噪声环境下语音多子带端点检测方法  

Sub-bands Endpoints Detection of Noisy Speech

作　　者：林兴华[1] 肖熙[1] 

机构地区：[1]清华大学电子工程系,北京100084

出　　处：《电声技术》2011年第9期34-37,共4页Audio Engineering

摘　　要：为了克服传统端点检测算法在有噪环境下的不足,提出了一种新的语音多子带端点检测算法,它能根据子带背景噪声大小动态设置门限。实验结果表明,所提出的算法较传统的检测方法有更好的检测性能。In order to overcome the shortcoming of the traditional speech endpoints detection algorithm in noisy environment, a new multiple sub-bands speech endpoints detection algorithm is proposed in this paper, which can dynamically adjust the threshold according to sub-bands background noise. Experimental results show that the proposed method has good detection performance compared with the other baseline algorithms.

关 键 词：语音信号处理 端点检测 语谱图 子带 

分 类 号：TN912[电子电信—通信与信息系统]