Spherical Sound Scene Analysis
Ramani Duraiswami, Nail A. Gumerov, and Adam E. Donovan.
The sound arriving at a point is directional, and a spherical setting is the natural one to model it, as well as to measure it. In this talk I will explore two representations of directional sound - in terms of spherical wave functions and in terms of plane wave expansions. Error bounds for accurate representation,
and the conversion between various basis will be presented. I will then discuss the capture of sound using spherical microphone arrays, and the conversion of the captured sound to these basis functions. Manipulation of the captured sound for various applications, such as HRTF based rendering for music and telepresence applications will be presented. In addition, application of the spherical array form to spatially selective sound capture (beamforming) will be discussed. Finally, I will present the use of spherical sound capture to create audio images. The registration of these images with visual images, and their joint use for audio-visual scene analysis will then be demonstrated. The talk will conclude with a discussion of open theoretical questions, and possible future applications.