Physical Modelling for Sound Synthesis - Game Sound Technology and Player Interaction: Concepts and Developments - page 353

by decreasing the amount of modes used for syn-
thesis and obviously the aim is to do this with the
minimum degradation of sound quality.
The first technique, which they call mode
compression , exploits the fact that humans find it
difficult to discriminate between nearby frequen-
cies (Sek & Moore, 1995). Therefore, a number of
nearby frequencies are lumped together to reduce
computational expense. The second technique,
mode truncation , considers that an oscillator
requires processing cycles regardless of how
much its output contributes to an overall sound.
To improve efficiency, a threshold is introduced
below which an oscillator's output is deemed to
be unimportant and therefore is no longer cal-
culated. The third technique, quality scaling, is
concerned with synthesizing sound for multiple
objects simultaneously. Each sounding object in
a scene is given a processing time-quota within
which to perform modal synthesis with more
important objects given more time than those
of lesser importance. At each audio callback,
modal synthesis is carried out starting with the
most important object and finishing with the least
important, allowing each object their full time
quota if required. This means the more important
objects will be rendered with a higher level of
sonic detail than those of less importance. Results
indicate that each of these techniques yields an
efficiency gain and, when used together, sound
can be synthesized for a large scene with hundreds

of interacting objects “with little loss in perceived
sound quality” (Raghuvanshi & Lin, 2006, p. 108).
Returning to the research of Menzies, 2007
saw the publication of Physical Audio for Virtual
Environments, Phya in Review (Menzies, 2007).
This paper described a project now known as
Phya: a library that facilitates physical modelling
for sound synthesis in tandem with a physics en-
gine. In this work, the author underlines the need
for creative thinking in such a project in order to
produce a “powerful synergistic perceptual effect”
(Menzies, 2007, p. 1) by combining realistic audio
with visuals. This may entail relaxing the physical
constraints on the sound production process and
giving sound designers some freedom to enhance
the characteristic sounds of a scene. Ways in which
a level of control may be given to a sound designer
are highlighted throughout the work. There is also
an emphasis on using robust, efficient techniques
and on creating a system that can be easily scaled
up to handle large environments.
By combining a knowledge of techniques
described in previous projects with new innova-
tions, Phya facilitates sound synthesis for many
different types of contact (Figure 2.). These in-
clude: simple impacts; multiple impacts, which
happen at a rate not captured by most physics
engines; scraping and rolling, including the effect
of contact jumps where the objects momentarily
break contact; grazing sounds, which occur when
an object simultaneously bounces and skids off
a surface; stick and slip, due to friction between

Figure 2. Screenshot of Phya synthesising the sounds of multiple objects. (© 2010 Dylan Menzies. Used
with permission.)