Challenges in VoLTE
Terminal Testing and
Interworking
HEARING IS A FASCINATING SENSATION
ITU Workshop on “Voice and Video over LTE”
Geneva, Switzerland, 1 December 2015
Dr.-Ing. H. W. Gierlich
Managing Director Telecom Division
© HEAD acoustics GmbH
Outline
A variety of challenges exist in VoLTE testing and
interworking - The following selected topics are addressed
in this presentation:
LTE Setup for Voice Communication - Obstacles
Delay and Delay Testing
Network Impairments, Delay & Jitter Profile Testing
Challenges in Superwideband & Fullband
Positional Dependant Testing of Mobile Terminals
Summary
H. W. Gierlich
Advanced speech quality testing
VoLTE Terminal Testing and Interworking
©HEAD acoustics GmbH
© HEAD acoustics GmbH
2
1
Outline
Introduction
LTE Setup for Voice Communication - Obstacles
Delay and Delay Testing
Network Impairments, Delay & Jitter Profile Testing
Challenges in Superwideband & Fullband
Positional Dependant Testing of Mobile Terminals
Summary
H. W. Gierlich
Advanced speech quality testing
© HEAD acoustics GmbH
3
Call Handling 4G in a Test Setup
Test Equipment
With Reference
Access
Signal
Generation
&
Analysis
VoLTE
Network Simulator
DUT
VoLTE call
Codec
Codec
IMS
IMS registration
IMS registration
• Both DUT and reference gateway for testing are registerd in the test
IMS server
• VoLTE call takes place between DUT and reference gateway
• Reference gateway takes care of encoding/decoding
H. W. Gierlich
Advanced speech quality testing
VoLTE Terminal Testing and Interworking
©HEAD acoustics GmbH
© HEAD acoustics GmbH
5
2
Difficulties in VoLTE Call Setup
Network provider flash their own firmware (Branding) with their special
VoLTE support => special procedures for IMS Signaling/Registration are
needed.
Network providers use different LTE Cell/Signaling settings/parameters
(e.g. LTE Band, MNC/MCC, Security Settings, Shared-Key, etc…). These
parameters often are confidential and unknown if not provided.
SIM-Lock, the device under test needs a special - provider specific -UICC
(SIM card) for testing.
DUTs normally are locked to their home network. In order to connect to the
radio tester the correct MNC/MCC settings must be known, otherwise the
roaming procedure does not allow VoLTE connections.
The SIP/IMS registration is very sensitive to errors or unknown SIP
messages depending on the implementation of the IMS in the VoLTE
network simulators.
VoLTE settings in DUTs may be fixed (e.g. P-CSCF Address set to a fixed
address) -> the address used in the radio tester cannot be configured.
H. W. Gierlich
Advanced speech quality testing
© HEAD acoustics GmbH
6
Difficulties VoLTE Call Setup
Buggy VoLTE implementation in devices under test (e. g. SIP call
negotiation with octet-aligned mode but DUT is sending the speech
packets in bandwidth efficient mode).
If no IMS APN (access point name) or matching PDP (packet data
protocol) contexts is available in the DUT => no IMS registration => no
VoLTE call.
IPsec parameters are needed if device under test only work with a secure
connection (IMS-AKA).
No access to VoLTE settings in DUT for changing required parameters for
audio tests.
How to handle eSIM?
H. W. Gierlich
Advanced speech quality testing
VoLTE Terminal Testing and Interworking
©HEAD acoustics GmbH
© HEAD acoustics GmbH
7
3
Difficulties in VoLTE Call Setup
=>
Most of these issues not only lead to difficulties in testing –
the differences in IMS settings prevent interworking between
network operators in general!
Solution for testing:
Define a generic test mode and IMS setting implemented in all
terminals.
Potential solution for interworking:
Define a minimum set of IMS settings used in interworking.
