We describe the use of artificial life (genetic) algorithms to generate auditory behaviors within 3D graphical environments. Small, high-quality soundfiles are algorithmically recombined and their behavior altered to create rich, non-looping sound environments. These environments consist of sound effects as well as musical tones; for example, the sound of a single water drop or a single ocean wave can be replicated and regenerated according to algorithmic rules to create a long-running nonlooping sound effect. Sonic parameters such as pitch, start time, intensity, and apparent spatial location are given initial values which are then changed in realtime as the Java-based genetic algorithms generate new values. This approach is particularly useful in Web-based 3D environments like VRML or Java3D, as it can dramatically reduce the necessity for downloading large soundfiles. It can also be used in real-world systems to produce long-running, interactive, non-looping sound environments. Thi...
Maribeth Back, Maureen C. Stone