基于子带广义旁瓣相消器的麦克风阵列语音增强  

Microphone Array Speech Enhancement Based on Subband Generalized Sidelobe Canceller

作　　者：何成林[1] 杜利民[1] 马昕[1] 

机构地区：[1]中国科学院声学研究所语音交互研究中心

出　　处：《计算机应用研究》2006年第4期208-210,共3页Application Research of Computers

基　　金：国家"973"计划资助项目(G1998030505)

摘　　要：为了加快基于广义旁瓣相消器的麦克风阵列语音增强系统的收敛速度,将其自适应模块的输入信号分解到子带以进行处理,并将多通道维纳滤波器引入广义旁瓣相消器的非自适应支路,以更有效地抑制非相干噪声。实际测试结果表明,相对于基于全带广义旁瓣相消器的麦克风阵列语音增强系统,采用该子带广义旁瓣相消器结构的语音增强系统具有更快的收敛速度和更高的输出信噪比。In order to accelerate the convergence of the microphone array speech enhancement system based on Generalized Sidelobe Canceller （ GSC）, the input signals of the adaptive block are decomposed into＇ theirs subbands to perform the adaptive multi-channel noise cancelling in each subband. And the integration of multi-channel wiener filter into the fixed branch of GSC can suppress the uneorrelated noise efficiently. Experiments indicate that the microphone array speech enhancement system based on this subband GSC has higher converging speed and better output signal-to-noise ratio than its fullband counterpart.

关 键 词：语音增强 麦克风阵列 子带自适应滤渡 广义旁瓣相消器 

分 类 号：TN912.3[电子电信—通信与信息系统]