Behaviour, Structure and Causality in Procedural Audio - Game Sound Technology and Player Interaction: Concepts and Developments - page 329

unified scheme. Diversity and openness are the
watchwords for the future of game audio.

We have an innate understanding of size and
scale as revealed by sound. Modal parameters
that change with size for constant shapes (scaled
Eigenvectors) are shown by some experiments
(Kunkler-Peck & Turvey, 2000) to be universally
interpreted. Changes of parametric scale (such as
speeding up, slowing down playback rate, or shift-
ing formants in a fixed ratio) to indicate changes
of size, is a common sound design technique. For
modal synthesis this technique is understood as
a means to change the apparent size of simple
rigid bodies.
For an excellent annotated bibliography of
psychoacoustic aspects see Giordano (2001).
Moving on to more complex sonic objects, large
scale structures (in which propagation cannot be
taken as uniform), heterogeneous, polymorphic,
composite, and variable size objects presents a
challenge. To summarise the work above, and that
presented in Designing Sound , it is sufficient to say
that we can understand and synthesise sounds from
objects that change in shape or size (like poured
liquids), or have non-linear discontinuities (such
as the twanged ruler against a table).
As the size and complexity of an object in-
creases we can no longer treat it as a collection
of modal resonators connected without concern
for propagation. We must consider the journey
of sound waves through a series of sub-objects,
from some excitation point or temporal origin,
towards the listener's ear through a radiation
surface, intervening medium and acoustic context.
At this point we introduce causality and the flow
of energy into our model. Space, viz. size, now
becomes relevant to the modal frequencies and to
the time domain structure. An example is a ticking
clock model in which power is transferred from
a sprung store of elastic potential energy along a
series of interconnected cogs and wheels towards
a final radiator which is the face and hands of the
device. Each sub-object can be modelled modally,
using a variety of methods (additive, subtractive,
or non-linear), but the overall behaviour that makes
the sound object a clock, as opposed to a collection

strUctUrE, sPAcE
AND cAUsALItY IN
PrOcEDUrAL sOUND

We now move on to a deeper discussion of sound
object design. In chapter 3 of Designing Sound , the
basic concepts of physical sound were explored,
at least as they pertain to rigid body vibrations
in objects whose size and shape remains fixed.
Within the framework of current game physics
almost all such sources are taken to be the result
of a collision and thus the energetic source is ki-
netic. Shape (Kunkler-Peck & Turvey, 2000) (van
den Doel & Pai, 1998) and material constitution
as a starting point for the synthesis of idiophonic
sounds is established. That a particular sound cor-
responds to a set of material properties, shape, size
and excitation pattern and position, is confirmed
by many whose work in modal and waveguide
methods produces excellent results.
A deeper investigation of the role of structure
would make a fascinating thesis on its own. The
question of whether such a correspondence is
strict (injective/one-to-one and surjective/onto) is
left aside. Benson (2007, pp. 119-120) references
the work of Gordon, Webb, and Wolpert (1992)
regarding the Dirichlet spectrum of homomor-
phic plates. In short, it's possible for different
shapes, similarly excited at different points, to
sound identical. This leads to a useful simplifi-
cation where for some object we can ignore the
physical arrangement of different sub-parts and
consider only their connectivity, like a net-list
that represents only their logical relationships,
not their spacial relationships. This is one essen-
tial feature of modal synthesis, an overview and
bibliography for which is given in Adrien (1991)
with a more recent discussion of modal methods
in Bilbao (2009).