含噪语音信号端点检测  

Detection of the end points of noisy speech signal

作　　者：罗海涛[1] Luo Haitao(School of Informatics, Guangdong University of Foreign Studies, Guangzhou, Guangdong 510420, China)

机构地区：[1]广东外语外贸大学信息学院,广东广州510420

出　　处：《计算机时代》2017年第10期13-15,共3页Computer Era

摘　　要：MATLAB是一个功能强大的数学软件,可以进行数值运算、函数或信号处理、绘制图形。处理含噪语音信号,通常需要首先消除噪声,与语音信号相比,交通噪声频率较低。根据Daubechies方法,构造小波,并据此导出小波包。再用小波包对含噪语音信号进行分解。该算法检测语音端点,获得良好效果。MATLAB is powerful mathematical software that performs numerical calculations, function or signal processing, and draws graphs. Processing noisy speech signal usually needs to eliminate noise first. Compared with speech signal, traffic noise frequency is lower. In this paper, Daubechies wavelet is constructed, and a wavelet packet is derived from the constructed scaling and wavelet functions. Then the wavelet packet is used to decompose the noisy speech signal. Detecting the starting point and ending point of speech with the algorithm achieves a good result.

关 键 词：MATLAB 语音信号处理 小波包 端点检测 

分 类 号：TP311.11[自动化与计算机技术—计算机软件与理论]