一种基于软判决的立体声回声抵消算法

Stereophonic Acoustic Echo Suppression based on Soft Decision

提出了一种新的基于维纳滤波的频域算法来解决立体声回声抵消问题,该算法不需要对立体声信号预处理,从而最大程度地保证了近端语音质量。并且它具有很好的鲁棒性,很快的收敛速度和跟踪速度,因而具有一定的实用价值。引入了语音增强中的软判决方法来进一步提高算法的性能。新的算法在保证近端语音质量的同时达到了更好的回声压制效果。仿真实验证明了新算法的良好性能。

Recently, a stereophonic acoustic echo suppression （SAES） algorithm based on Wiener filter in the frequency domain is proposed. This method does not require the signal preprocessing, and thus obtains good speech quality. Also, the algorithm is robust and has fast convergence and tracking speed. In this paper, a soft decision approach to further improve the echo cancellation performance is introduced. The new scheme has a good tradeoff between the speech quality and echo - return -loss enhancement （ERLE） performance. Computer simulations demonstrate the effectiveness of the new algorithm.