Spatial sound rendering for dynamic virtual environments

Abstract

We present the details of a virtual sound rendering engine (VSRE) that is being developed for virtual environments and serious games. The VSRE incorporates innovative graphics processing unit-based methods to allow for the approximation of acoustical occlusion/diffraction and reverberation effects at interactive rates. In addition, the VSRE includes a GPU-based method that performs the one-dimensional convolution allowing for the incorporation of head-related transfer functions also at interactive rates. The VSRE is being developed as a research tool for examining multi-modal (audio-visual) interactions through the simple manipulation of the acoustic environment and audio parameters (sound quality), that will, through a series of human-based experiments, allow for the testing of the effect of varying these parameters may have on immersion, engagement, and visual fidelity perception within a virtual environment. Finally, we also provide a running time comparison of several one-dimensional convolution implementations.