摘 要：语音增强是解决语音噪声污染的一种有效手段。本文简单介绍了现有的基于语音短时对数谱的最小均方误差（MMSE-LSA）进行估计的语音增强算法。利用帧间平滑定义平滑系数来对先验信噪比进行连续估计，在减少语音失真机会的同时，能够有效地抑制残留噪声。文章还通过MATLAB 仿真分析其优劣，并结合算法的优点来达到语音增强的最佳效果，从而提高语音信号的可懂度和识别率。
关键词：语音增强；MMSE-LSA；帧间平滑；MATLAB 仿真
中图分类号：TP391.42；TN912.3 文献标识码：A 文章编号：2096-4706（2018）02-0118-02
Research and Implementation of MMSE-LSA Speech Enhancement algorithm
ZHAO Hongzhi，AN Pengbo，DU Lixia
（School of Electronics and Information Engineering，Lanzhou Jiaotong University，Lanzhou 730070，China）
Abstract：Speech enhancement is an effective means to solve speech noise pollution.This paper briefly introduces the existing speech enhancement algorithm based on the minimum mean-squared error （MMSE-LSA） estimation of speech short-time log spectrum，then improve the speech enhancement algorithm based on MMSE-LSA，to improve the effect of speech enhancement，and the enhanced speech is more in line with human hearing.Using inter-frame smoothing to define the smoothing coefficient continuously estimates the a priori SNR，which can effectively suppress the residual noise while reducing the opportunity for speech distortion.The advantages and disadvantages are analyzed through MATLAB simulation，and the advantages of the algorithm are combined to achieve the best effect of speech enhancement，thereby improving the intelligibility and recognition rate of the speech signal.
Keywords：speech enhancement；MMSE-LSA；inter-frame smoothing；MATLAB simulation
参考文献：
[1] 朱先和，杨世平. 基于小波变换对信号噪声的处理研究 [J].电子科技，2016，29（6）：128-131.
[2] Hu Y，Loizou P C. Speech enhancement based on wavelet thresholding the multitaper spectrum [J].IEEE Transactions on Speech & Audio Processing，2004，12（1）：59-67.
[3] 李轩，张瑜，关庆阳. 卫星信道MMSE 和LS 均衡算法性能比较 [J]. 电子科技，2015，28（7）：84-86.
[4] Hasan M K，Zilany MSA，Khan M R. DCT speech enhancement with hard and soft thresholding criteria[J]. Electronics Letters，2002，38（13）：669-670.
[5] 杨龙. 几种单通道的语音增强算法研究 [J]. 科技视界，2015（26）：155+222.
[6] Wang Y，Han K，Wang D L. Exploring Monaural Features for Classification-Based Speech Segregation [J].IEEE Transactions on Audio Speech & Language Processing，2013，21（2）：270-279.
[ 7 ] Xu T，Wa n g W. A b l o c k - b a s e d c o mp r e s s e d sensing method for underdetermined blind speech separation i n c o r p o r a t i n g b i n a r y m a s k [ C ] / / I E E E I n t e r n a t i o n a l Conference on Acoustics Speech and Signal Processing，2010，130（5）：2022-2025.
作者简介：赵宏志（1990-），男，汉族，江苏涟水人，研究生，研究方向：数字信号处理，无线通信。