基于LMS算法的自适应语音除噪性能研究  

Performance study of adaptive speech noise canceling based on LMS algorithm

作　　者：朱冲[1] 梁小朋[1] 

机构地区：[1]桂林电子科技大学信息与通信学院,广西桂林541004

出　　处：《桂林电子科技大学学报》2008年第4期298-301,共4页Journal of Guilin University of Electronic Technology

基　　金：广西自然科学基金项目(桂科自0832246);广西青年科学基金项目(桂科青0832084)

摘　　要：语音除噪是自适应信号处理研究的重点,通过利用matlab分别实现了基于时域定步长LMS算法、时域变步长LMS算法、基于FFT技术的LMS频域快速算法(FLMS)的自适应语音除噪仿真,并成功应用到带噪语音信号除噪方面。经实验表明,时域变步长LMS算法比时域定步长LMS算法的改善性噪比高1dB左右,与FLMS算法除噪性能相当,但FLMS算法的速度明显快于传统时域算法,而具体时间比的大小与总的样本数目有相关联系。Speech noise canceling is an important research field of adaptive signal processing. This paper simulates the performance of adaptive speech noise canceling based on LMS algorithms using matlab. These LMS algorithms contain time domain fixed steplength LMS, time domain variable steplength LMS, and frequency domain LMS based on fast fourier transform（FFT） technology. Experimental results proved that the improved signal-to-noise ratio of time domain variable steplength LMS was the same as FLMS, and ldB highter than time domain fixed steplength LMS algorithm. The efficiency of FLMS is higher than traditional time domain LMS algorithms,and time ratio is related to total sample number.

关 键 词：时域LMS算法 变换域LMS算法 自适应语音除噪 

分 类 号：TP273.2[自动化与计算机技术—检测技术与自动化装置]