An adaptive noise decorrelation technique for stereophonic acoustic echo cancellation

Abstract

In this paper, a new signal decorrelation technique, based upon adaptive noise addition is presented for stereophonic acoustic echo cancellation (SAEC), with insignificant additional computational complexity. The added noise level is adjusted according to the characteristics of the speech inputs, hence, the quality of the decorrelated stereo speech signals are kept undisturbed. Test results on recorded speech signals indicate improved performance in both objective and subjective manners.