【语音识别】基于MFCC和MEL倒频系数实现声纹识别附matlab代码

1 内容介绍

提出了以Mel频率倒谱系数(Mel Frequency Cepstrum Coefficients,MFCC)和MEL倒频系数作为特征提取技术,以KNN作为分类器的语音识别方法,实验结果表明,对于容易混淆的英文单词,该方法语音识别准确率较高.​

2 部分代码

function f=enframe(x,win,inc) 

%ENFRAME split signal up into (overlapping) frames: one per row. F=(X,WIN,INC) 

% 

% F = ENFRAME(X,LEN) splits the vector X up into 

% frames. Each frame is of length LEN and occupies 

% one row of the output matrix. The last few frames of X 

% will be ignored if its length is not divisible by LEN. 

% It is an error if X is shorter than LEN. 

% 

% F = ENFRAME(X,LEN,INC) has frames beginning at increments of INC 

% The centre of frame I is X((I-1)*INC+(LEN+1)/2) for I=1,2,... 

% The number of frames is fix((length(X)-LEN+INC)/INC) 

% 

% F = ENFRAME(X,WINDOW) or ENFRAME(X,WINDOW,INC) multiplies 

% each frame by WINDOW(:) 

%    Copyright (C) Mike Brookes 1997 

%      Version: $Id: enframe.m,v 1.3 2005/02/21 15:22:12 dmb Exp $ 

% 

%   VOICEBOX is a MATLAB toolbox for speech processing. 

%   Home page: http://www.ee.ic.ac.uk/hp/staff/dmb/voicebox/voicebox.html 

% 

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 

%   This program is free software; you can redistribute it and/or modify 

%   it under the terms of the GNU General Public License as published by 

%   the Free Software Foundation; either version 2 of the License, or 

%   (at your option) any later version. 

% 

%   This program is distributed in the hope that it will be useful, 

%   but WITHOUT ANY WARRANTY; without even the implied warranty of 

%   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the 

%   GNU General Public License for more details. 

% 

%   You can obtain a copy of the GNU General Public License from 

%   ftp://prep.ai.mit.edu/pub/gnu/COPYING-2.0 or by writing to 

%   Free Software Foundation, Inc.,675 Mass Ave, Cambridge, MA 02139, USA. 

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 

nx=length(x); 

nwin=length(win); 

if (nwin == 1) 

   len = win; 

else 

   len = nwin; 

end 

if (nargin < 3) 

   inc = len; 

end 

nf = fix((nx-len+inc)/inc); 

f=zeros(nf,len); 

indf= inc*(0:(nf-1)).'; 

inds = (1:len); 

f(:) = x(indf(:,ones(1,len))+inds(ones(nf,1),:)); 

if (nwin > 1) 

    w = win(:)'; 

    f = f .* w(ones(nf,1),:); 

end 

3 运行结果

4 参考文献

[1]李玲俐. 一种基于MFCC和SVM的语音识别方法[J]. 软件导刊, 2012, 11(3):2.

博主简介：擅长智能优化算法、神经网络预测、信号处理、元胞自动机、图像处理、路径规划、无人机、雷达通信、无线传感器等多种领域的Matlab仿真，相关matlab代码问题可私信交流。

部分理论引用网络文献，若有侵权联系博主删除。