基于特征值合并的语音增强算法

系统仿真学报 ›› 2016, Vol. 28 ›› Issue (7): 1622-1627.

基于特征值合并的语音增强算法

陆慧娟¹, 刘亚卿¹, 刘砚秋¹, 关伟²

1.中国计量学院信息工程学院,浙江杭州310018;
2.中国计量学院现代科技学院,浙江杭州 310018

收稿日期:2014-08-12 修回日期:2015-02-01 出版日期:2016-07-08 发布日期:2020-06-04
作者简介:陆慧娟(1962-),女,浙江,博士,教授,研究方向为物联网、模式识别与云计算。
基金资助:
国家自然科学基金(61272315,60842009),浙江省自然科学基金(Y1110342)

Eigen-value Combination Approach for Speech Enhancement

Lu Huijuan¹, Liu Yaqing¹, Liu Yanqiu¹, Guan Wei²

1. College of Information Engineering, China Jiliang University, Hangzhou 310018, China;
2. College of Moder Science and Technology, China Jiliang University, Hangzhou 310018, China

Received:2014-08-12 Revised:2015-02-01 Online:2016-07-08 Published:2020-06-04

摘要/Abstract

摘要： 为进一步抑制噪声,提出一种基于特征值合并的语音增强算法。在经典的内嵌式预白化子空间方法的基础上,用特征值合并来提高语音质量。研究发现,对含噪语音的协方差矩阵进行特征值分解后,大特征值分量主要包含语音信息,而小特征值分量主要包含噪声,特征值分量按特征值从小到大排序后,剔除相邻的小特征值分量,可有效抑制噪声,提高语音质量。相比于其它方法,基于特征值合并的语音增强算法能有效工作于各种噪声环境中,显著提高信噪比,并有更好的语音可懂度。

关键词: 语音增强, 子空间方法, 特征值分解, 语音质量

Abstract: In order to further suppress noise, a kind of speech enhancement algorithm based on Eigen-value merge was proposed. Eigen-values combination was used to improve the speech quality on the basis of the classic embedded pre-whitening subspace methods. The study shows that, after decomposing the covariance matrix of speech signals with noise, the larger Eigen-value component mainly includes speech information, and the smaller Eigen-value component mainly contains noise. Sorted by Eigen-values from small to big, the adjacent large Eigen-value component replaces with small Eigen-value component, which can effectively suppress noise and improve the quality of speech. Compared with other speech enhancement algorithm, this algorithm based on Eigen value merger can work effectively in a variety of noisy environment, significantly improves the SNR, and has better speech intelligibility.

Key words: speech enhancement, subspace method, eigen-value decomposing, speech quality

中图分类号:

TN912.3

陆慧娟, 刘亚卿, 刘砚秋, 关伟. 基于特征值合并的语音增强算法[J]. 系统仿真学报, 2016, 28(7): 1622-1627.

Lu Huijuan, Liu Yaqing, Liu Yanqiu, Guan Wei. Eigen-value Combination Approach for Speech Enhancement[J]. Journal of System Simulation, 2016, 28(7): 1622-1627.

参考文献 16

[1]	Boll S.Suppression of Acoustic Noise in Speech Using Spectral Subtraction[J]. IEEE Transactions on Acoustics, Speech and Signal Processing (S0096-3518), 1979, 27(2): 113-120.
[2]	Ephraim Y, Malah D.Speech Enhancement Using a Minimu m Mean Square Error Short-time Spectral Amplitude Estimator[J]. IEEE Transactions on Acoustics, Speech and Signal Processing (S0096-3518), 1987, 32(6): 1109-1121.
[3]	Wiener N. Extrapolation, Interpolation,Smoothing of Stationary Time Series [M]. USA: The MIT Press, 1964.
[4]	Plapous C, Marro C, Scalart P.Improved Signal-to-Noise Ratio Estimation for Speech Enhancement[J]. IEEE Transactions on Audio, Speech and Language Processing,(S1558-7916), 2006, 14(6): 2098-2108.
[5]	Ephraim Y, Van-Trees H L. A Signal Subspace Approaches for Speech Enhancement[J]. IEEE Transactions on Speech and Audio Processing (S1063-6676), 1995, 3(4): 251-266.
[6]	Yi H, Loizou P C.A Generalized Subspace Approach for Enhancing Speech Corrupted by Colored Noise[J]. IEEE Transactions on Speech and Audio Processing,(S1063-6676), 2003, 11(4): 334-341.
[7]	Borowicz A, Petrovsky A.Signal Subspace Approach for Psychoacoustically Motivated Speech Enhancement[J]. Speech Communication (S0167-6393), 2011, 53(2): 210-219.
[8]	Mittal U, Phamdo N.Signal/noise KLT Based Approach for Enhancing Speech Degraded by Colored Noise[J]. IEEE Transactions on Speech and Audio Processing (S1063-6676), 2000, 8(2): 159-167.
[9]	Dendrinos M, Bakamidis S, Garayannis G.Speech Enhance-ment from Noise: A Regenerative Approach[J]. Speech Communication (S0167-6393), 1991, 10(2): 45-57.
[10]	Jensen S H, Hansenn P C, Hansen S D, et al.Reduction of Broad-band Noise in Speech by Truncated QSVD[J]. IEEE Transactions on Speech and Audio Processing,(S1063-6676), 1995, 3(6): 439-448.
[11]	Rezayee A, Gsxot S.An Adaptive KLT Approach for Speech Enhancement[J]. IEEE Transactions on Speech and Audio Processing (S1063-6676), 2001, 9(2): 87-95.
[12]	Li Z B, Zhao X X.A Kind of the Inverse Eigenvalue problem for Tridiagonal Matrices[C]// Advancing knowledge discovery and data mining technologies, Proceedings of International Conference on Knowledge Discovery and Data Mining. Moscow, Russia: KDDM, 2009: 68-71.
[13]	Loizou P C.Speech Enhancement: Theory and Practice [M]. Boca Raton, FL, USA: CRC Press Inc, 2013.
[14]	Stapleton O, Melvin H, Pocta P.Quantifying the effectiveness of PESQ(Perceptual Evaluation of Speech Quality),in coping with frequent time shifting[C]// 10th International Conference and On-Line Workshop on Measurement of Speech, Audio and Video Quality in Networks (MESAQ-IN). Czech Republic: Czech Technical Univ Prague, 2011: 10-13.
[15]	Yi H, Loizou P C.Evaluation of Objective Quality Measures for Speech Enhancement[J]. IEEE Transactions on Speech and Audio Processing (S1063-6676), 2008, 16(1): 229-238.
[16]	Boh L S, Yit C T, Chang J S, et al.A Parametric Formulation of the Generalized Spectral Subtraction Method[J]. IEEE Transactions on Speech and Audio Processing (S1063-6676) 1998, 6(4): 328-337.