DT2410
Audio Technology
Autumn 2007
School of Computer Science and Communication
Department of Speech, Music and Hearing (TMH)
Web page: http://www.csc.kth.se/utbildning/kth/kurser/DT2410/autek07/
Course news are posted on this web page.
Course leader: Prof. Sten Ternström, stern@kth.se
Visiting hours Thursdays 16.00-17.00
Lindstedtsvägen 24, level 4.
DT2410 Audio Technology
Autumn 2007
1
Welcome to the course in Audio Technology!
In this paper you will find all the information that is available when the course starts.
Course news are posted when necessary on the course web page.
Please fill in the following for your own reference:
I belong to group number ___________
together with
____________________________________________
and
____________________________________________
and the topic we have chosen for our assignment is
____________________________________
DT2410 Audio Technology
Autumn 2007
2
To be examined
Written exam 4.5 hp, laboratories 1.5 hp, group assignment 1.5 hp.
Literature
Textbook
The book of choice is John Watkinson: The Art of Digital Audio, third edition. Focal Press 2001,
ISBN 0-240-51587-0. This is the best book for the course, but it is quite expensive at almost
SEK 1000, so I am not counting on many participants to buy it. It is recommended for those
who are particularly interested or wealthy. This book will be referred to as WArt.
The second choice is John Watkinson: An Introduction to Digital Audio, second edition. Focal Press
2002, ISBN 0-240-51643-5. It is about SEK 415 kr at the KTH student bookstore. This book is
adequate but somewhat thin and introductory. Analogue audio technology is entirely missing (as
in WArt). This book will be referred to as WIntro.
The third choice is John Watkinson: The Art of Sound Reproduction. Focal Press 1998, ISBN 0-24051512-9. It is about SEK 640 kr at the KTH student bookstore, which may have some copies
left. This book is older and contains sections on analog techniques which some may find
interesting, although they will not be dealt with in the course. This book will be referred to as
WSound.
The core chapters in these three books are however very similar. Your interests (or your wallet)
can guide your choice. None of the books contains much on spatial sound (lectures F2-F4 and
tutorial 1). That topic will be covered by substantial handouts and web links.
2. Audioteknik i praktiken: selected articles written by earlier course participants in 2002-2004.
Some of these are in English. These provide interesting insights into special areas and real-life
installations. Download from the web page.
3. The Swedish broadcasting corporation (Sveriges Radio) has published a pair of excellent
overview books in Swedish called Digital teknik i ljudproduktion and Digital ljudlagring på CD och
DVD. These will be distributed free of charge, courtesy of SR.
Reading guide to WIntro: Watkinson, The Art of Digital Audio
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
Why digital? Read through.
Some audio principles. Repetition, on your own.
Digital principles. Applied theory, on your own.
Conversion. Important F6, F7, Ö2.
Compression. Important F11.
Dig. rec & transmission principles. On your own.
Error correction. On your own.
Transmission. Lecture F10.
Digital audio tape recorders. Legacy technology: on your own.
Disk drives. Lecture F7.
Audio editing. On your own.
Digital audio in optical disks. Lecture F12.
Sound quality considerations. Lecture F13, parts.
DT2410 Audio Technology
Autumn 2007
3
Reading guide to WIntro: Watkinson, An Introduction to Digital Audio
The whole book is worth a good read through, but all chapters are not on the curriculum. “On
your own” means that the material is part of the course, but will not be treated in the lectures.
Chapter 1 Overview of the book. Read before the first lecture.
Chapter 2 Elementary acoustics and perception. Voluntary repetition from Medieteknik gk:Ljud
för Media, suitable for a quick read for those from the E and D programmes.
Chapter 3 A quick sweep of things audio related in digital technology fundamentals, transforms
and discrete mathematics. The beginning is easy but it advances rapidly. Good for
filling in your prior knowledge. On your own.
Chapter 4 Conversion. Important, corresponds to lectures F6 and F7.
Chapter 5 Data reduction. The principles are important, corresponds to lecture F11.
Chapter 6 Coding principles, error handling. The technique is important, but it is mostly the
error handling that is audio-specific. The principles of Reed-Solomon-codes are
important but their implementation details are out of scope (some of you may
know them already).
