Co-channel Speech Detection Algorithm And Useble Speech Reconstruction

Co-channel speech occurs when one speaker's speech is corrupted by another speaker's speech. A co-channel detection system could provide information to suspend the operation of a speech processing system whose operation would be degraded if it were processing co-channel speech.A new methods of co-channel speech detection based on CAMDF (Circularaverage magnitude difference function)are developed in this thesis. The algorithm not only simplifies the co-channel speech detection ,but also effciently improves the precision of co-channel speech detection.Lots of experiment show that the performance of the algorithm is better than other methods that are based on the wavelet transform or the SAPVR..Speech that is corrupted by interfering speech, but is still usable for applications such as speaker identification is referred to as "usable speech". The speech segments are declared "usable" for speaker identification based upon a Target-to-Interferer energy ratio (TIR). There is none of method about usable speech segments reconstruction under co-channel conditions.A new ideal and method of usable speech reconstruction based on wavelet transform is developed and investigation of the method reconstruct usable speech above 62%.