基于盲信号分离的自适应回声抵消算法

Adaptive Echo Cancellation based on Blind Signal Separation

在视频会议和免提通信系统中,扬声器和麦克风之间的声耦合严重影响语音通信系统的质量。文中提出了一种用基于盲信号分离(BSS)的自适应声回波对消(AEC)方法,可有效解决回声和噪声对近端语音信号的影响。该方法不仅能减少背景噪声对回波对消的影响,而在双向通话时,可直接利用盲信号分离技术分离出近端语音。

The echo and noise arising in applications such as teleconferencing and free-hand communication seriously affect the quality of speech communication system. The contributions of the loud-speaker signals and the background noise to the microphone signals need to be cancelled. In this paper, a method for echo cancellation based on blind signal separation is proposed, which can effectively reduce the influence of echo and noise on the speech signal. During double talk, the near-end speaking signals are accurately recovered from the mixture signals combining noise and echo directly by BSS.