Chapter 7 Transmission. Lecture F10.
Chapter 8 Digital recorders. Read as an orientation, not part of the course requirements.
Chapter 9 Hard disks and audio. Lecture F9, parts.
Chapter 10 Editing digital audio. Read in preparation for Lab B.
Chapter 11 Optical discs. Lecture F12. More detailed than we have seen in earlier courses. The
Sveriges Radio booklet is perhaps an easier read, but not as detailed.
Reading guide to WSound: Watkinson, The Art of Sound Reproduction
The whole book is worth a good read through, but all chapters are not on the curriculum. “On
your own” means that the material is part of the course, but will not be treated in the lectures.
Chapter 1 Introduction. Read it through.
Chapter 2 General fundamentals. Perhaps superfluous for E, but a useful recapitulation for
Media and D. On your own.
Chapter 3 Elementary acoustics and perception. Voluntary repetition for Media and E, partially
new for D. On your own.
Chapter 4 Sound sources, is a bit like a quick preview of DT2212 Music Acoustics, for those
who are interested. Not part of the course.
Chapter 5 Microphones, more applied than in DT2400 Electroacoustics. On your own.
Chapter 6 Loudspeakers, more applied than in DT2400 Electroacoustics. On your own.
Chapter 7 Stereophony. Read as a preliminary to Dr Murphy’s lectures.
Chapter 8 Digital signals. Included.
Chapter 9 Analog techniques. On your own if you are interested.
Chapter 10 Digital recording.
Chapter 11 Optical discs. Lecture F12. More detailed than we have seen in earlier courses. The
Sveriges Radio booklet is perhaps an easier read, but not as detailed.
Chapter 12 Sound editing. Read in preparation for Lab B.
Chapter 13 Audio processing. Sections 13.1-13.15 will be handed out, to be read in preparation
for Lab A.
Chapter 14 Sound quality measurements. This is somewhat dated; better to read the blue booklet
from Sveriges Radio.
DT2410 Audio Technology
Autumn 2007
4
Additional reading
The course also has a substantial block on multichannel sound and surround sound, which is not
covered by Watkinson. Dr Damian Murphy’s handout is the main source for this block,
complemented by the article series on Surround Sound at http://www.sospubs.co.uk/ (of Sound
on Sound magazine).
Other relevant books for those who are interested:
Pohlmann, Ken C. Principles of Digital Audio. ISBN 0-672-22388-0.
Everest, F. Alton. The Master Handbook of Acoustics. ISBN 0-8306-4437-7. (mostly on the design of
studios)
Davis, Don & Davis, Carolyn. Sound System Engineering. ISBN 0-240-80305-1. (construction and
planning of large installations.)
Rumsey, Francis. Spatial Audio. Focal Press Music Technology Series, ISBN 0-240-51623-0
(Surround Sound, 2D- and 3D-reproduction)
DT2410 Audio Technology
Autumn 2007
5
Lectures and tutorials in DT2410 Audio Technology, 2007
The lecturers will assume that you have read the chapters listed below in advance. If you think of
questions after the lecture, do not hesitate to ask, we will discuss them in the tutorials.
This schedule is preliminary, and does not include the laboratory sessions.
F1
Mon 29/10 1012 Fantum
F2
Tue 30/10
10-12 Fantum
F3
Wed 31/10
10-12 Fantum
Mon 5/11
8-10 Fantum
Mon 5/11
15-17 Fantum
Tue 6/11
8-10 D33
Wed 7/11
8-10 D33
Mon 12/11
13-15 D33
F4
Ö1
F5
F6
F7
F8
Ö2
Ö3
F9
F10
Ö4
F11
Ö5
F12
Ö6
F13
F14 +
F15
Tent
Tue 13/11
8-10 D33
Wed 14/11
10-12 Fantum
Thu 15/11
10-12 D33
Mon 19/11
13-15 Fantum
Wed 21/11
10-12, D35
Thu 22/11
10-12, D33
Mon 26/11
13-15, E34
Tue 27/11
8-10, D33
Wed 28/11
13-15, D33
Mon 3/12
13-15, (D35)
Thu 6/12
10-12, D33
Mon 10/12
13-17 Fantum
Wed 19/12
10-11 D33
Mandatory attendance. Course overview. Division into groups.
