Abstract

This paper proposes a new method of signal extraction from noise-added signal, addressing on the problem of segregating two acoustic sources as a model of acoustic source segregation. This method models some constraints of auditory scene analysis (ASA) and makes it possible to segregate two acoustic sources using the amplitude envelope and the phase deviation (input and output phase) obtained from the output of a wavelet filterbank. Using these three physical clues, the amplitude envelope and the output phase are determined; then, the input phase is determined using the physical constraints translated from heuristic regularities, changes in an acoustic event and gradualness of change, as proposed by Bregman. As an example of segregation using the proposed method, we seek to provide a solution for the problem of segregating two acoustic sources in which a sinusoidal signals added to a bandpassed noise. In particular, if the parameters of the proposed model are set to the human auditory properties, it can be a computational model of co-modulation masking release, which makes extraction of sinusoidal signals when such signals are added to AM bandpassed noise simpler while making the extraction of sinusoidal signals difficult when such signals are added to bandpassed random noise.