一种改进的语音去噪方法(英文)  

An improved speech denoising method

作　　者：王彪[1] 王钟斐[1] 李建文[2] 

机构地区：[1]宝鸡文理学院数学系,陕西宝鸡721013 [2]陕西科技大学电气与信息工程学院,陕西西安710021

出　　处：《安徽大学学报（自然科学版）》2013年第6期23-29,共7页Journal of Anhui University(Natural Science Edition)

基　　金：Supported by the National Natural Science Foundation of China(60672001);the Education Department of Shanxi Province Fund Project(05JC03)

摘　　要：为了提高语音信号去噪效果,首先改进了小波包算法,提出了一种基于改进小波包的语音去噪方法.该方法将语音信号进行改进小波包分解,为每个终端结点提供一个阈值进行去噪处理.仿真实验表明,该方法比传统硬、软阈值方法更有效也更优越,能够比较准确地去除语音噪声.In order to improve the effect of speech signal denoising, the wavelet packet algorithm was improved, which a speech denoising method based on was proposed. In this method, the speech signal was decomposed by improved wavelet packet, a threshold was provided for each terminal node to denoise. Simulation experiment showed that this method was more effective and more superior than traditional hard, soft thresholding methods, which could remove noise of speech signal accuratly.

关 键 词：小波包 去噪 硬阈值 软阈值 语音信号 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]