计算机工程与应用 ›› 2018, Vol. 54 ›› Issue (7): 44-48.DOI: 10.3778/j.issn.1002-8331.1611-0366

• 理论与研发 • 上一篇    下一篇

谐波重构先验信噪比估计算法

王  杰1,杨程程1,莫嘉永2,王敦泽1,王谢谢1   

  1. 1.广州大学 机械与电气工程学院,广州 510006
    2.广州市信息安全测评中心,广州 510635
  • 出版日期:2018-04-01 发布日期:2018-04-16

A priori SNR estimator based on harmonic regeneration

WANG Jie1, YANG Chengcheng1, MO Jiayong2, WANG Dunze1, WANG Xiexie1   

  1. 1.College of Mechanical and Electric Engineering, Guangzhou University, Guangzhou 510006, China
    2.Guangzhou Information Technology Security Evaluation Center, Guangzhou 510635, China
  • Online:2018-04-01 Published:2018-04-16

摘要: 先验信噪比单通道语音增强算法在信噪比较高时能有效地去除噪声,但在信噪比较低时语音高次谐波失真较为严重。针对此提出了一种基于谐波重构的先验信噪比估计算法,对增强后的信号加权求平方,进行功率谱的二次谱处理,以加强语音信号的周期性;再进行谐波重构,提升谐波分量。实验研究表明,该算法在低信噪比时能够有效地增强语音谐波分量,相对于先验信噪比估计的语音增强算法能够改善语音质量,减少语音失真。

关键词: 语音增强, 先验信噪比, 谐波重构, 二次谱处理

Abstract: A priori SNR estimator of single-channel speech enhancement can effectively remove the noise when the SNR is high, but the speech distortion is more serious when the SNR is low. In this paper, a priori SNR estimator using harmonic regeneration is proposed. In the algorithm, the power spectrum reprocessing(including weighting and square)can be used to process speech signals, which can enhance the periodicity of the speech signal; and then the harmonic regeneration is carried out, thus the harmonic component is enhanced. Experimental results show that the proposed algorithm can effectively enhance the harmonic component of the speech signal in low SNR, improve the performance of this algorithm and reduce speech distortion than a priori SNR estimator of speech enhancement.

Key words: speech enhancement, a priori SNR, harmonic regeneration, power spectrum reprocessing