基于软硬阈值折中的小波包语音增强算法的研究  

Speech enhancement algorithm based on compromise for soft and hard thresholds

作　　者：王芳[1] 刘祖润[1] 吴海辉[1] 

机构地区：[1]湖南科技大学信息与电气工程学院,湘潭411201

出　　处：《铁路计算机应用》2010年第7期8-10,共3页Railway Computer Application

基　　金：国家自然科学基金项目(60673119);湖南省科技计划项目(2006GK3071)

摘　　要：由于语音信号的非平稳性,传统去噪方法将会不可避免地造成有用语音信号的损失,小波包分析能同时对信号的低频部分和高频部分进行分解,与小波分析相比,对信号的分析能力更强。文中提出一种新的去噪方法,它对噪声的清除更加干净,仿真结果表明,这种方法优于软、硬阈值法。Due to the non-stationarity of speech signals,traditional method of de-noising would inevitably cause the loss of the useful speech signal.By using wavelet packet the speech signal could be decomposed into the low frequency part and the high frequency part,compared with wavelet analysis,the analytical capacity of the signal was stronger.A new de-noising method was proposed,which could de-noise the speech signal more completely.Simulation results showed that this method was superior to the soft and hard threshold method.

关 键 词：小波包变换 语音消噪 语音增强 阈值函数 

分 类 号：TN911.72[电子电信—通信与信息系统]