Allocation of laboratory schedule. Introduction to the group assignments.
Sten Ternström (WIntro chap 1-3, WSound chap 1).
”You are surrounded!” Background & theory for spatial sound. (WSound
chap 7.11+extra handout material.) Guest lecturer dr Damian Murphy,
dept of Electronics, University of York.
Recording for stereo and spatial sound. Damian Murphy. Repeat
microphones in WSound chap 5 and stereo in WSound chap 7.
Alternative 3D sound systems: Ambisonics, wavefield synthesis and
binaural/transaural presentation. Damian Murphy.
Numerical methods for room acoustics simulations. DM, ST
Software architectures for audio, part 1. How audio is handled in different
operating systems, in many layers of protocol. ST
A/D-D/A-conversion part 1. Sampling and quantisation for audio.
(WIntro chap 4, WSound chap 8.1-8.9) Svante Granqvist, TMH Music Acoustics
A/D-D/A-conversion partl 2. Dither, convertors. (WIntro chap 4, WSound
kap 8.10-8.15) Oversampling, noise-shaping, single-bit-converters (chap
8.16-8.21) Svante Granqvist
Software architectures for audio, part 2. Common high-level tools and
applications. (Status check on group assignments.) ST
Exercise on A/D-D/A. Methods for sampling-rate conversion. SG, ST
Hardware for audio part 1: IC's and buses for digital audio, construction
examples (WSound chap 10 selections; handout materials). ST
Hardware for audio part 2: overview of modern pro audio equipment
(WIntro chap 9) ST
Audio transmission: types of cables and networks (WIntro chap 7) ST
Group assignments: status reports, problem solving.
ST
Compression of audio signals (WIntro chap 5, WSound chap 13.16-13.25)
Svante Stadler, KTH-S3 Sound & Image Processing Lab (SIP)
Round table discussion on design criteria for different types of audio
installations ST
Optical discs in audio: CD, DVD, MiniDisc. DVD-Audio and SACD.
(WSound chap 11, WIntro chap 11, Swedish Radio's green book).
Field trip
(Your exam problem to be submitted electronically by 18.00)
Audio in broadcasting: DAB, surround, current distribution
technologies (WIntro 7.12-7.19+ SR's blue book). Sound quality
aspects. (WSound chap 14). Lars Jonsson, technical strategist, Swedish
Broadcasting Corporation.
Assignment presentations! Open to all audiophiles at KTH. Course
conclusion. The home exam is handed out. Sten Ternström
Deadline for turning in the written exam.
DT2410 Audio Technology
Autumn 2007
6
Tutorials with Q/A sessions and weekly reports
Audio technology is a smörgåsbord of engineering. The ingredients include mathematical
methods from Fourier analysis, the Laplace transform, source coding, signal theory,
combinatorics, and more. The potential for mathematical detail is immense, and from your time
at KTH you will know where to look. However, a more important objective of this course is to
build an overall perspective at a qualified level, and to show how different technical systems work
together. The textbook is fairly extensive, but it does not contain exercises, and not much
mathematics. To Swedes, Watkinson’s English may appear somewhat concentrated, opinionated
in places, and not always easy to assimilate.
Questions: In our tutorial sessions we therefore make extra room for questions and answers.
This is structured as follows: first agree within your group on things that need to be clarified
(often someone else in the group will be able to explain). Unsolved issues can be submitted to
me, collected in one weekly e-mail per group. Please put "Audio? Group xx" in the subject line, it
will help me a lot. This will give me some time to prepare a proper answer for the next time we
meet.
Signs of life: Even if you do not have any questions, please send me exactly one e-mail per
week from each group to let me know how your assignment is going. Put "Audio! Group xx" in
the subject line, please.
We will have tutorial sessions at least on spatial audio, sampling rate conversion and system
design considerations. Two tutorial sessions will be available for field trips.
Laboratories
The laboratory sessions will take place in the weeks number 46-49, Fri 16 Nov to Fri 7 Dec. The
lab rooms are located on the basement floor (level one) in the TMH building, Lindstedtsvägen
24. On each occasion, two groups of three people will be active. The lab groups are the same as
the assignment groups.
The list of lab sessions will be posted in the ground floor entrance of the TMH building
(LV24), and it will be closed before the first lab session. You must make sure that your group is
signed up once for each of the four labs A-D. Please note also that lab A must be done before
lab B. Lab instructions will be provided for download from the course web page.
Lab A is about signal manipulation with the most usual tools: level faders, equaliser, dynamics
processor, and reverb unit. You will make measurements and listen to what the machines
do to the sound.
Lab B is a downmix session in which you will apply what you learned in Lab A. A multitrack
material of moderate quality needs to be mixed down to stereo. Your task is to make it
sound as good as possible.
Lab C is about illusory reproduction of monophonic speech in a room. Given a selection of
microphones, loudspeakers and equalisers, try to create the illlusion that a recorded voice
is actually someone talking in the room. This is an exercise in listening and applying
technical knowledge.
Lab D is about 3-D sound reproduction with Ambisonics. You will test various loudspeaker
configurations with up to eight channels, and try to assess the localisation of sounds in
different directions.
Lab crew
Lab A and B: Kahl Hellmer
Lab C and D: Marco Fabiani
DT2410 Audio Technology
hellmer@kth.se
himork@csc.kth.se
Autumn 2007
7
Group assignment
(1 p)
Goals
A.
Be able to include audio input and output (recording and playback) in your own
programs.
B.
Be able to modify an audio signal in real time using high-level tools.
Task A: Soundfile I/O
You need first to choose three things:
1. Operating System
Windows XP / Vista – Mac OS – Linux – other.
Choose what you like, as long as you can demonstrate it in class.
The TMH student lab has Windows workstations. You do not need to use these; in fact, it is
probably easier to work on your own computer if you have one.
2. Audio API
•
Portaudio library (all OS)
•
OpenAL (all OS)
•
CoreAudio (Mac)
•
MME or DirectX or MCI (Media Control Interface, only simple calls) (Windows)
3. Programming language
C, C++, C#, Java, Python, Smalltalk, Delphi, other.
Choose one that you like and that is available to you.
The purpose is not to learn a programming language, but to learn how to program for audio.
Not all combinations of the above are practical or possible. Your programs should demonstrate
as many as possible of the following actions (from easier to more difficult):
1.
2.
3.
4.
5.
6.
Play a given soundfile in its entirety, using a single high-level function call.
Play user-selected portions of a given soundfile, with user-selectable sampling rate. The
program should display level meters in decibels that move appropriately during playback.
Play a long soundfile, with continuous event-driven double-buffering of the disk I/O.
Display real-time dB meters of two live inputs.
Record a short soundfile in its entirety.
Record a long soundfile, with continuous event-driven double-buffering of the disk I/O.
The program must check for available disk space. The program must ask the user to
confirm overwrite of an existing file.
Task B: real-time modification of audio with remote control over MIDI
Tools – choose one of
• PureData (freeware, Windows – Mac – Linux)
• AudioMulch (Windows only, free 60-day demo)
• Reaktor (1 licence only at TMH)
• Aladdin Interactive DSP (requires additional DSP hardware; only in the course lab room.)
Equipment for the assignment
The course laboratory with WinXP+PC, diverse audio software and MIDI-controlling hardware
can be booked for half a day at a time, unless the room is booked for other courses. If this is not
practical, you may borrow audio equipment from TMH for a few days.
DT2410 Audio Technology
Autumn 2007
8
Alternative assignment: Write a technical article
If you do not want to do the programming assignment, you may instead choose to write a
technical article on some specialised topic of your own choice within audio technology. Sound
examples to go with it are of course encouraged. Examples of such topics are given below or can
be found in the collections from earlier years. The assignment should be started on as soon as
possible, and the idea is that you do your own research in the field, on the web and in libraries.
The TMH library contains some material of interest.
The prime readership for these articles is the participants of this course. The text must contain a
list of references like that at the end of each chapter in the textbook. There must also be an
account of how the different subtasks were allocated to the group’s members. The assignment
text may be written in Swedish if you so prefer, but the presentation on 10 December must be
given in English, including any slides. There is a Microsoft Word document template with a
suggested format for the articles available on the course web page.
