自适应门限处理双话的双滤波回声抵消算法

A dual-filters echo canceller using adaptive threshold to handle double talk

针对当前双话处理算法对未知或可变回声路径性能恶化以及复杂度过高等问题,采用一种基于短时能量比较,根据已估计的回声路径信息自适应更新判决门限的双话检测算法,为确保更新在滤波器收敛到一定程度后进行,引入了表征滤波器调整程度的参数和残留回声与参考信号的能量比来控制更新速率,同时达到检测路径变化的作用。文中算法结合双滤波器,能处理误检带来的收敛速率下降的问题。实验证明,该回声抵消器既能有效消除回声又能保护近端语音,同时能够处理路径变化。

As existed algorithms of double talk detection are vulnerable to variable or unknown echo path or complexity is too high for real time implementation. This paper takes use of short-term average energy and dynamically adapts the detection threshold according to the estimated echo path. The short-term average energy ratio and a parameter which can be used to detect path change are introduced to control the update of threshold, ensuring that DT detection are adopted only when the filter has converged. This algorithm combined with the dual filter can handle the problem of decline of convergence rate caused by false alarms. The experiments show that the echo canceller can effectively eliminate echo, protect the near-end voice, and work when echo path changes