrated framework, making
possible to put together in a single application MPEG-4 Part 9 with MPEG-7 Part 7 and
AVC (MPEG-4 Part-9) software and the addition of new HDL modules.
 The review of the new HDL module and associated documentation submitted for
integration in Part 9.
2. The contribution to the Reconfigurable Video Coding (RVC) activity reporting the results
for the on going core experiments.
3. The evaluation of the final 5 proposals for the finite precision DCT/IDCT specification
considering the results of performance and complexity for all the metrics agreed in
Klagenfurt meeting and now obtained by the experiments of the common testbed.
Input contributions to ISG group w.r.t. the above items are summarized according to the following
table:
Contributions to ISG
M13751
Robert Turney (Xilinx) Marco
Mattavelli (EPFL)
AHG report on MPEG-4 Part 9
Reference Hardware Description
Phase 1 and 2”
M13795
University of Calgary, Calgary,
Alberta, Canada.
Author:
Choudhury A.
Rahman and Wael Badawy
Telecommunications InstituteUniversity of Aveiro-Portugal
Author:
M. Santos and A.
Navarro
A HW Block For H.264/AVC Context
Adaptive Variable Length Coding
(CAVLC).
M13989
2.
Hardware implementation of full
search H.264 motion estimation
Detailed Report
2.1.
The progress in the development of the MPEG-4 “Part 9 Reference
Hardware Description”
241
The ISG activity at the Hangzhou meeting has mainly been devoted to
 the review of the two contributions,
 the editorial work for third edition of the technical report,
Contribution M13795 presented a new HDL block implementing for the AVC part the Context
Adaptive Variable Length Coding (CALVC). The new contribution is completed in its
documentation and has been included in the study document of the Third edition of the PDTR of
Part 9.
Contribution M13989 presents an HW block for the full search estimation of motion vectors for the
macroblock partitioning of AVC. The contribution has been included in the study document of the
Third edition of the PDTR of Part 9.
The ad-hoc group on the development of MPEG-4 Part 9 (N8520) has been re-established with the
ususal mandate, including a specific mandate for the specification and development of the
demonstration platform and the mandate for continuing the investigating the hardware reference
description for DCT/IDCT.
The ad-hoc schedule includes 3 telephone conferences before next meeting. Phone conferences are
planned on 23th November, 21st December, 11th January at 4 p.m. CET (3 p.m. GMT). Tel: (from
US 1-877-582-3182, from outside US 1-770-970-4161, participant code 9202060193).
2.2.
The contribution to the activity on Reconfigurable Video Coding (RVC).
An important part of the ISG activity in Hangzhou has been spent in joint meetings with Video,
MDS, Systems and the RVC subgroup. The main issue was the evaluation of the results of the ongoing core experiments and of the choices between the two proposal in the WD. The major
outcome of the discussion was the agreement of the need of RVC to provide an “abstract model” of
the “decoder description” that is a machine executable non-ambiguous model for conformance point.
At the moment the only proposal satisfying this condition is the proposal for a CAL based
description of FUs and the associated decoder description language.
2.3.
Contributions to the review of the proposals received as answer to the call
for specification for finite precision IDCT
Essentisl review of finite precision DCT/IDCT contributions
M113912
Author:
The Drift Problem of Fixed-Point IDCT on News Sequence
Zhibo Ni, Lu Yu
5 CD IDCT has been tested with encoder at double precision and QP 1,2,3,4, evident drift artefacts
at 50th image are already visible not only for QP = 1. 13979, 13799 similar artefacts are found.
Three proposals showed high PPE drift errors.
Conclusion of the contribution is that IDCT that do not bound drift should not go into the CD. A
check verified that even when the Morrison test was implemented the drift was still present.
M13927
Author:
Anti-IDCT for IDCT Drift Test
Cixun Zhang , Lu Yu, Yuriy A. Reznik
242
Contribution about the possible differences of implementations in the 1180 accuracy range. Anti
IDCT is the symmetrical IDCT versus the floating point. So encoder has been implemented with
anti IDCT and decoder. Three proposals show PSNR drift up to 10 dBs when matched at the
encoder with their anti-IDCT implementation.
