【作者】 沈亚强； 程仲文； 任德官；

【Author】 Shen Yaqiangi Chen Zhongwen Ren deguan (Zhejiang Normal University)

【机构】 浙江师大微电子技术应用研究所；

【摘要】 本文介绍了低信噪比带噪话音信号中语音、非语音段辨别的几种方法,并指出用短时自相关系数和时间序列的分数维方法在区别被附加噪声污染的语音信号中的语音段、非语音段是有效的,而且还具有算法简单高速的优点．更多还原

【Abstract】 In this paper , several methods that disciminate whether speech or non- speech on the low SNR noisy speech are proposed , and we indicate that the two algorithems of short -time autocorrelation coeffient and fraotal characterization of time sequences are very efficient to discriminate whether speechor non-speech in the speech degraded by stationary additional noise, and the algorithms are not only simple but also of high speed.更多还原