高噪声环境下基于自适应滤波语音降噪技术的研究  

Research of noise reduction technology for voice signal based adaptive filter in high noise environment

作　　者：郑海啸[1] 刘珩[1] 

机构地区：[1]中国农业大学信息与电气工程学院,北京100083

出　　处：《电子测量技术》2007年第7期16-19,23,共5页Electronic Measurement Technology

摘　　要：为了在高噪声环境下得到较清晰的语音信息,本文提出一种基于线性预测的FIR自适应语音滤波器,对LMS算法作出改进,提高了算法的收敛性。在多种噪声环境下,MATLAB仿真实验结果表明,改进的算法具有更好的降噪效果,显著提高了信噪比,并保持了语音的自然度。To get the clear voice signal under the situation of high-noise, this paper introduces a type of adaptive speech FIR filter based on linear prediction. The algorithm improves LMS algorithm, and has better convergence. After MATLAB simulating under several different noise, the result shows that the improved LMS algorithm reduces more noise than the classical LMS algorithm, evidently improves SNR, and keeps the naturalness of the speech.

关 键 词：自适应滤波 线性预测 语音信号 降噪 

分 类 号：TN912.35[电子电信—通信与信息系统]