A Versatile Software Architecture for Virtual Audio Simulations

Existing real-time audio rendering architectures provide rigid development frameworks which are not adapted to a wide range of applications. In particular, experimenting with new rendering techniques is virtually impossible. In this paper, we present a novel, platform-independent software architecture that is well suited for experimenting with multichannel audio mixing, geometrical acoustics and 3D audio processing in a single framework. Our architecture is divided into two layers. A low level DSP layer is responsible for streaming and processing audio buffers using a general filter-based formalism. Built on top is an audio rendering layer responsible for general geometry-based audio rendering and configuration of the rendering setup. In particular, we introduce sequence objects which we use to define and control arbitrary sound propagation paths in a geometrical 3D virtual environment. We discuss implementation details and present a variety of prototype applications which can be efficiently designed within the proposed framework.