计算机工程与应用 ›› 2008, Vol. 44 ›› Issue (32): 6-9.DOI: 10.3778/j.issn.1002-8331.2008.32.002

• 博士论坛 • 上一篇    下一篇

基于短时谱估计的语音增强研究

雷建军1,杨 震2,刘 刚3,郭 军3   

  1. 1.天津大学 电子信息工程学院,天津 300072
    2.北京工业大学 计算机学院,北京 100124
    3.北京邮电大学 信息工程学院,北京 100876
  • 收稿日期:2008-07-21 修回日期:2008-08-18 出版日期:2008-11-11 发布日期:2008-11-11
  • 通讯作者: 雷建军

Research on speech enhancement based on short-time spectrum estimation

LEI Jian-jun1,YANG Zhen2,LIU Gang3,GUO Jun3   

  1. 1.School of Electronic Information Engineering,Tianjin University,Tianjin 300072,China
    2.College of Computer Science,Beijing University of Technology,Beijing 100124,China
    3.School of Information Engineering,Beijing University of Posts and Telecommunications,Beijing 100876,China
  • Received:2008-07-21 Revised:2008-08-18 Online:2008-11-11 Published:2008-11-11
  • Contact: LEI Jian-jun

摘要: 基于短时谱估计的语音增强算法具有良好的降噪性能,算法高效且易于实现。本文对谱减法、维纳滤波、最小均方误差估计等此类算法进行系统的论证,结合实验,分析比较了它们的性能差异,并指出了它们各自的优缺点及适用环境。

关键词: 语音增强, 短时谱, 谱减, 维纳滤波, 最小均方误差

Abstract: Speech enhancement based on short-time spectrum estimation has good noise robustness,and it is efficient and easy-to-implement.In this paper,spectral subtraction,Wiener filter,and minimum mean square error estimator are demonstrated.Then,their respective advantages and disadvantages of the approaches are outlined.Finally,the applicable field and localization of each method are also pointed out.

Key words: speech enhancement, short-time spectrum, spectral subtraction, Wiener filter, Minimum Mean Square Error