METHOD AND SYSTEM FOR SEPARATING ACOUSTIC SIGNALS

摘要

A system for separating acoustic signals is adapted for separating a plurality of acoustic mixtures into at least one single sound source. The system for separating acoustic signals includes: a pitch prediction module for removing temporal correlation in each acoustic mixture to result in a corresponding meta-signal; a linear prediction module connected electrically to the pitch prediction module for further removing the temporal correlation in each meta-signal to obtain a corresponding residual signal; and an independent component analysis processing module connected electrically to the linear prediction module for receiving the residual signals. The independent component analysis processing module calculates a separated matrix from the residual signals, and performs convolution of the separated matrix with the acoustic mixtures for separation into said at least one single sound source.