M14004
Authors:
On clipping and dynamic range of variables in IDCT designs
Yuriy A. Reznik
The contribution reports information on required dynamic range and clipping for MPEG systems.
Previous publication (1997) shows that the minimum dynamic range to prevent clipping is at least
+/-1805. Some additional ways (transmit dynamic range interval to decoder) to prevent clipping are
recommended for MPEG conformance.
M13916
Test Results for Technical Selection of Committee Draft of ISO/IEC
23002-2 Fixed-Point IDCT
Author:
Zhibo Ni, Cixun Zhang , Lu Yu
Comparison of 5 proposals in terms of test, adders and shift needed. Comparisons of PSNR loss.
M13990
Author:
Performance in MPEG-4 of five submitted integer IDCTs for CD
Antonio Navarro and Antonio Silva
Results of the 5 proposals for drift and PSNR. The best proposal with the lowest drift is m13797.
However, if we also take into account the computational complexity, the contribution suggests that
best proposal is M13803 if a uniform weighting is applied between complexity and PSNR results.
M13941
Author:
Test Results for Selection of Committee Draft of ISO/IEC 23002-2 Fixed-Point
IDCT
Honggang Qi, Wen Gao, Debin Zhao and Siwei Ma.
The contribution reports the comparison between the 5 candidates. Consistent with previous results
even if PSNR results are provided for separated sequences and cannot be compared directly. An
additional table for pruning has been added.
Results reported on the same form has been confirmed the cross check of other proponents and the
document has been uploaded on the web site.
M13993
Source:
Lazar Bivolarski, Yuriy A. Reznik
Connex Technology, QUALCOMM Incorporated
The contribution reviews some modern platforms widely deployed and recognized to verify
implementation of IDCT. Execution flows and instruction sets are analyzed in particular multiply
and multiply accumulate using parallel multipliers. The conclusion is that IDCT candidate should
include implementation of multiply and shift that are compatible with extended instruction set that
implement parallel 16 bit multiply and adds and should avoid scaling that are not compatible with
extended instruction sets.
M13847
Author:
FastVDO 16-bit IDCT Proposal for CD: Performance and Comparison
Trac D. Tran, Lijie Liu, Pankaj Topiwala
243
The contribution suggests further test for improved core experiments.
M13997
Authors:
On the Cost and Performance of IDCT Implementations in Hardware
Joanna J. Eastment, Arianne T. Hinds
This contribution reports considerations on complexity and performance for 3 DCT designs on two
platforms one microprocessor and custom IC design. For 13784 algorithm without and with
multiply ARM and Intel are compared in terms of cycles of latencies. No parallelisation is
considered.
For custom IC rules for size and delay are used to estimate multiply with carry look-haead adders
and adders implementation of the various algorithms. Conclusion the IBM Qualcomm proposal is
claimed to be simpler for full custom HW implementations that other candidates.
M13914
Authors:
Analysis of Hardware Implementation Cost of Fixed-Point IDCT
Dandan Ding, Zhibo Ni,Lu Yu.
Report on HW implementation of 4 of the candidate implementations. Sythesis of the butterfly
stage is used as estimation of the integration area plus the scaling (797 require 7 time area versus
784). FPGA results for Virtex4 are compared for multiplierless implementations. Conclusion 797 is
the higher while 799 and 784 have the lowest complexity. (13803 proposal has not been considered).
However usage of multiplication available on Virtex4 has not been considered.
M14005
Authors:
Additional information on IDCT CD candidates and proposed Core
Experiments
Yuriy A. Reznik (Qualcomm), Arianne T. Hinds (IBM), Cixun Zhang, Lu Yu,
Zhibo Ni (ZJU), Lazar Bivolarski (Connex Tecnology), Honggang Qi, Siwei Ma
(CAS)
The contribution reports a proposal for further core experiments for optimizing current candidate
versus minimization of multiplications or minimization of adders. Performances can be between 2
times better than 1180 up to 50 times increasing the number of additions from the minimum to 42 to
54. candidates can be implemented using multipliers or switch to multiplierless. Results of drift are
reported for QP=1 for MPEG-2.
