Time Frequency Masking Strategy for Blind Source Separation of Acoustic Signals Based on Optimally-Modified LOG-Spectral Amplitude Estimator

The problem of Blind Source Separation (BSS) of convolved acoustic signals is of great interest for many classes of applications such as in-car speech recognition, hands-free telephony or hearing devices. The quality of solutions of ICA algorithms can be improved by applying time-frequency masking. In this paper, a number of time-frequency masking algorithms are compared and a post-processing algorithm is presented that improves the quality of the results of ICA algorithms by applying a modified speech enhancement technique. The proposed method is based on a combination of "classical" time-frequency masking methods and an extended Ephraim-Malah filter. The algorithms have been tested on real-room speech mixtures with a reverberation time of 130 - 159 ms, where a SIR-improvement of up to 23dB has been obtained, which was 11 dB above ICA performance for the same dataset.