计算机工程与应用 ›› 2008, Vol. 44 ›› Issue (5): 215-217.

• 工程与应用 • 上一篇    下一篇

一种Bark子波变换的电子耳蜗语音增强算法

黄雅婷,顾济华,陶 智,赵鹤鸣,吴 迪   

  1. 1.苏州大学 物理科学与技术学院,江苏 苏州 215006
    2.苏州大学 电子信息学院,江苏 苏州 215006
  • 收稿日期:2007-06-07 修回日期:2007-08-24 出版日期:2008-02-11 发布日期:2008-02-11
  • 通讯作者: 黄雅婷

Speech enhancement in cochlear implant based on bark wavelet transform

HUANG Ya-ting,GU Ji-hua,TAO Zhi,ZHAO He-ming,WU Di   

  1. 1.School of Phys. Sci. & Tech.,Suzhou Univ.,Suzhou,Jiangsu 215006,China 2.Dept. of Electron,School of Infromation,Suzhou Univ.,Suzhou,Jiangsu 215006,China
  • Received:2007-06-07 Revised:2007-08-24 Online:2008-02-11 Published:2008-02-11
  • Contact: HUANG Ya-ting

摘要: 提出了一种Bark子波变换的电子耳蜗语音增强算法。该算法首先引入与人耳听觉系统更为适应的Bark子波变换来进行电子耳蜗CIS语音信号处理,然后在每个Bark通道中利用非线性谱减法对其进行语音增强,谱减法的参数由人耳隐蔽阈值来控制。结果表明:即使在低信噪比的情况下,信噪比也能提高16 dB左右,合成的语音对于电子耳蜗使用者具有较好的清晰度和可懂度。

关键词: Bark子波, 电子耳蜗, 语音增强

Abstract: A speech enhancement in cochlear implant based on Bark wavelet transform is proposed.Bark wavelet transform,which is more suitable to human ear’s auditory system,is first used in this new algorithm to realize speech in electronic cochlear implant.Then non-linear spectral subtraction algorithm is used to enhance speech in each channel,and the parameter of spectral subtraction is controlled by the masking threshold of human auditory.The result shows that the SNR output can be improved by 16 dB even if the SNR input is low and the enhanced speech is more clearness and intelligibility.

Key words: Bark wavelet, electronic cochlear implant, speech enhancement