一种改进的基于人耳听觉掩蔽效应的语音增强算法  

An improved algorithm of speech enhancement based on human auditory masking effect

作　　者：蔡萍[1] 

机构地区：[1]闽江学院物理学与电子信息工程系,福建福州350108

出　　处：《闽江学院学报》2012年第2期70-72,共3页Journal of Minjiang University

摘　　要：提出一种结合几何谱减法的基于人耳听觉掩蔽效应的算法.通过仿真证明,该算法与一般的基于人耳听觉掩蔽效应的算法相比,能更好地减少音乐噪声,提高信噪比.在去噪的同时,减小了纯净语音的畸变,达到良好的效果.This paper puts forward a new algorithm combined with geometric spectral subtraction, which is based on human auditory masking effect. By carrying out simulation comparing this method with other methods based on auditory masking properties, conclusion is driven that this method has better performance on music noise reduction as well as higher speech-to-noise radio. Meanwhile, it reduces speech distortion when noise is cleared. In a word, this method achieves better effect than the others.

关 键 词：听觉掩蔽 谱减 几何谱减 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]