作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2011, Vol. 37 ›› Issue (3): 178-179,182. doi: 10.3969/j.issn.1000-3428.2011.03.063

• 人工智能及识别技术 • 上一篇    下一篇

非因果先验信噪比估计的LSA算法改进

陈国冻a,何良华a,b   

  1. (同济大学a. 电子与信息工程学院;b. 国家高性能计算中心同济分中心,上海 201804)
  • 出版日期:2011-02-05 发布日期:2011-01-28
  • 作者简介:陈国冻(1986-),男,硕士研究生,主研方向:语音识别,语音增强;何良华,副教授、博士后
  • 基金资助:
    上海市国际科技合作基金资助项目(062107037, 07510 7005)

Improvement of LSA Algorithm on Noncausal A Priori SNR Estimation

CHEN Guo-dong a, HE Liang-hua a,b   

  1. (a. College of Electronic and Information Engineering; b. National High Performance Computing Center, Tongji Branch, Tongji University, Shanghai 201804, China)
  • Online:2011-02-05 Published:2011-01-28

摘要: 对于大多数的语音增强算法,先验信噪比及背景噪音频谱估计的准确与否,对语音增强的效果影响至关重要。为此,在传统MMSE-LSA算法的基础上,提出一种基于非因果先验信噪比估计的LSA改进算法,较好地弥补了传统LSA算法在先验信噪比上估计的不足,同时采用平滑系数动态更新噪音频谱值,使估计值能更好地跟踪噪音的变化。实验结果表明,改进算法能有效减少残余噪音量,提高语音分段信噪比,改善语音质量。

关键词: 非因果先验信噪比估计, 背景噪音估计, MMSE-LSA算法

Abstract: For most speech enhancement process, the accuracy of the a priori SNR and the background noise estimation has great influence. This paper derives traditional MMSE-LSA algorithm based on noncausal a priori SNR estimation. In contrast to the LSA algorithm, it introduces noncausal estimation for the a priori SNR, and utilizes smoothing parameters to change the value of noise spectral which can adaptively adjust to the noise environment. Experimental results show the new algorithm yields a higher improvement in the segmental SNR, lower residual noise and better enhancement of speech quality.

Key words: noncausal a priori SNR estimation, background noise estimation, MMSE-LSA algorithm

中图分类号: