reported
by: Mr. Christopher Yap
Secretary (Term
1998/99)
On Friday, 29 January 1999 at
7.00pm, 13 members and 24 guests of the AES Singapore Section gathered at
Kent Ridge Digital Labs (KRDL) to listen to a talk on "Interactivity: Audio
in the New Millenium" by Dr Pete Kellock. A senior research staff at KRDL,
and also Member of the AES, Dr Kellock is a pioneer in the field of
interactive sequencing. He helped introduce multimedia to Singapore by
founding and managing the first local series of corporate short courses in
the field, and has been co-leading the development of the FlexEffex system
at KRDL, a global first in the area of interactive sound effects and music,
for the last three years.
The meeting started with an
opening address by Mr Jibby Jacob, the Chairman of the AES Singapore
Section. He noted that there was an increasing number in the audience who
were first timers at an AES Singapore Section activity. After the
introduction, Dr Kellock began his talk by giving an introduction to
Interactive Audio. He explained that I-Audio is any Music, Sound Effects(SFX),
or speech that, playing from a computer, would respond to a user's action
(or multiple users' actions) in real time. Although many musicians and audio
producers are already working with audio tools which are interactive, the
main promise of I-audio is that it will increasingly allow the "audio
consumer" to become the participant. I-audio has many applications including
computer games, virtual environments, infotainment/edutainment (CD-ROMs,
multimedia-kiosks, presentations, etc.), web sites, interactive movies,
interactive music systems, music education, and many more.
Dr Kellock questioned the view
that I-Audio is, and will remain, "a niche". He presented some interesting
facts about the existing state of the art (for example that the computer
games industry is already bigger than the television industry on some
measures) and floated some questions that may help us imagine the future for
this emerging field: Does music always fit your mood from moment to moment?
Would you play a role in a movie if you could? How many people fancy
themselves as musicians? Do humans ever resist the temptation to get more
control of their environments? The realisation of these would be dreams come
true, and represent a revolution in audio. I-Audio will certainly be the hot
technology in time to come.
|

Mr Jibby
Jacob(left) presenting Dr Pete Kellock with an AES Singapore Section
plaque in appreciation for his talk on I-Audio -
photograph by Robert Soo |
Dr Kellock next explained the
technology of I-Audio, comparing it to the traditional Linear Audio systems
of today. He explained that Linear Audio systems (for example, recorded
audio on a CD) are very rigid and allow little interaction - little more
than the select_Track/Play/Stop controls on a CD player. I-Audio gives the
user much more intimate control, for example by allowing much smaller
building blocks to be combined in real time allowing numerous "paths"
through the audio experience. Of course, such interactivity requires lots of
intelligence in I-Audio systems compared to the relatively dumb nature of
traditional Linear Audio systems: much of the composition, performance and
production, traditionally done by people in studios needs to be done "on the
fly" in the computer system, (which, by the way drives a need to move the
entire studio into the PC.)
Speaking on the challenges that
I-Audio faces, Dr Kellock cited three universal challenges. Firstly, there
is presently a trade-off between flexibility and realism. Highly flexible
systems, such as, for example a physical model of a flute (which computes
the resultant sounds according to given parameters) generally fall short of
delivering the realism of the actual flute sound. Conversely, a CD track or
even technologies such as wavetable synthesis, provide relatively realistic
sounds, but lack flexibility. The challenge is to make the sound both real
and flexible.
The second universal challenge is
the human interface. A person from the professional audio industry will be
familiar with operating sophisticated audio equipment, but these complicated
technicalities will baffle the everyday "audio consumer". Thus I-Audio tools
must provide "handles" which the average person can understand - a model of
footsteps might provide "walking speed" and "shoe type", not "filter
resonance".
The last challenge, but not at all
the least, is the problem of morphing a sound from Sound A to Sound B. This
process must generate a smooth and natural transition in which one sound
becomes the other, not a cross-fade from one into the other. This can be
useful in applications such as games where the sound morphs pleasingly when
moving from one location to another, or in web applications when the user
navigates
between web pages. Because there are
numerous paths possible in a morph, producing a good morphing algorithm
really requires not only technical but also audio and/or musical expertise.
Of the I-Audio technologies
available today, 3D Sound, or Spatial Audio as Dr Kellock prefers to term
it, has been enjoying much attention within the community. He noted that 3D
Sound is not necessarily interactive, but is often associated with
interactivity because it is such an important addition to the visual
elements of 3D virtual environments.