The group assignments must be handed in no later than 4 December. I will arrange for the
collected articles to be printed, so that you can receive them all on 10 December when you
collect your materials for the home exam.
Extra literature
The Internet is all good and well, but some knowledge is found only in libraries. The TMH
library has a reasonable stock of sound books and journals. The library on level 4 contains
journals and on level 5 you will find books. Please note that it is not permitted to remove any
journal or book from the library – the material you need must be photocopied. For access to the
library, please contact Sten.
Suggested topics for an article
Earlier topics have included:
Existing sound installations: Cosmonova, the Royal Opera, the Royal Dramatic Theatre, the Globe
Arena, cinemas.
Broadcasting: surround sound at SR,
Types of sound systems: in churches, theatres, cinemas, cars, radio stations
Sound in computer games, Speech in noisy environments, Audio on computer networks,
Reverb simulation techniques, Data reduction techniques, Pro audio versus consumer audio,
Binaural synthesis, HRTF’s, Mastering, Comparing DVD-Audio to SACD, Archiving sound
You are welcome to re-use these topics if you like, if you can improve on the previous paper.
Some topics we have not yet seen
• Train/underground: sound for stations and carriages
• Civil aviation – PA and entertainment systems in airplanes
• Conference audio: speaker mikes, participant mikes, feedback suppression
• Teleconferencing (from conventional loud telephones to full virtual reality)
• Sports arenas
• Language training systems
• Systems for simultaneous interpreters
• Covert information technology (spy mikes etc)
• A selection of interesting recent audio patents
DT2410 Audio Technology
Autumn 2007
9
The article should always account for
• Conditions that are special for the application
• Specific requirements for the sound
• Other specific requirements, such as reliability, coordination, security aspects
If you are describing a type of installation there should be a case study of a representative system,
including approximate costs of acquisition, installation, operation and maintenance.
Vocational information
• Is there such a thing as an audio labour market? Who works with audio and how many?
• An overview of professional journals and magazines
• Book review
Improving the course laboratories
If you have suggestions for a new lab session, why not make it your assignment and work it
through?
Presentation of group assignment works
The presentation session for both software and technical article assignments is on 10 December,
13.15 until finished, probably about 15.30. The assignment presentations are to be about 20
minutes in duration. The presentation session will be open to all interested, and will be posted in
the main TMH seminar series. This means that the presentations must be of good quality and
should be of interest also to audio enthusiasts who have not followed the course.
DT2410 Audio Technology
Autumn 2007
10
Examination
Write an exam problem, and do a written exam at home.
As part of the course examination, you must write a proposed problem for the home exam. Each
problem is to be worth 10 points. A suggested problem must include
1. The problem text, with figures if necessary
2. A complete solution
3. Detailed criteria for corrections (what earns points?)
If the solution requires independent research (which is admissible) then sources should be
suggested. If you prefer to write the problem in Swedish, that is acceptable, but I will translate it
into English when composing the final exam.
The exam problem is an individual task which you must do on your own. By submitting a
problem you also attest that it is your own work, and that you intend to take the written exam
(Sw. tentamensanmälan). Your proposed problem must be submitted to me by e-mail no later than
Monday 3 December at 18.00. Don’t forget to put your name on each page, even in electronic
documents. Please also put your name in the file name of the e-mail attachment, if used. It is also
admissible to enter submissions on paper, or by fax to 08-790 7854, attention Sten Ternström.
I will select the best proposed problems, perhaps adjust them if necessary, and will compile a
home exam with five problems for you to do at home. Hence the maximum total will be 50
points. As you can imagine, the bonus for having your proposed problem selected is that you can
solve it very quickly. Look out, though, since I may have tweaked it a little! Preference will be
given to compound problems that have a coherent theme and that require the solver to think
independently. A problem whose answer is simply a quote from the book is not very interesting.
The solutions may be narrative, or mathematical, or both. Multiple choice problems are allowed
if they are non-trivial and if the probability of guessing the right answer is low. Be creative!
The final exam text and the cover paper will be handed out at the end of the final project
presentation on Monday, 10 December. The personal envelope you pick up there will contain
also my comments to the problem you proposed; and a course evaluation sheet, which you are
asked to complete.
