Acoustic Scene Analysis Using Modal Array Signal Processing

Abstract

In this chapter, the concepts of classical acoustics and wavefield decomposition as derived in Chapters 2 and 3 on the one hand, and classical array signal processing, as discussed in Chapter 4, on the other hand, are combined to yield novel solutions for sensor array signal processing tasks. As before, the signal processing tasks considered are beamforming, i.e. waveform estimation (WE), and the localization of possibly multiple acoustic sources, i.e. parameter estimation (PE). In addition, the problem of how to determine the number of active sources in a wavefield is addressed. One of the central observations to be made here, which makes the combination of the two paradigms possible, is that the individual circular or spherical harmonics resulting from the wavefield decomposition step can be regarded as individual sensors in the classical sensor array processing framework. In the following, the circular and spherical harmonics are jointly denoted as eigenbeams, a concept first applied to acoustic signal processing by Elko et al. in [EKM03] and [ME04].