毕业论文--应用谱减法进行语音去噪的算法.doc

毕业设计(论文)
题目:应用谱减法进行语音去噪的算法
研究与仿真
系别
#######
专业班级
#########
学生姓名
##
指导教师
###
应用谱减法进行语音去噪的算法研究与仿真
摘要
语音在通话过程中要受到各种外界其他声音尤其噪音的影响,因此要对语音信号进行去噪处理,本课题就通过利用一种通用的语音去噪算法谱减法来进行语音去噪研究。传统谱减法虽然简单易行,但其处理过的语音含有音乐噪声且输出语音可懂度、质量仍有待提高,所以对传统谱减法进行了改进,通过仿真结果证明,改进的方法处理后的带噪语音可懂度和质量都有了很大的提高。
本课题首先研究传统谱减法然后对其进行改进研究。传统谱减法是在带噪语音频谱中减去噪声频谱,从而得到纯净的语音信号。在改进的谱减法中,添加了分帧加窗处理、噪声估计、增加谱减系数、平滑处理、阈值计算、去除音乐噪声处理等。最后仿真对最后结果进行分析,通过改进前后波形、输入输出的声音文件以及前后信噪比可以看出,改进后语音质量和语音可懂度都有了较大的提高。
关键词:语音去噪;谱减法;语音质量;语音可懂度

THE APPLICATION OF RESEARCH AND SIMULATION OF SPECTRAL SUBTRACTION ABOUT REMOVING THE NOISE
Abstract
Speech during munication will be affected by outside noise interference, so it’s important to reduce noise. This topic will research spectral subtraction one of the mon speech denoising algorithm. Traditional spectral subtraction simple and easy,but the speech its processed contains music noise and, the intelligibility and the quality of the output speech needs a further improvement. After the improvement, we make a simulation, though the results, it’s obviously that the quality of the speech and the intelligibility of the speech has improve a lot.
This subject first study traditional spectral subtraction and then the improvement research. Traditional spectral subtraction is minus the spectrum of the noise from the spectrum of the speech with noise, and get a clean speech. In the improving processing adds the frame and window treatment, noise estimates, increases the spectrum factor, runs smooth processing, threshold calculation, removes music noise, etc. Using matlab to simulation, through the input/output waveform, the sound file and the SNR, we can see the improved speech quality and intelligibility have greatly improved.
Keywords: Removing the noise; Spectral subtraction; the intelligibility of ;the speech the quality of the speech
目录
摘要 I
Abstract II
1绪论 1
1
1
本文各章节要点介绍 2
2 语音去噪的理论基础 3
3
3
评价标准 4
本章小结 5
3 应用谱减法