A comparison of sound localization methods in a room environment using a spherical microphone array.
Colin L. Barnhill, James E. West · The Journal of the Acoustical Society of America · 2010
Enclosures are one of the most challenging environments for source localization, especially for teleconferencing where the main goal is to enhance desired speech signals while suppressing noise. This can be accomplished using beamforming but the speech source locations must be known. Three passive methods will be compared for room environment localization: generalized cross-correlation (GCC), spherical harmonic-based multiple signal classification (MUSIC), and spatial power signal characteristics and gradients (SPSGs). SPSG is a new localization algorithm designed for room environments based on wideband third order beamforming and power signal characteristics, differences, and gradients. MUSIC is a subspace method that relies on an eigen-decomposition of the beam correlation matrix and GCC is a time delay estimation algorithm based on microphone signal correlation. MUSIC and GCC are methods that perform well in their optimal environments but suffer in noisy and reverberant conditions. Results will be shown for multiple source situations with reverberation and noisy interferers.