Behaviour, Structure and Causality in Procedural Audio - Game Sound Technology and Player Interaction: Concepts and Developments - page 323

Depth of Model and the
Question of realism

exposing the concept of deferred form which,
I believe, is a vital facet of procedural audio.
In this analogy the performance is “played to
the audience”, perhaps a little different on each
night. It's adaptive and plastic. So, my favorite
procedural sound metaphor is a football game. It
encapsulates the essence of football, the players,
the rules, and the ball, the spectators and, at each
juncture, it is entirely causal: Every move sets up
every subsequent move, yet its unfolding form and
precise outcome are unpredictable even though the
essential experience of it remains football . Natural
sounds are like this. Given identical contexts,
no two snapping twigs would produce the same
sound pattern. Game sound is itself, like a game.

Clothes you put on don't change who you are.
A soldier is more than a uniform and a doctor is
more than a stethoscope. Thus it is with sampled
sounds that suffer the weakness of exposing only
surface features which will not withstand harder
examination. One of the things people mean when
describing sounds as “two dimensional” is an
inability to work in more than one or two tightly
constrained contexts. The idea of multi-sampling
only makes sense insofar as all entity interactions
can be enumerated. With combinatorial growth this
clearly becomes a nonsense. Thus, sampled sounds
are straw men, exposed by cursory examination.
They capture the look without the feel, the appear-
ance without the behaviour. To use a computer sci-
ence term, they are brittle . The breakdown comes
quickly under the functionalist gaze (if ears can
cast such a thing) of a listener who cannot refrain
from causal analysis (causal listening in Chions
sense of seeking an identifying mechanism behind
the surface signal). The senses are jarred and the
familiarity of being as change (as Heidegger might
have it) deeply upset by hearing exactly the same
thing twice (it's completely unnatural), especially
in rapid temporal proximity where it may still be
in echoic memory. This is the experience of the
sound sample. A whole generation has grown used
to it. And yet, with a sound photograph (sample)
the best, most expensive microphones, studios
and playback technology only help to expose
the fakery as soon as the same sound plays for a
second time. Doesn't this make bogus the whole
quest for “realism” in game audio?
Such inflexibility has long been the bane
of game audio developers struggling to avoid
repetition and uniformity (comically depicted in
an episode of “The Simpsons” where Sideshow
Bob repeatedly steps on garden rakes, trapped in
a “simulacra hall of mirrors”, repeating the same
experience and briefly reducing the story to the
level of a cheap computer game). Many systems

behavioural breadth

The football game is bounded by immutable con-
straints, such as the players not leaving the field
and flying about. It is a fair model for physical
vibrations and acoustic propagation where unfath-
omable complexity means that no two time domain
waveforms will ever be the same, yet underlying
circumstances mean that two indistinguishable
strikes can sound from the same bell hit in dif-
ferent places. This non-contradiction requires a
model with constraints that allow a wide enough
resulting range for the domain of stimulations
(behavioural parameters), and narrow enough for
the sound to be perceptually defined. We could
call this the reality window .
In the idiophonic case, the breadth of behav-
ioural parameters may extend to include different
excitation points or methods. For example, we
may wish to construct a tin can model that can be
impacted on its base or sides, be scraped, rolled
or crushed, and yet still remain recognisable as
the same object under different conditions of
behavioural stimulation (see Vicario, 2001).