语声通信降噪研究

A research on voice communication de-noising

语声通信系统中,语声通过信道传输将不可避免地引入码间串扰和信号畸变,同时受到噪声污染。该文在分析自适应常数模盲均衡算法和改进盲均衡算法的基础上,考虑到自适应盲均衡技术在语声噪声控制方面能力有限,将自适应盲均衡技术与小波包掩蔽阈值降噪算法联合使用,形成一种基带语声增强新方法。仿真试验结果显示自适应盲均衡技术可以使星座图变得清晰而紧凑,有效减小误码率。研究证实该方法在语声信号码间串扰和畸变严重情况下,在白噪声及有色噪声的不同噪声环境中都具有稳定的降噪能力,消噪同时可获得汉语普通话良好的听觉效果。

In the voice communication system,the voice signal is transmitted through the channel,which introduces the inter-symbol interference(ISI)and signal distortion.At the same time,it is polluted by noise.Based on the analysis of adaptive blind equalization algorithm CMA(constant modulus algorithm)and improved blind equalization algorithm,and considering the limited ability of adaptive blind equalization in speech noise control,in this paper,a new base-band voice enhancement method is developed by combining adaptive blind equalization with wavelet packet masking threshold noise reduction algorithm.The simulation shows that the adaptive blind equalization technology can make the constellation map clear and compact,and reduce the error rate effectively.The research verifies that the method has stable noise reduction ability in different noise environments,such as white noise and colored noise,under the condition of ISI of speech signal and serious distortion.At the same time,it can get good auditory effect in Mandarin Chinese.