时-频变换的语音信号分析与处理
2022-12-23 10:46:32
论文总字数:15785字
摘 要
作为语音信号处理的前提,语音信号分析技术是我们处理语音信号的基础。通过对语音信号的频域信息和时域信息进行分析,咱们可以得到语音信号的本质特征,这些特征能让我们对语音信号的消噪处理有更直观的理解。
迄今为止,时频分析技术一直就是人们关注的一个重点,平稳信号的分析用傅里叶变换是比较方便并且合适,而本文要分析和处理的语音信号是非平稳信号。短时距傅里叶变换克服傅里叶变换只能整体分析的弊端,通过移动时窗在时间轴上的的位置来截取平稳信号来得出信号的时频关系,从而再得到整体的关系。但在利用短时傅里叶变换分析的时候,由于高频和低频的时间局部化的窗函数是固定的,所以不能同时被满足。而小波变换可以针对非平稳信号进行分析,并且得到每一个时刻的频谱分量它能够自适应的去处理,能用不固定的时窗截取变化。本文主要就是来研究非平稳语音信号利用这两种时频分析方法来对语音信号进行时域和频域信息的分析,然后对加噪声的信号进行同样的分析,最后消噪再分析一次,对比这些处理过的信号,进行总结和学习。
关键词:短时傅里叶变换(STFT);连续小波变换(CWT);语音信号处理;时频分析;matlab仿真
Speech Signal Analysis and Processing of Time-frequency Conversion
Abstract
As the premise of speech signal processing, speech signal analysis technology is the basis of speech signal processing.By analyzing the frequency domain information and time domain information of speech signals, we can get the essential features of speech signals, which can make us have a more intuitive understanding of speech signal denoising.
Up to now, time-frequency analysis technology has been a focus of people's attention. Fourier transform is convenient and appropriate for the analysis of stationary signals, while the speech signals to be analyzed and processed in this paper are non-stationary signals.The short-time Fourier transform overcomes the disadvantage that the Fourier transform can only be analyzed as a whole, intercepts the stationary signal by moving the position of the window on the time axis to obtain the time-frequency relation of the signal, and then obtains the overall relation.However, in the analysis of short-time Fourier transform, the window function of high and low frequency time localization is fixed, so it cannot be satisfied at the same time.The wavelet transform can be used to analyze the non-stationary signal, and obtain the spectral component of each moment. It can be adaptive to process, and can intercept the change with the fixed time window.In this paper, the two time-frequency analysis methods are used to analyze the time-domain and frequency-domain information of non-stationary speech signals. Then, the same analysis is carried out for the signal with noise. Finally, the noise reduction is analyzed again.
Key words:Short-time Fourier transform,Continuous wavelet transform,Speech signal processing,Time-frequency analysis,The matlab simulation
目 录
摘 要 I
Abstract II
第一章 引言 1
1.1 课题的研究背景 1
1.2 课题的现实意义 1
1.3 研究方向的国内外现状 2
1.4 本文的内容以及安排 2
第二章 语音信号分析的基本原理 4
2.1 短时傅里叶变换(STFT) 4
2.1.1 傅里叶变换处理语音信号的局限 4
2.1.2 短时傅里叶变换的基本原理 4
2.1.3 短时傅里叶变换的特点 5
2.1.4 短时傅里叶变换的两种常见窗函数基本原理 9
2.2 小波变换(CWT) 11
2.2.1 小波变换的基本原理 11
2.2.2 小波变换的优缺点 13
2.2.3 小波基的具体函数举例 13
2.2.4语音信号分析的三种方法的特征和小结 14
第三章 语音信号分析处理仿真结果及分析 15
3.1 语音信号的时频分析仿真及结果分析 15
3.1.1 语音信号快速傅里叶变换(FFT)仿真与分析 15
3.1.2 语音信号短时傅里叶变换(STFT)仿真与分析 16
3.1.3 语音信号连续小波变换仿真与分析 19
第四章 总结以及展望 21
4.1 总 结 21
4.2 展 望 21
致 谢 22
参考文献 23
引言
1.1 课题的研究背景
自社科人文得到重视以来,咱们就可以通过远距离通信来把消息传递给远方。所以作为信号载体的语音信号被分析和处理进而得到发展是十分有意义的。
对有用的语音信息进行处理,比如去噪、滤波等这些常用处理是一项十分必要的工作,所以能够有效的获得语音信号的时频信息这一点显得尤为重要。
剩余内容已隐藏,请支付后下载全文,论文总字数:15785字