基于高斯混合模型的语音带宽扩展算法的研究  

Speech wideband extension based on Gaussian mixture model

作　　者：张勇[1] 胡瑞敏[1] 

机构地区：[1]武汉大学国家多媒体软件工程技术研究中心,武汉430072

出　　处：《声学学报》2009年第5期471-480,共10页Acta Acustica

基　　金：国家自然科学基金重点资助项目(60832002)

摘　　要：为了降低高带谱失真,研究了带宽扩展算法中特征参数与高带谱包络的互信息和高带谱失真之间的函数关系,并在此基础上提出了一种扩展高斯混合模型带宽扩展算法。首先,算法选择与高带谱包络互信息大的参数构成特征矢量,并根据高斯混合模型计算特征矢量与高带谱包络的联合概率密度。其次,采用Expectation-Maximization(EM)算法估计高斯分量模型参数并计算后验概率。最后,通过后验概率估计高带谱包络。实验结果表明,与传统的高斯混合模型带宽扩展算法相比,本文算法可降低0.3 dB的高带平均谱失真,将谱失真大于10dB的语音帧减少了50%以上。To decrease the spectral distortion of highband envelope, the function of spectral distortion and mutual information which is between feature vector and highband envelope is studied, and an extended Gaussian Mixture Model （GMM） wideband extension algorithm is proposed based on the research. The feature parameters which have higher mutual information with highband envelope are selected to constitute feature vector, and the GMM is adopted to compute the joint probability density of the feature vector and highband envelope. Then the highband envelope is estimated via the posterior probabilities computed from the model parameters estimated by Expectation-Maximization （EM） Mgorithm. The experimental results show that the spectral distortion is inferior to the Mgorithm, such as the traditional algorithm based on GMM, by 0.3 dB and the number of frames with spectral distortion over 10 dB sharply reduced over 50%.

关 键 词：高斯混合模型 扩展算法 语音帧 带宽 联合概率密度 特征矢量 概率估计 谱失真 

分 类 号：O423[理学—声学]