计算机工程与应用 ›› 2009, Vol. 45 ›› Issue (36): 145-147.DOI: 10.3778/j.issn.1002-8331.2009.36.043

• 数据库、信号与信息处理 • 上一篇    下一篇

小波变换的语音去噪方法

李从清1,2,孙立新1,龙 东3,李洙梁1,戴士杰1   

  1. 1.河北工业大学 机械学院,天津 300130
    2.天津城市建设学院 能源与机械工程系,天津 300384
    3.中国人民解放军 总参陆航研究所八室,北京 101101
  • 收稿日期:2008-07-15 修回日期:2008-11-27 出版日期:2009-12-21 发布日期:2009-12-21
  • 通讯作者: 李从清

Method of speech denoising using wavelet transform

LI Cong-qing1,2,SUN Li-xin1,LONG Dong3,LI Zhu-liang1,DAI Shi-jie1   

  1. 1.School of Mechanical Engineering,Hebei University of Technology,Tianjin 300130,China
    2.Department of Energy Sources and Mechanical Engineering,Tianjin Urban Construction Institute,Tianjin 300384,China
    3.Army Aviation Research Institute of PLA,Beijing 101101,China
  • Received:2008-07-15 Revised:2008-11-27 Online:2009-12-21 Published:2009-12-21
  • Contact: LI Cong-qing

摘要: 提出一种改进的语音去噪处理方法:二次小波分解全局阈值法。该方法不同于传统阈值滤波方法,首先对语音信号的高频部分进行二次分解,然后应用阈值滤波的方法对信号进行去噪处理。该方法在MATLAB上进行了模拟实验,实验结果表明该方法提高了信噪比,去除了大部分噪声,相当完整地保留了有效信号能量,很好地解决噪声对语音信号干扰的问题。

关键词: 小波变换, 语音去噪, 阈值滤波

Abstract: An improved method of speech denoising processing is put forward:Second-decomposition wavelet global threshold.This method is different from the traditional threshold filtering.First,it decomposes high frequency twice,and then denoises the signal by threshold filtering.Simulated test is made on MATLAB.Test results show that the algorithm improves signal to noise ratio,and wipes off the most noise with full effective energy.The interference problem of noise signal is absolutely solved.

Key words: wavelet transform, speech denoising, threshold filtering

中图分类号: