一种高性能的小波框架静音检测算法  

A High-performance Detection Algorithm Wavelet Frame Mute

作　　者：刘勇 谢映海 

机构地区：[1]广州海格通信集团股份有限公司,广东广州510663

出　　处：《科技视界》2016年第22期85-86,共2页Science & Technology Vision

摘　　要：利用语音信号和噪声分帧信号在频域上的小波框架分解后能量分布特征的明显差异性,本文提供了一种高性能的静音检测算法,算法引入了一种分段阈值和判决平滑机制,获得了较好的静音检测效果。仿真结果表明,在噪声功率时变和低信噪比的情况下,检测算法仍具备高准确率,性能明显优于一些传统算法。The energy distribution of the speech signal and noise sub-frame wavelet frame in the freqoenev donmin decoomposition characteristics of obvious differences,this article provides a high-performance detection algorithm silence,the algoritbm introduces a fragmentation threshold and smoothing judgment mechanism to obtain better silence detection Simuiation results show that the situation has changed and under low signal to noise ratio at the time of the noise power detection algorithm still have a high degree of accuracy.the performance was better than some of the traditional algorithms.

关 键 词：最小能量小波框架 分段阈值 判决平滑 静音检测 

分 类 号：TN912.3[电子电信—通信与信息系统]