首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于两级判别法的端点检测算法研究
引用本文:李娟.基于两级判别法的端点检测算法研究[J].洛阳师范学院学报,2012,31(5):62-64.
作者姓名:李娟
作者单位:运城学院物理与电子工程系,山西运城,044000
摘    要:本文介绍了一种基于两级判别法的端点检测算法,该算法利用短时能量和短时平均过零率相结合的方法,继承了二者的优点,弥补了单独使用短时能量端点检测容易丢失清音、拖尾音的问题,通过短时平均过零率的第二级判别进一步扩大语音段的范围,在高信噪比和低信噪比情况下分别验证.实验证明:这种算法简单易行,大大提高了端点检测的准确率,并具有较好的抗噪性.

关 键 词:端点检测  短时能量  短时平均过零率  两级判别法  低信噪比

A modified Noise-Robust Method for Pitch Period Detection
LI Juan.A modified Noise-Robust Method for Pitch Period Detection[J].Journal of Luoyang Teachers College,2012,31(5):62-64.
Authors:LI Juan
Institution:LI Juan(Department of Physics and Electronic Engineering,Yuncheng Univernity,Yuncheng 044000,China)
Abstract:An endpoint detection algorithm based on two stage method is presented.The algorithm uses an integrated approach of short-time energy and short-time zero-crossing rate,inheriting their merits and made up for shortcomings of losing voiceless sound and tailing sound if only using short-time energy method.It expands the spectrum of the voice by the second stage measurement of short-time average zero-crossing rate method.Verification is done both in high and low signal to noise ratio.The experiment shows that the algorithm is simple and feasible,and enhances the accuracy rate in endpoint detection and has the better anti-noise performance.
Keywords:endpoint detection  short-time energy  short-time zero-crossing rate  two stage method  low sig-nal to noise ratio
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号