一种利用人耳听觉特性的双声道回波抵消改进算法

An improved method of acoustic echo cancellation using auditory properties

本文提出了一种利用人耳听觉特性的双声道回波抵消改进算法。在该系统中,利用人耳的听觉特性给输入信号加上一个小的扰动来改善输入信号的自相关特性和条件数,从而达到改善系统收敛特性的作用。本文讨论了在双声道回波抵消系统中遇到的输入信号强相关从而引起系统性能恶化的问题,提出了采用输入信号加扰来改善系统性能的算法。并且利用人耳听觉特性使这种扰动信号隐藏在语音信号中,使其对语音信号的干扰最小。从本文的仿真试验来看,该方法对双声道回波抵消系统的性能有一定的改善。

This paper propose an improved method of AEC （acoustic echo cancellation） using auditory properties. In this method, a small disturbance has been added to improve the condition number and autocorrelation of the input signals by use of auditory properties. By this mean, system convergence performance can be improved. This paper discusses the problem of system performance degradation which is incurred by strong correlation of input signals in AEC system, and proposes an improved algorithm by adding disturbance to input signals. Simultaneously,we hide this disturbance into speech signals by using auditory properties. So,this disturbance can make least interruption to speech signals. According to the simulation results, this method can make some improving to the performance of stereophonic acoustic echo cancellation.