基于小波降噪的改进型OSF话音激活检测算法

Modified voice activity detection based on wavelets using order statistics filter

提出了一种基于小波降噪的改进型顺序统计滤波器(OSF)语音端点检测算法。在介绍了浊音分离、小波降噪和基于顺序统计滤波器(OSF)端点检测算法的原理基础上,给出了改进型算法的流程图和实验的仿真图。从结果可以看出,算法对信噪比大于0dB的语音信号,检测性能较为稳健。

This paper presents a modified voice activity detection using order statistics filter with wavelets denoising.It describes the principles of voiced-sound separation,noise reduction using wavelets and voice activity detection based on OSF.It gives a flow-process diagram of the modified algorithm,and gets the result of simulation through several experiments.It shows that the performance of the modified algorithm is robust when the signal-to-noise ratio greater than 0 dB.