14005 is an extension of 13784 that well behave with respect to drift 44 additions and 18 shifts with
factors that are only 8 bits. PPE error is also reduced versus more complex algorithms including the
current candidates.
Proposal is to carry on core experiments on that variation to further check performances.
M14003
Authors:
Cross check of proposed additional (CE-stage) IDCT designs
Antonio Navarro
The contribution is a cross check of candidate core experiments presented in 14005. The results
confirm that all proposals except the linearity test where 4 of the 6 did not pass the test.
M13996
Authors:
On the Usage of of High Precision IDCTs in Existing MPEG Products
Honggang Qi, Arianne T. Hinds
244
Survey of SW codec products available on the market. Floating point DCT is always one option that
can be used to encode video content, thus avaiding drift from floating point implementation is
fundamental for finite precision specifications.
M14006
Authors:
Examples of existing IDCT designs
Yuriy A. Reznik
Document reporting implementation of 16 bits IDCT, publicly available, some sources provides
several IDCT implementation examples optimized in assembler for most common platforms.
M13846
Author:
FastVDO 16-bit IDCT Proposal for CD: Performance and Comparison
Trac D. Tran, Lijie Liu, Pankaj Topiwala
Proposes a core experiment for tuning accuracy and complexity for the family of lifting scheme
implementations.
M13930
Author:
CNNB comments on the work of fixed-point 8x8 IDCT transform
CNNB
Request of disqualifying candidate that did not respect procedural rules. Request has not been
accepted.
M14000
Authors:
On the Complexity Analysis of IDCT Algorithms for CD Selection
Lazar Bivolarski (Connex Technology),
The contribution analyzed two proposals 13799 with 13791 in terms of complexity and the
conclusion is that 13791 is less complex. The difference in complexity is 10 %.
Evaluation of results
A table of all results has been assembled by the contributions reporting results and cross check of
results. Two candidate algorithms are clearly outperforming the others. He two candidates presents
a small difference in complexity.
Candidate 13784 presents a 33% less complexity and slightly lower results at the very high ranges
of quality around 50 dBs, fdor these reasons it has been selected with the consensus of the group as
better performing specification and has been included in the CD.
New core experiments have been defined to evaluate performances for all selected metrics for
additional three bits of widths so as to consider a higher dynamic range as remarked and suggested
by one contribution and to investigate possible further reductions of the complexity.
3.
Resolutions
The above activities have led to the following resolutions and output document approval.
245
MPEG-4
5.6 Part 9 – Reference Hardware Description
5.6.1 The implementation studies subgroup recommends approval of the following documents
No.
Title
14496-9 Reference Hardware Description
N8518 Status of HDL submissions and commitments for MPEG-4 Part-9
Study of “ISO/IEC PDTR 14496-9 3rd Edition Reference
N8519
Hardware Description”
TBP Available
No
No
Yes
Yes
MPEG-C
10.2 Part 2 – Fixed point implementation of DCT/IDCT
10.2.1 The ISG and the video subgroups recommend approval of the following documents
No.
