期刊文献+

基于扩展谱减法语音增强系统的DSP实现 被引量:1

Implementing a Speech Enhancement System with ESS on DSP
下载PDF
导出
摘要 针对语音通信经常受到背景噪声干扰的问题,本文采用扩展谱减法(ExtendedSpectral Subtraction)在TI公司的TMS320VC5410DSP芯片上实现语音增强系统。实验结果表明,当p取0.85时,获得最佳增强效果,增强后的语音可懂度明显得到改善。由于背景噪声是非平稳的,在低信噪比下语音信号有所损伤,但是没有出现大量的“音乐噪声”,其可懂度也没有显著下降。扩展谱减法可以有效地消除背景噪声,抑制“音乐噪声”,信噪比越高,增强效果越好。即使当信噪比较低(0 dB)时也能很好的提高语音质量。 In speech communication environments often occurs the situation that the speech signal is superosed by background noise. To overcome it we implement a speech enhancement system with Extended Spectral Subtraction (ESS) algorithm on TMS320VC5410 DSP. The results of experiments show that this method can suppress the noisy speech very well even at 0dB SNR with a low level of musical noise. The system gets the best performance when p is 0.85. Because the backround noise is nonstationary, the enhanced speech gets worse when SNR gets lower.
出处 《青岛大学学报(工程技术版)》 CAS 2005年第4期61-64,共4页 Journal of Qingdao University(Engineering & Technology Edition)
关键词 语音增强 谱减 扩展谱减 DSP speech enhancement spectral subtraction ESS DSP
  • 相关文献

参考文献2

  • 1Philip,Loizou.A Modified Spectral Subtraction Method Combined with Perceptual Weighting for Speech Enhancement[M].Dallas:The University of Texas,2002:23-50.
  • 2Sovka P,Pollakp P,Kybic J.Extended Spectral Subtraction[C]//Proceedings of EUSIPCO'96,Vol.2.Italy:EURASIP,1996:963-966.

同被引文献4

  • 1钱国青,赵鹤鸣.基于改进谱减算法的语音增强新方法[J].计算机工程与应用,2005,41(35):42-43. 被引量:18
  • 2Martin R. Noise power spectral density estimation based on optimal smoothing and minimum statistics[J].IEEE Transactions on Speech and Audio Processing,2001,(05):504-512.
  • 3Jax P,Vary P. Artificial bandwidth extension of speech signals using MMSE estimation based on a hidden Markov model[J].IEEE International Conference on Aeeousties Speech and Sigllal Proeessing,2003,(01):680-683.
  • 4李媛,铁勇,那顺乌力吉,李树华,王强.基于扩展谱减法的语音增强算法及实现[J].内蒙古大学学报(自然科学版),2008,39(1):97-101. 被引量:2

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