H. W. Gierlich
Advanced speech quality testing
© HEAD acoustics GmbH
8
Outline
Introduction
LTE Setup for Voice Communication - Obstacles
Delay and Delay Testing
Network Impairments, Delay & Jitter Profile Testing
Challenges in Superwideband & Fullband
Positional Dependant Testing of Mobile Terminals
Summary
H. W. Gierlich
Advanced speech quality testing
VoLTE Terminal Testing and Interworking
©HEAD acoustics GmbH
© HEAD acoustics GmbH
9
4
3GPP & GSMA: Delay Testing
Requirement 4G:
Roud trip delay MTSI < 190 ms (min. Reqirement)
Performance goal < 150 ms
Delay measurement direction by cross correlation
Cross correlation avg. FFT Size:65536 Overlap:0,0% Envelope Normalized
%
100
80
60
33 ms
40
20
0
0
H. W. Gierlich
25m
50m
75m
Advanced speech quality testing
t/s
0.125
0.15
0.175
0.2
© HEAD acoustics GmbH
10
Delay in VoLTE - Not Stable, Delay Drift due to Clock Drift
t/s
0.125
0.124
0.123
RCV(2)*Source
0.122
0.121
0.12
0.119
0.118
Receiving
Direction
0.117
0.116
0.115
0
20
26
40
27
28
t/s
%
Drift: 815 µs/min
H. W. Gierlich
Advanced speech quality testing
VoLTE Terminal Testing and Interworking
©HEAD acoustics GmbH
80
29
100
30
120
31
13,6 ppm
© HEAD acoustics GmbH
11
5
Delay Results VoLTE – Clock Drift Compensation
t/s
0.125
0.124
Clock Drift
0.123
RCV(2)*Source
0.122
0.121
0.12
Receiving
Direction
0.119
0.118
0.117
Not compensated:
815 µs/min
48000 Hz
- Compensated:
0 µs/min
48000,67 Hz
-
t/s
0.116
0
20
26
40
27
28
t/s
%
80
29
0.115
120 0.144
100
30
31
RCV(2)*Source
0.142
Synchronized testing
possible
0.14
0.138
0.136
0
29
H. W. Gierlich
20
30
40
31
32
t/s
%
34
80
35
100
36
37
Advanced speech quality testing
120
38
© HEAD acoustics GmbH
12
VoLTE Terminal Delay: Call by Call Variation
Call by call delay variation and theoretical background:
• Clock syncronization absolutely required to avoid delay drift during
testing
• Clock sychronization does NOT allow packet stream synchronization
Phase shift
between packet
streams from call
to call
H. W. Gierlich
Advanced speech quality testing
VoLTE Terminal Testing and Interworking
©HEAD acoustics GmbH
© HEAD acoustics GmbH
13
6
Delay Testing: Results in Sending
Call by call delay variation: measurement examples in sending
• Delay variation measured over 30 calls
• Delay variation within the expected range of 20 ms (packet size)
=> In 3GPP:5 Tests,
take maximum to
address call to call
variation
H. W. Gierlich
Advanced speech quality testing
© HEAD acoustics GmbH
14
Jitter & Packet Loss – Simulation in Testing
HMS II.3
HHP III.1
Testsystem
Analog
Analog/Digital
Interfavce (MFE VI)
Acoustic
AES/EBU
USB
USB
IP Reference
Gateway (MFE VIII.1)
Ethernet
USB
Jitter & Packet
Loss by
TCN - Profiles
H. W. Gierlich
Network Impairment
Simulator (MFE IX)
Delay & loss
profiles from
TS 26.131, Rel. 12,
RF
synchronized
to
(LTE / 4G)
audio data
Ethernet
VoLTE Network
Simulator
Advanced speech quality testing
VoLTE Terminal Testing and Interworking
©HEAD acoustics GmbH
DUT
= Speech
= Control
© HEAD acoustics GmbH
15
7
Simulated Jitter & Packet Loss: Typical Test Results
delay_profile_20msDRX_10pctBLER_e2e
Delay [ms]
80
Example
for one terminal: 20 ms DRX cycle, 10% BLER
Delay
Drops
60
40
20
Delay [s]
0
0
0.2
0.15
DUT
20 #3 - Delay
40 vs. Time
60(delay_profile_20msDRX_10pctBLER_e2e)
80
100
120
140
160
Time [s]
Delay vs. Time: Profile 1
Delay vs. Time: Clean Channel
0.1
Increased mean
0.05to jitter
delay due
buffer increase
0
DUT
20 #3 - POLQA
40 vs. Time
60 (delay_profile_20msDRX_10pctBLER_e2e)
80
100
120
140
POLQA [MOS]
5
H. W. Gierlich
Profile 1
Clean Channel
4
POLQA scores
almost
unchanged
3
2
1
0
160
Time [s]
20
40
60
80
Time [s]
Advanced speech quality testing
100
120
140
160
© HEAD acoustics GmbH
16
Outline
Introduction
LTE Setup for Voice Communication - Obstacles
Delay and Delay Testing
Network Impairments, Delay & Jitter Profile Testing
Challenges in Superwideband & Fullband
Positional Dependant Testing of Mobile Terminals
Summary
H. W. Gierlich
Advanced speech quality testing
VoLTE Terminal Testing and Interworking
©HEAD acoustics GmbH
© HEAD acoustics GmbH
17
8
Superwideband & Fullband with EVS
How does it sound?
