Residual cross-talk suppression for convolutive Blind Source Separation

In this paper, the problem of crosstalk suppression in convolutive Blind Source Separation (BSS) is investigated using a post-processing unit. We propose a precise technique to estimate the power spectral densities of the cross-talk components and then apply the Wiener filtering as a post-processing unit. Simulation results show that the proposed method suppresses the recognition of the crosstalk components with adding no distortion to the target signals.