听觉模型反演方法及其应用  

Auditory model inversion and its application

作　　者：赵鹤鸣[1] 王永琦[1] 陈雪勤[1] 

机构地区：[1]苏州大学电子信息学院,苏州215021

出　　处：《声学学报》2005年第6期530-535,共6页Acta Acustica

基　　金：国家自然科学基金(60172016)江苏省计算机信息处理重点实验室开放课题(KJS010021)资助项目

摘　　要：听觉模型已应用于语音信号处理的许多方面,并已取得了较好的结果。论文根据目前应用较为广泛的听觉模型,提出了模型中各部分的逆变换。首先通过对自相关图谱逆变换并经迭代运算重构相位信息恢复神经发放率,再经半波整流逆运算恢复负半部分信号,最后对描述内毛细胞、突触模型的相关方程进行逆运算和Gammatone逆滤波,构成整个听觉模型的反演方法。作为应用论文提出了一种基于听觉模型逆变换的含噪语音增强方法。实验结果表明,该方法对含噪语音有很好的降噪效果,特别是当信噪比较低时,该方法较常用的方法更为有效。论文提出的听觉模型逆变换方法可应用于语音增强等领域。Auditory model has been applied to several aspects of speech signal processing field, and appears to be effective in performance. This paper presents the inverse transform of each stage of one widely used auditory model. First of all it is necessary to invert correlogram and reconstruct phase information by repetitious iterations in order to get neural firing rate. The next step is to obtain the negative parts of the signal via the reverse process of the HWR （Half Wave Rectification） stage. Finally the functions of inner hair cell/synapse model and Gammatone filters have to be inverted. Thus the whole auditory model inversion has been got. An application of noisy speech enhancement based on auditory model inversion is mentioned in this paper. Many experiments show that this method is effective for noisy speech, especially for low SNR. The method of auditory model inversion given in this paper could be used widely in speech enhancement field.

关 键 词：听觉模型 反演方法 应用 语音信号处理 逆变换 信息恢复 迭代运算 半波整流 相关方程 

分 类 号：TN912.3[电子电信—通信与信息系统]