Real-Time Convolutive Blind Source Separation Based on a Broadband Approach

In this paper we present an efficient real-time implementation of a broadband algorithm for blind source separation (BSS) of convolutive mixtures. A recently introduced matrix formulation allows straightforward simultaneous exploitation of nonwhiteness and nonstationarity of the source signals using second-order statistics. We examine the efficient implementation of the resulting algorithm and introduce a block-on-line update method for the demixing filters. Experimental results for moving speakers in a reverberant room show that the proposed method ensures high separation performance. Our method is implemented on a standard laptop computer and works in realtime.