SoundScapeGenerator: soundscape modelling and simulation

Abstract

This paper describes SoundScapeGenerator, a generic sys- tem for modelling and simulating soundscapes both in real and non-real time. SoundsScapeGenerator features algorithms for three-dimensional sound-localisation and is able to model complex spaces. The system relies on abstract rule descriptions (generators) to simulate deterministic, “car- toonified” (loosely modelled) sequencing of sound events. It further supports virtual P.O.H. (point of hearing) and is able to render the final output in various channel con- figurations. Based on generative algorithms, the Sound- ScapeGenerator implementation allows real-time user interaction with ever-changing soundscapes, and it can easily communicate with other applications and devices. Finally, we introduce SoundScapeComposer, a higher level module developed for the SoDA project (dedicated to soundscape automatic generation), that has been built on top of SoundScapeGenerator, hiding most of the latter’s details to the user.