Abstract:

Microphone arrays can be used for high-quality sound pickup in reverberant and noisy environments. The beamforming capabilities of microphone array systems allow highly directional sound capture, providing a superior signal-to-noise ratio (SNR) when compared to single microphone performance. Recent work [Rabinkin et al., ``Optimum microphone placement for array sound capture,'' Proc. SPIE 8/97] has addressed the issue of microphone placement for optimized array performance. A Monte Carlo procedure was described to evaluate average sound capture SNR for a given microphone configuration based on geometry-related array performance statistics. Numerical optimization was performed for particular classes of sensor geometries based on the evaluated SNR. A uniformly distributed uncorrelated additive noise model was used to evaluate performance. An improved performance model is described which accounts for the correlated nature of additive noise generated by acoustic reflections of the source in a reverberant environment. The Allen and Berkley image method is used to generate the reflective noise portion. Optimization is performed for various acoustic conditions and array performance is compared for different reverberation levels.