Microphone Array Post-Filter for Target Speech Enhancement Without a Prior Information of Point Interferers

The post-filter for microphone array speech enhancement can effectively suppress noise including point interferers. However, the suppression of point interferers relies on the accurate estimation of the number and directions of point interferers, which is a difficult task in practical situations. In this paper, we propose a post-filtering algorithm, which is independent of the number and directions of point interferers. Specifically, we assume that the point interferers are continuously distributed at each direction of the plane but the probability of the interferer occurring at each direction is different in order to calculate the spatial covariance matrix of the point interferers. Moreover, we assume that the noise is additive and uncorrelated with the target signal to obtain the power spectral densities (PSDs) of the target signal and noise. Finally, the proposed post-filter is calculated using the estimated PSDs. Experimental results prove that the proposed post-filtering algorithm is superior to the comparative algorithms in the scenarios where the number and directions of point interferers are not accurately estimated.