Physically informed stochastic modal sound synthesis
Perry R. Cook · The Journal of the Acoustical Society of America · 2001
A general and flexible sound synthesis method is described, which is based on controlled random excitations and parameter modifications of resonant filters. Parameters can be derived from simulations of particle/friction systems, from data extracted from physical experiments, from signal analysis (a developing yet still open topic of research), or from simple experimentation and ‘‘tweaking.’’ The algorithm has proven convincing and useful for modeling the sounds of many common objects and interactions in music and everyday life. Particle system examples include maracas, windchimes, coins in a pocket, ice cubes in a glass, shoveling gravel, breaking glass, etc. Friction systems include coins rolling along a table, squeaking floor panels and hinges, etc. Aside from being flexible and realistic, the algorithm is extremely efficient by nearly any standard of computation or memory use. Most exciting is the ability to drive and modify the parameters in real time, resulting in virtual objects and systems which are ‘‘playable’’ in musical performance, or easily useable for sound effects in games and virtual reality.