Frequency response tolerance schemes under discussion;
EVS FB @ 16,4 kbit/s
H. W. Gierlich
Advanced speech quality testing
© HEAD acoustics GmbH
18
Interoperability - Superwideband & Fullband with EVS
How does it sound?
….when combining 2 terminals performing within the limits
For comparison narrowband:
Optimum
Worst highpass
Worst lowpass
Fullband:
Optimum
Worst highpass
Worst lowpass
H. W. Gierlich
Advanced speech quality testing
VoLTE Terminal Testing and Interworking
©HEAD acoustics GmbH
© HEAD acoustics GmbH
19
9
Outline
Introduction
LTE Setup for Voice Communication - Obstacles
Delay and Delay Testing
Network Impairments, Delay & Jitter Profile Testing
Challenges in Superwideband & Fullband
Positional Dependant Testing of Mobile Terminals
Summary
H. W. Gierlich
Advanced speech quality testing
© HEAD acoustics GmbH
20
© HEAD acoustics GmbH
21
Artificial Head & Artificial Ear
• HATS acc. to
ITU-T P.58
• Artificial ear
acc. to ITU-T
P.57
• Positioning
acc. to ITU-T
P.64
H. W. Gierlich
Advanced speech quality testing
VoLTE Terminal Testing and Interworking
©HEAD acoustics GmbH
10
HHP IV
Video
H. W. Gierlich
Advanced speech quality testing
© HEAD acoustics GmbH
Positional Robustness
3rd octave FFT Size:4096 Overlap:0,0% Hanning
Ch.1 - Ch.2
22
1 / 2 Without Background Noise
L/dB
10
5
normal Position
dow nw ards
0
-5
-10
-15
-20
3rd octave FFT Size:4096 Overl ap:0,0% Hanning
-25
-30
200
500
f/Hz
2000
L/dB
10
5
normal Position
dow nw ards
-35
100
Ch.1 - Ch.2
0
5000
-5
-10
-15
-20
-25
-30
-35
100
H. W. Gierlich
Advanced speech quality testing
VoLTE Terminal Testing and Interworking
©HEAD acoustics GmbH
200
500
© HEAD acoustics GmbH
f/Hz
2000
5000
23
11
Positional Robustness
3rd octave FFT Size:4096 Overl ap:0,0% Hanning
Ch.1 - Ch.2
2 / 2 With Background Noise
L/dB
10
5
normal Position
dow nw ards
0
-5
-10
-15
-20
-25
3rd octave FFT Size:4096 Overl ap:0,0% Hanning
Ch.1 - Ch.2
L/dB
10
-30
-35
100
200
500
f/Hz
2000
5
normal Position
dow nw ards
5000
0
-5
-10
-15
-20
-25
-30
-35
100
H. W. Gierlich
Advanced speech quality testing
200
500
f/Hz
2000
5000
© HEAD acoustics GmbH
24
Summary
Advanced mobile terminals require more advanced and more
realistic testing procedures
Still many unresolved issues in VoLTE call setup and IMS-based
interworking
VoLTE introduces more variability in communicational quality –
e.g. for delay and delay variation
Terminal interworking concerning sound quality in superwideband and fullband from mouth to ear is critical and still an
unresolved topic
Testing taking into account time variant user behavior and signal
processing gets more important – e.g. positional robustness
testing
H. W. Gierlich
Advanced speech quality testing
VoLTE Terminal Testing and Interworking
©HEAD acoustics GmbH
© HEAD acoustics GmbH
25
12
Dr.-Ing. H. W. Gierlich
Head of Telecom Division
info@head-acoustics.de
www.head-acoustics.de
© Copyright HEAD acoustics GmbH
VoLTE Terminal Testing and Interworking
©HEAD acoustics GmbH
13