首页 | 官方网站   微博 | 高级检索  
     

跟踪共振峰的小波包语音增强仿真研究
引用本文:田玉静,周立俭,左红伟.跟踪共振峰的小波包语音增强仿真研究[J].烟台大学学报(自然科学与工程版),2011(4).
作者姓名:田玉静  周立俭  左红伟
作者单位:青岛理工大学现代教育技术中心;青岛理工大学信息工程学院;青岛理工大学土木工程学院;
基金项目:山东省高等学校科技计划资助项目(J09LG03)
摘    要:通过仿真实验,研究和分析了语音共振峰的子波时频特性,发现在各子带可跟踪并找到共振峰,结合听觉感知,提出了一种跟踪共振峰的小波包自适应阈值语音降噪新算法.该算法能拟合人耳听觉特性,紧密跟踪语音信号的能量变化,通过系数实时调节子带阈值,消噪同时能有效减少语音弱分量损失.仿真实验表明,相对较传统算法,新算法在低信噪比输入时,能有效提高增强语音的输出信噪比、保持较小的波形失真度、获得较好的听觉效果.

关 键 词:共振峰  听觉感知  自适应阈值  小波包分析  

Simulation of Wavelet Package Speech Enhancement Based on Tracing Formant
TIAN Yu-jing,ZHOU Li-jian,ZUO Hong-wei.Simulation of Wavelet Package Speech Enhancement Based on Tracing Formant[J].Journal of Yantai University(Natural Science and Engineering edirion),2011(4).
Authors:TIAN Yu-jing  ZHOU Li-jian  ZUO Hong-wei
Affiliation:TIAN Yu-jing1,ZHOU Li-jian2,ZUO Hong-wei3(1.Modern Education and Technology Center,Qingdao Technological University,Qingdao 266033,China,2.School of Communication and Electronics Engineering,3.School of Civil Engineering,Qingdao Technology University,China)
Abstract:We study and analyze the time-frequency characteristics of wavelet sub-band speech formant in the way of computer simulation experiments.It is obtained that every speech formant can be traced and found out in the sub-band.Combined with the human auditory perception,this paper presents a new algorithm of wavelet package threshold speech enhancement based on tracing the formant.The new algorithm can fit human auditory,closely trace the energy change of speech signal,and reduce the damage of weak speech signal...
Keywords:formant  auditory perception  adaptive threshold  wavelet packet analysis  
本文献已被 CNKI 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司    京ICP备09084417号-23

京公网安备 11010802026262号