Prospective View on Sound Synthesis BCI Control in Light of Two Paradigms of Cognitive Neuroscience - Guide to Brain-Computer Music Interfacing - page 64

natural way, led to numerous interactive applications in musical contexts ( Journal
of New Music Research, special issue “ Enaction and Music ” 2009). A general

review on fundamental research in the field of enactive music cognition can be

found in (Matyja and Schiavio 2013 ).

4.3 Sound Synthesis Processes

To date, two approaches to synthesize sounds could be highlighted: sound synthesis
based on the modelling of physical sources (from either physical or signal per-
spectives) and sound synthesis based on the modelling of perceptual effects.
Interestingly, these synthesis approaches could be linked to the two paradigms
related to our perception of the surrounding world (i.e. approaches inspired by the
representational-computational and the enactive paradigms, cf. Fig. 4.2 ) described
in the previous section.

4.3.1 Two Approaches for Sound Synthesis

4.3.1.1 Modelling the Physical Sources
In the case of sound synthesis based on the modelling of physical/vibrating sources,
either the mechanical behaviour or the resulting vibration of the sound source is
simulated.
Physical synthesis models that simulate the physical behaviour of sound sources
can either be constructed from the equations describing the behaviour of the waves
propagating in the structure and their radiation in air (Chaigne 1995 ) or from the
behaviour of the solution of the same equations (Karjalainen et al. 1991 ; Cook
1992 ; Smith 1992 ; Bilbao 2009 ). Physical models have been used to simulate a
large number of sound sources from voice signals to musical instruments. Several
synthesis platforms based on physical modelling are now available, such as
Modalys that is based on modal theory of vibrating structures that enable the
simulation of elementary physical objects such as strings, plates, tubes, etc. These
structures can further be combined to create more complex virtual instruments
( http://forumnet.ircam.fr/product/modalys/?lang=en)n.d ) . Cordis-Anima is a mod-
elling language that enables the conception and description of the dynamic
behaviour of physical objects based on mass-spring-damper networks ( http://www-
acroe.imag.fr/produits/logiciel/cordis/cordis_en.htmln.d ) . Synthesis models for
continuous interaction sounds (rolling, scratching, rubbing, etc.) were proposed in
previous studies. In particular, models based on physical modelling or physically
informed considerations of such sounds can be found (Gaver 1993a ; Hermes 1998 ;
van den Doel et al. 2001 ; Pai et al. 2001 ; Rath and Rocchesso 2004 ; Stoelinga and
Chaigne 2007 ). In particular, Avanzini et al. ( 2005 ) developed a physically based
synthesis model for friction sounds. This model generates realistic sounds of
continuous contact between rubbed surfaces (friction, squeaks, squeals, etc.). The