Your completed exam is due no later than 11.00 on Wednesday, 19th December. I will be
sitting in room D33 between 10.15 and 11.00 to receive the exams. This coincides with the exam
time given on the central schedule. Exams that are returned too late will not be graded above C.
Exams may not be returned by e-mail or by fax - your signature on paper is needed. The
solutions may be written in English or in Swedish, but not both (excepting certain technical
terms, which may not be available in Swedish, such as “dither.”)
When you communicate as an engineer to other people, it is important to be able to explain
things correctly in words. Because you will have time to think about your answers, I will be more
particular about the wording of your solutions. This means that I may subtract points if your
wording is incorrect, unclear or incomplete, even if I can more or less understand what it is that
you actually mean. Remember to cast the narrative into your own words, and do not copy the
running text of others (the CSC Code of Honour applies). Always identify your sources.
The preliminary grading scale is A: 50-45, B:44-40, C:39-35, D:34-31, E: 28-30, F:25-27, Fx: 2224. Exams from earlier years will not be made available until after the deadline, on request.
DT2410 Audio Technology
Autumn 2007
11
Earlier students have found this form of examination quite challenging, and also that it is a good
incentive for really penetrating the subject.
To pass the laboratories you must be present at all four labs. If you are absent because of illness
or other very compelling reasons, we will arrange an extra session.
To pass the group assignment, all members of the group must be actively involved, and the
course leader must approve the result. A good quality criterion is that your material should be
usable “as is” in the course materials for the following year.
The final grade on the course will be determined both (a) from the points you reach in the
written exam, (b) from the quality of your assignment work and the presentation.
DT2410 Audio Technology
Autumn 2007
12
Course schedule DT2410 Audio Technology, period 2, autumn 2007
Lab times are preliminary; they will be adjusted for your schedules
in the first week of the course.
v 44
må 29/10
08:00
09:00
10:00 F1 Intro
11:00 Fantum
12:00
13:00
14:00
15:00
16:00
17:00
v 45
08:00
09:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
må 5/11
F4 Surround 3
Fantum
ti 30/10
on 31/10
F2 Surround 1
Fantum
F3 Surround 2
Fantum
ti 6/11
F5 Software 1
D33
on 7/11
F6 A/D-D/A 1
D33
to 1/11
fr 2/11
to 8/11
fr 9/11
to 15/11
fr 16/11
Ö1 Surround
Fantum
v 46
må 12/11
08:00
09:00
10:00
11:00
12:00
13:00 F7 A/D-D/A 2
14:00 D33
15:00
16:00
17:00
ti 13/11
F8 Software 2
D33
v 47
må 19/11
08:00
09:00
10:00
11:00
12:00
13:00 F9 Hardware 2
14:00 Fantum
15:00
16:00
on 14/11
Ö2 Sampling
Fantum
Ö3 Hardware 1 L1
D33
L2
ti 20/11
on 21/11
to 22/11
fr 23/11
L3
F10 Transmiss. Ö4 Group work
D35
D33
L4
(continued)
DT2410 Audio Technology
Autumn 2007
13
(continued)
v 48
må 26/11
ti 27/11
on 28/11
08:00
Ö5 Round table
09:00
D33
10:00
11:00
12:00
13:00 F11 Compress.
F12 Optical disc
14:00 E34
D33
15:00
16:00
to 29/11
v 49
må 3/12
08:00
09:00
10:00
11:00
12:00
13:00 Ö6 Field trip
14:00 (D35)
15:00
16:00
17:00
to 6/12
ti 4/12
L5
L6
on 5/12
fr 7/12
L9
F13 Broadcast
D33
F
D33
L8
L7
v 50
må 10/12
ti 11/12
09:00
10:00
Lab
11:00
12:00
13:00 F14-15
14:00 TMH Fantum
15:00 Project present.
16:00
17:00
v 51
09:00
10:00
11:00
fr 30/11
må 17/12
DT2410 Audio Technology
ti 18/12
L10
on 12/12
to 13/12
fr 14/12
to 20/12
fr 21/12
F
D33
on 19/12
Turn in exam D33
Autumn 2007
14