基于麦克风阵列的语音增强系统设计  

DESIGNING SPEECH ENHANCEMENT BASED ON MICROPHONE ARRAY

作　　者：朱兴宇[1] 万洪杰[1] 

机构地区：[1]北京化工大学信息科学与技术学院,北京100029

出　　处：《计算机应用与软件》2013年第3期240-243,共4页Computer Applications and Software

摘　　要：针对目前在近场语音环境中语音增强系统噪声干扰大、难以清晰还原信号等局限性,设计并实现一种基于麦克风阵列的新型语音增强系统。系统硬件设计新颖,采用低功耗设计,最大支持12路语音信号同时采集,同时对噪声的处理采用自适应差分脉冲编码调制和卡尔曼滤波算法相结合分开处理的新型方式。通过Matlab仿真和封闭近场环境下的实际测试,结果表明该系统性能稳定可靠,可以提高语音清晰度,实现短距离的多路语音增强。Due to the limitation of intense noise interference and difficult to restore signals of those the speech enhancement system has in near field speech environment,we design and implement a microphone array-based speech enhancement system.The novel designed system hardware adopts low-power consideration,and supports up to twelve channels＇ speech acquisition simultaneously.Meanwhile,it uses a novel approach to deal with noises,which adopts a way integrating the adaptive difference pulse code modulation（ADPCM） with Kalman filtering algorithm but processing separately.MATLAB simulations and practical tests in closed near field environment are carried out.Results demonstrate that the system performance is stable and reliable,it can clearly augment speech quality,and achieve multi-channel speech enhancement in short distance.

关 键 词：语音增强 麦克风阵列 数字信号处理器 自适应差分脉冲编码调制 卡尔曼滤波 

分 类 号：TP368.1[自动化与计算机技术—计算机系统结构]