论文部分内容阅读
在车载系统、电话会议与多媒体会议等语音处理系统中,由于受到混响、背景噪声及干扰等因素的影响,麦克风拾取的信号通常为带噪的语音信号。这样,不仅影响语音的可懂度,而且影响语音处理系统的整体性能。因此,需要进行有效的噪声抑制,以增强语音信号的质量。本文中重点介绍了采用LMS算法的自适应波束形成语音增强系统。自适应语音增强算法较其它方法相比,采用了一个参考噪声作为辅助输入,从而能获得比较全面的关于噪声的信息,因而可以得到更好的降噪效果。在分析研究LMS算法的基础上,本课题用Matlab语言对LMS算法实现自适应语音增强进行了仿真实现。实验结果表明,该算法在一定程度上可有效地抑制噪声,提高信噪比,减少失真。
In speech processing systems such as in-car systems, teleconferences and multimedia conferencing, signals picked up by the microphone are usually noisy speech signals due to factors such as reverberation, background noise and interference. This not only affects speech intelligibility but also affects the overall performance of the speech processing system. Therefore, effective noise suppression is required to enhance the quality of speech signals. This article focuses on the adaptive beamforming speech enhancement system using LMS algorithm. Compared with other methods, adaptive speech enhancement algorithm uses a reference noise as an auxiliary input, so as to obtain more comprehensive information about noise, so that a better noise reduction effect can be obtained. On the basis of analyzing and studying LMS algorithm, this subject uses Matlab language to simulate the LMS algorithm to achieve adaptive speech enhancement. Experimental results show that the proposed algorithm can effectively suppress noise, improve signal-to-noise ratio and reduce distortion to a certain extent.