8479
8480
8481
Title
23002-2 Fixed point implementation of DCT/IDCT
ISO/IEC CD 23002-2 Fixed point IDCT and DCT
Description of Core Experiments on Fixed-Point DCT/IDCT
Software Testbed for fixed-point DCT/IDCT V 5.0
246
TBP Available
No
No
No
06/10/27
06/10/27
06/12/01
Annex M– Liaisons report
Source: Kate Grant, Nine Tiles
The following documents were reviewed in the Liaison meeting:
Liaison Statements
13768
Liaison Statement from JTC1 SC37
From London SC37 meeting: report on Special Group on Face Identity Data
and information of new scope of 19794-12, draft LS considered in Klagenfurt
13780
Liaison Statement from ITU-R SG6 WP 6J
Editorial comments on video amendments concerning support of colour spaces
13792
Liaison Statement from IEC TC100
Ballot text of CDV of IEC 62455 IP and transport stream based service access
13798
Liaison Statement from W3C
Response to outgoing Klagenfurt statement on correct use of namespaces and
new events in LASeR
13808
Liaison Statement from ITU-R SG6 WP 6Q
Final call for proposals on extension of Recommendation ITU-R BS.1387-1 to
address the measurement of multi-channel audio signals
13809
Liaison statement from 3GPP2
Requesting establishment of Category A liaison and attaching documents re
3GPP2 project, working procedures etc
Liaison Statement from SMPTE on ISO Base Media File
13840
Format
Request for comments on attached CD “VC-1 Bitstream Storage in the ISO
Base Media File Format”
Liaison Statement from SMPTE Constraints on High
13841
Profile
Request WG11 to consider definition of 2 new Intra profiles: High 10 Intra
profile and High 4.2.2 Intra profile
Liaison Statement from SMPTE on new profile for
13842
production
Highlighting opportunity for a new profile intended for high quality production
applications designed for minimal computational complexity
13885
Liaison Statement from FLO Forum
Requests advice and guidance on how to progress Work Items relating to Rich
Media that could leverage aspects of the LASeR
Liaison statement from W3C Multimedia Semantics
14015
Incubator Group (W3C MMSem-XG)
Project to study existing multimedia metadata standards (inc MPEG-7 and
MPEG-21) and identify how semantic web technologies can deal with
interoperability problems through a number of use cases
14016
Liaison Statement from ITU-T SG 9
Developing a draft Question on the Free-viewpoint TV (FTV) system future
standardization especially from a view point of transport system aspect
14029
Liaison Officer report on IEC TC100 meeting
Information on formation of 2 new TAs (TA9: Audio, video and multimedia
247
applications for end-user network and TA10 Multimedia e-publishing and ebook) and other projects with work relevant to MPEG
Additional documents received after the Liaison meeting were also considered during the week.
14031
Liaison Statement from DVB
Asking if codepoints could be assigned for their BiM profiles
14032
Liaison Statement from OMA BAC MAE to W3C (cc to WG11)
RME has identified the need for clipping functionality that provides pixel
aligned clipping defined as a transformable rectangle implementable on a
device with limited resources. (LASeR rectclip fulfils the RME requirement)
The following documents were issued (see resolution 16.2.1 in N8432)
Liaison Statements
8526 Liaison Statement to UHAPI concerning M3W
8527 Liaison Statement to ITU-T FG/IPTV WG 6 concerning M3W
8528 Liaison Statement to 3GPP2
Liaison Statement to ITU-R SG6 WP 6J concerning colour space
8529
amendments
8530 Liaison Statement to ITU-R SG6 WP 6Q on Call for Proposals
8531 Liaison Statement to SMPTE
Liaison Statement to SMPTE on 4:2:2 and 4:2:0 Intra-only profiles
8532
of AVC
8533 Liaison Statement to SMPTE on 4:4:4 Intra-only profile of AVC
8534 Liaison Statement to FLO Forum
8535 Liaison Statement to IEC TC100
8536 Liaison Statement to W3C MMSem-XG
8537 Liaison Statement to ITU-T SG9 concerning FTV ad MVC
8538 Liaison Statement to OMA BAC MAE
8539 Liaison Statement to DVB
8540 Liaison Statement to ISO TC184 SC4
8542 Liaison Statement to SCTE
8543 Liaison Statement to WG1 (JPEG)
8544 Liaison Statement to Khronos
Liaison Statement to ITU-T FG/IPTV WG 6 concerning work on
8545
IPTV
8546 Liaison Statement to ITU-T SG16 Q23
8702 Liaison Statement to 3GPP
Requests for establishment of the following liaisons were prepared and approved:
(see resolution 16.2.2 in N8432)
Request for Establishment of Liaison
8548 Request for establishment of Category A liaison with 3GPP2
8549 Request for establishment of Category B liaison with AES
8550 Request for establishment of Category C liaison with Khronos
The Liaison Group compiled the following response to National Bodies: (see resolution 16.2.4 in
N8432)
248
8551
Response to National Bodies
The Liaison Group revised the List of Organisations with which MPEG has a liaison
relationship: (see resolution 16.2.3 in N8432)
8552
List of Organisations with which MPEG entertains liaisons (as of
October 2006)
249