语音增强中谱减法的改进应用  

Improved Application of Spectral Subtraction in Speech Enhancement

作　　者：何强[1] 罗晓松[1] 何巧[1] 万众[1] 

机构地区：[1]中国电子科技集团公司第三研究所,北京100015

出　　处：《电声技术》2008年第7期57-60,共4页Audio Engineering

摘　　要：针对谱减法在语音增强中的应用,提出了几种改进措施,包括预加重、初始噪声谱的估计、能量VAD、保护帧等。非正式试听表明改进后算法的增强处理效果好于基本谱减法。Aiming to the application of spectral subtraction in speech enhancement, several improved measures such as pre-emphasize,estimate original noise spectrum, energy threshold VAD and protecting frame are proposed. Informal listening tests indicated that the proposed methods has better result in speech enhancement.

关 键 词：语音增强 谱减 VAD 分频带 保护帧 

分 类 号：TN912[电子电信—通信与信息系统]