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

一种改进的小波域语音去噪方法研究
引用本文:刘娟花,李福德.一种改进的小波域语音去噪方法研究[J].西安工程科技学院学报,2006,20(1):92-95.
作者姓名:刘娟花  李福德
作者单位:[1]西安工程科技学院电子信息学院,陕西西安710048 [2]西安理工大学自动化与信息工程学院,陕西西安710048
摘    要:分析了信号和噪声在小波域的不同特征表现,并根据语音中浊音和清音的特点,提出了一种改进的多尺度多阈值的小波域语音去噪方法.该方法采用软限幅函数对浊音和清音信号的小波变换系数作不同的阈值处理,既抑制了噪声,又减少了语音段信息的损失,提高了信噪比.仿真结果表明,这是一种有效的语音增强方法.

关 键 词:小波变换  语音去噪  随机噪声  阈值
文章编号:1671-850X(2006)01-0092-04
修稿时间:2004年12月15

Study on a modified method of wavelet domain speech denoising
LIU Juan-hua,LI Fu-de.Study on a modified method of wavelet domain speech denoising[J].Journal of Xi an University of Engineering Science and Technology,2006,20(1):92-95.
Authors:LIU Juan-hua  LI Fu-de
Abstract:According to the different characteristics between signal and noise on wavelet transform domain,also considering the voiced and unvoiced speech has different features,a modified method of speech denoising which is using a changing threshold at different scales is proposed.That is,using a soft thresholding to remove noise components from the wavelet coefficients of the voiced and unvoiced speech in noisy speech respectively in a different way,which is not only removing noise but also is preventing the quality degradation of the unvoiced sounds and enhancing the signal-noise ratio.Simulation results demonstrate the method is very effective in speech enhancement.
Keywords:wavelet transform  speech denoising  random noise  threshold
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