改进的后滤波波束形成器语音增强算法

被引:4
作者
阎兆立
杜利民
机构
[1] 中国科学院声学研究所
关键词
语音增强; 传声器阵列; 波束形成器;
D O I
暂无
中图分类号
TN912.35 [语音增强];
学科分类号
0711 ;
摘要
该文提出了一种具有后滤波的波束形成器的语音增强改进算法。该算法主要解决维纳滤波器的理想信号功率谱估计,结合自功率谱减法和互功率谱减法计算出尽可能多的功率谱估计值,以使平均结果更接近于真实值,同时修正了声源移动引起的互功率谱变化。实验结果信噪比提高5dB以上,汽车环境中基于隐含马尔可夫模型(HMM)的小词汇量短语识别达到84%。从信噪比、平均谱距离和语音识别率可以看出该算法有效去除了原始算法中易残留的低频噪声,减少了语音信号失真。
引用
收藏
页码:2269 / 2272
页数:4
相关论文
共 10 条
[1]  
Speech enhancement based on the general transfer function GSC and postfiltering. Gannot S,,Cohen I. IEEE Trans.on Speech and Audio Processing . 2004
[2]  
A microphone array with adaptive post–filtering for noise reduction in reverberant rooms. Zelinski R. IEEE Int.Conf.Acoustic Speech Signal Process . 1988
[3]  
Microphone array post-filter for diffuse noise field. McCowan I A,,Bourland H. IEEE Int.Conf.Acoustic Speech Signal Process . 2002
[4]  
Speech enhancement based on a cobined multi-channel array with constrained iterative and auditory masked processing. Zhang Xianxian,,Hanson J H L,Arehart Kathryn. IEEE Int.Conf.Acoustic Speech Signal Process . 2004
[5]  
An alternative approach to linearly constrained adaptive beamforming. Griffiths L J,,Jim C W. IEEE Transactions on Antennas and Propagation . 1982
[6]  
Adaptive null-forming scheme in digital hearing aids. Luo Fa-Long,,Yang Jun,Pavlovic C,Nehorai A. IEEE Transactions on Signal Processing . 2002
[7]  
Multi-channel speech enhancement in a car environment using Wiener filtering and spectral subtraction. Meyer J,Simmer K U. IEEE Int.Conf.Acoustic Speech Signal Process . 1997
[8]  
Stochastic gradient-based implementation of spatially preprocessed speech istortion weighted multichannel Wiener filtering for noise reduction in hearing aids. Spriet A. IEEE Transactions on Signal Processing . 2005
[9]  
Image method for efficiently simulating small room acoustics. Allen J B,,Berkley D A. The Journal of The Acoustical Society of America . 1979
[10]  
Teleconferencing,system identification and array processing. hgren P. . 2001