Sound spatialization for virtual environments can be realized in software [Intel97] [PK97] or hardware [WSFF90] [AMY +98] [RM95]. The sound spatialization framework developed at our university supports various backends via a common interface, allowing application development in virtual environments for various platforms. There are many toolkits which support sound spatialization; the VRML2.0 [CB97] specification is one popular example. It provides sound spatialization using a simple sound node model for the World Wide Web, allowing one to gain a feeling for the space easily, although it does not model the size of the space or important reverberation parameters. The Java3D [SRD +97] API is a more advanced toolkit in that regard, the soundscape node featuring an application region and aural attributes. The aural attributes allow the following parameterization: Attribute Gain (amplitude and sound speed), Reverberation Model (coefficient and reverberation delay, reverberation order), Distance Filter (filtering), and Doppler Effect Model (scale factor and Doppler velocity). Each set of attributes corresponds to a single application region.

``A soundscape is a spatial region in which sound sources, sound sinks, and interactions between objects with sonic results can share common media, geometries, and spatial mappings.''[Her97]

kimitaka ishikawa