Dr Kellock then gave an overview
of the companies, organisations and musicians that are presently involved in
the advance of I-Audio technology. He said that sound card companies like
Creative Technology, who dominates the sound card market, are constantly
bringing pro-audio and music features into the PC. This is essential in his
opinion, as I-Audio requires some or all of the sound production to be done
in real-time on the computer. Other thrust areas of I-Audio include
Microsoft's and Intel's efforts in recent years to enable software synthesis
on the PC CPU. Microsoft's DirectSound programming interface, which provides
a common sound service for applications such as games, and also DirectMusic,
incorporating technologies like DLS and IMA, nudges I-Audio a step closer.
Dr Kellock went on to present two demos of Microsoft's DirectMusic, simple
programs in which the music changes to suit the game play interactively.
Other industry players Dr Kellock
cited are Headspace, Staccato, and Mixman. In the demo of a Mixman product,
he showed how a piece of dance music could be created almost
instantaneously. According to Dr Kellock, it is becoming possible, to an
increasing degree, for people without any formal musical training to compose
music. He also mentioned several professional musicians known for their work
incorporating I-Audio, of the more popular are Peter Gabriel (Xplora), Todd
Rundgren (No World Order), The Residents (Gingerbread Man), and his personal
favourite, Laurie Anderson (Puppet Motel). Professional organisations
involved include MMA-IASIG (I-Audio Special Interest Group of the MIDI
Manufacturer's Association), whose members include Creative, Roland, VLSI
and many others.
|

Dr Pete Kellock
(centre) posing with 37 AES Singapore Section members and guests at the
KRDL auditoriumlocated at the National University of Singapore (NUS)
after tha talk on 29 Jan 1999 - photograph by
Robert Soo |
The mood grew to be more exciting
as Dr Kellock talked about his group's work in I-Audio at KRDL. Using demos,
he illustrated their main work, the FlexEffex sound effects and music
system. Based on early work by Dr Lonce Wyse, joint leader of the KRDL team,
the FlexEffex system generates sound effects using algorithmic techniques,
producing convincingly realistic sounds. Classes of sound effects are
modeled and a high level of interactivity is provided for each model through
parameters that can be controlled - for example by virtual sliders. In a
footsteps model, the user can control the speed of the walk/run, the surface
type walked on, the weight of the person walking, the styles of the walk
(long strides, marching, etc.). No previous program has been able to create
"on the fly" footsteps on wood, then on concrete, then a combination of
concrete and mud, and so on under real-time control.
Furthermore, their algorithms
automatically introduce random variations into the footsteps, very
effectively increasing the realism of the generated sound. Dr Kellock went
on to demo, to the fascination of the audience, a guns model, a wind model,
and a bell model. In the guns model, the user had a wide variety of controls
such as the gun type, variable speed of fire, controllable pitch of the
shot, and the amount of reverberation following the shot. Each of the models
provided the ability to create extremely flexible sounds, simply by
changing parameters which would be easily understood by anyone, not just
audio experts. Dr Kellock said that the team at KRDL is presently continuing
to work on this system, increasing the number of models in their library.
In his prediction of future
trends, Dr Kellock said that while traditional Linear Audio had a strong
future, the growth rate of I-Audio would be faster, with more and more
titles produced in both versions. Over the next few years, he thinks that
I-Audio experiences will gradually approach the realism of movie experiences
today. Furthermore, with the rapidly decreasing cost of PCs, Dr Kellock
reiterated that the studio would gradually disappear inside the PC, a
crucial enabling factor for I-Audio.
Another trend will be the blurring
of the roles of producer and the consumer -the consumer will to some degree
create his own produtions drawing on material produced by others, in much
the same way some professional musicians uses drum loop libraries. In
conclusion, Dr Kellock assured the audience that most pro-audio skills will
be in demand for a long time to come, but the project goals may be different
with I-Audio requiring audio that includes behaviour. To achieve this, there
must be good communication between the pro-audio industry and the audio
software programmers.
In the question and answer
session, Dr Kellock told the audience that the team at KRDL had started this
project from scratch, and up to today had written computer code totalling
more than 10MB. Answering a question on the commercialisation of the
FlexEffex system, he mentioned that marketing work has already begun, with
the team at KRDL considering commercialisation through either a startup or
by licensing the technology. He said that because of the wide variety of
applications this technology has, FlexEffex has generated much interest when
it was shown to software houses, film companies, and others worldwide.
Following a few more very
interesting demos of morphing drum rhythms, the linking of FlexEffex from
Cakewalk, and a virtual environment game, the meeting ended at about
9:45pm.
The AES Singapore Section would
like to thank Dr Pete Kellock and his colleagues at KRDL for an eye-opening
talk, the use of the KRDL auditorium and their kind assistance during the
meeting. |