Auditory-based Subband Blind Source Separation using Sample-by-Sample and Infomax Algorithms

We present a new subband decomposition method for the separation of convolutive mixtures of speech. This method uses a sample-by-sample algorithm to perform the subband decomposition by mimicking the processing performed by the human ear. The unknown source signals are separated by maximizing the entropy of a transformed set of signal mixtures through the use of a gradient ascent algorithm. Experimental results show the efficiency of the proposed approach in terms of signal-to-interference ratio. Compared with the fullband method that uses the Infomax algorithm, our method shows an important improvement of the output signal-to-noise ratio when the sensor inputs are severely degraded by additive noise.