期刊文献+

基于级联式改进型谱减算法的语音去噪技术研究 被引量:2

Speech Denoising Based on Cascaded-Improved Spectrum Subtraction Method
下载PDF
导出
摘要 在研究基本谱相减法的基础之上,提出了一种级联式改进型谱减算法。通过适当调整三个基本谱减系数值的大小和多次级联谱相减的处理方式,该方法有效的去处了含噪语音中的环境噪声和由谱减法引起的"音乐噪声",从而取得了良好的语音去噪效果,同时也为语音增强技术的研究提供了重要参考。 This paper puts forward spectrum subtraction algorithm a cascaded-improved spectrum subtraction algorithm based on the research of the basic By properly adjusting the values of the three basic spectrum subtraction coefficient and times process modes of cascaded spectrum subtraction,the method effectively weakens the environmental noise in noisy speech and the musical noise caused by spectrum subtraction algorithm.This paper obtains a good effect of speech denoising,and at the same time it also provides an important reference to the research of speech enhancement technology.
出处 《工业控制计算机》 2014年第7期119-120,123,共3页 Industrial Control Computer
关键词 语音增强 级联谱减法 噪声消除 音乐噪声 speech enhancement,cascaded spectrum subtraction algorithm,noise elimination,musical noise
  • 相关文献

参考文献3

二级参考文献7

  • 1Boll S.Suppression of Acoustic Noise in Speech using Spectral Subtraction[J].IEEE Trans on Acoustic,Speech and Signal Processing,1979;27(2):113-120.
  • 2Kamath sunil,Loizou Philipos.A multi-band spectral method for enhancing speech corrupted by colored noise[C].In:IEEE International Conference on Acoustic,Speech and Signal Processing,2002; (4):4164.
  • 3Virag N.Single channel speech enhancement based on masking properties of human auditory system[J].IEEE Trans on Speech Audio process,1999; 7 (2):126-137.
  • 4Ghoreish M H,Sheikzadeh H.Hybird Speech Enhancement System Based on HMM and Spectral Subtraction[C].In:IEEE International Conference on Acoustic,Speech and Signal Processing,2000; (3):1855-1858.
  • 5高鹰,谢胜利.一种变步长LMS自适应滤波算法及分析[J].电子学报,2001,29(8):1094-1097. 被引量:400
  • 6陈紫强,黄冰,刘庆华.高性能语音增强算法的研究[J].电声技术,2002,26(10):4-7. 被引量:2
  • 7杨秋成,蔡斌,陈娅冰.应用谱减法及其改进型算法进行语音增强[J].无线电工程,2003,33(8):43-45. 被引量:5

共引文献20

同被引文献18

引证文献2

二级引证文献1

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部