摘要
传统语音增强算法在去除噪声的同时也导致语音受损,为了减小这种负面影响,结合了语音信号的稀疏表示算法与语音增强算法和自适应的获得训练字典,提出了一种基于自适应稀疏表示的语音增强算法。仿真实验结果表明该方法即使在低信噪比的条件下也能有效去噪,且去噪后能很好的分辨出原始语音信号。
Traditional speech enhancement algorithm cannot avoid damaging speech information,to reduce this negative impact,combining the speech sparse representation,speech enhancement algorithm with adaptively gaining the training dictionary,a speech enhancement algorithm based on adaptive sparse representation is proposed. Experiment results of simulated data show that the proposed approach could remove noise from signal effectively,keeping a satisfied recognition ability.
引文
[1] 赵力.语音信号处理[M].第2版.北京:机械工业出版社,2010:285-286.
[2] Paliwal K,Wojcicki K,Shannon B.The Importance of Phase in Speech Enhancement[J].Speech Communcation,2011,53(4):465-494.
[3] Gemmeke J F,Virtanen T,Hurmalainen A.Exemplar-Based Sparse Representations for Noise Robust Automatic Speech Recognition[J].IEEE Transactions on Audio,Speech,and Language Processing,2011,19(7):2067-2080.
[4] Michael Elad.Sparse and Redundant Representations[M].NewYork:Springer Science_ Business Media,2010:234.
[5] Rix A W,Beerends J G,Hollier M P,et al.Perceptual Evaluation of Speech Quality(PESQ)—A New Method for Speech Quality Assessment of Telephone Networks and Codecs[C]//Acoustics,Speech,and Signal Processing(ICASSP),USA,2001:749-752.
[6] Lee K.Application of Non-Negative Spectrogram Decomposition with Sparsity Constraints to Single-Channel Speech Ehnhancement[J].Speech Communication,2014,58:69-80.
[7] Xu Y,Du J,Dai L,et al.An Experimental Study on Speech Enhancement Based on Deep Neural Networks[J].IEEE Signal Processing Letters,2014,21(1):98-102.
[8] Li Jinyu,Deng Li,Gong Yifan,et al.An Overview of Noise-Robust Automatic Speech Recognition[J].Audio,Speech,and Language Processing,IEEE/ACM Transactions on,2014,22(4):745-777.