Matlab基于倒谱和EMD的语音基音周期的提取
关键字：基音周期 倒谱法 EMD8664
Title Pitch Period Extraction of Speech Signals based
on Cepstrum and EMD
Abstract
In voice signal processing, MFCC and some inherent characteristics of voice signals, such as the frequency of pitch. Cepstrum can be used to separate the excitation signal and channel response, and can represent channel response with only a dozen cepstral coefficients. As a result, it has been a very important role in voice signal processing. While cepstrum deconvolution method is largely influenced by additive noise,and anti-noise performance is not very good.The EMD method can be applied to decompose any type of signals,and thus,having a very distinct advantage in handing non-stationary and non-linear data.For this problem,in this paper,an improved algorithm of pitch period extraction of speech signals based on cepstrum and EMD is proposed. Its implementation in MATLAB are described in detail.
Key words: pitch period Cepstrum EMD
目 次
1 引言 1
1.1 背景 1
1.2 语音特性的提取 1
2 基于倒谱的基音周期的提取 2
2.1 基音的相关知识 2
2.1.1 基音周期 2
2.1.2 基音检测的难点 2
2.2 基音周期提取方法 3
2.3 倒谱和复倒谱 3
2.4 倒谱的特点 5
2.5 求倒谱的算法 6
2.6 倒谱提取基音周期算法的原理 7
2.7 MATLAB的设计与实现 8
3 EMD的相关知识 11
3.1 EMD 11
3.2 EMD的特点 12
3.2.1 本征模函数 12
3.2.2 EMD的分解 12
4 基于倒谱和EMD的基音周期的提取 13
4.1 倒谱和EMD 13
4.2 基于倒谱和EMD的基音提取流程图 13
4.3 MATLAB中的设计与实现 13
4.3.1 基于倒谱和EMD的基因提取新算法仿真 14
4.4 新算法与倒谱法的比较 14
4.4.1 简单音节的比较和复杂音节的比较 14
4.4.2 不同声调的单音节比较 17
4.4.3 同声母不同韵母的单音节比较 19
4.4.4 加噪下的单音节比较 25
结论 29
致谢 30
参考文献 31
上一篇：基于FPGA的信号发生器设计