Geometric source separation: merging convolutive source separation with geometric beamforming

Blind source separation of broad band signals in a multi-path environment remains a difficult problem. Robustness has been limited due to frequency permutation ambiguities. Increasing the number of sensors allows improved performance but introduces degrees of freedom in the separating filters that are not determined by separation criteria. We propose to further shape the filters and improve the robustness of blind separation by including geometric information such as sensor positions and localized source assumption. This allows us to combine blind source separation with adaptive and geometric beamforming leading to a number of novel algorithms collectively termed "geometric source separation". Performance comparisons on real room recordings for 2 and 3 simultaneous sources are presented.