mfccnew1.m

各种SVM分类算法

clear all
for g=1:20
  fname=sprintf('D:\\study\\师姐论文\\邢峰师姐\\邢峰论文程序\\邢峰论文程序\\audio files\\speech\\%d.wav',g);
  x=wavread(fname);
  x=x(:,1);
  [c,l]=wavedec(x,3,'db4');
  ca=appcoef(c,l,'db4',3);%近似系数
  x=ca;
%规一化mel滤波器组系数
bank=melbankm(24,128,22050);
bank=full(bank);
bank=bank/max(bank(:));
%DCT系数，12×24
for k=1:12
    n=0:23;
    dctcoef(k,:)=cos((2*n+1)*k*pi/(2*24));
end
%规一化倒谱提升窗口
w=1+6*sin(pi*[1:12]./12);
w=w/max(w);
%预加重滤波器
xx=double(x);
xx=filter([1-0.9375],1,xx);
%音频信号分帧
xx=enframe(xx,128,64);%对信号x分帧处理，每帧有128个采样点，相邻帧之间有50%的叠加率
%计算每帧的MFCC参数
for i=1:size(xx,1)
    y=xx(i,:);
    s=y'.*hamming(128);
    t=abs(fft(s));
    t=t.^2;
    c1=dctcoef*log(bank*t(1:65));
    c2=c1.*w';
    m(i,:)=c2';
end
[row col]=size(m);
mfcc=zeros(1,12);
  for i=1:col
      for j=1:row  
        mfcc(1,i)=mean(m(j,i));
      end
  end
mfccspeech(g,:)=mfcc;
end
mfccspeech


clear all
for g=1:20
  fname=sprintf('D:\\study\\师姐论文\\邢峰师姐\\邢峰论文程序\\邢峰论文程序\\audio files\\music\\%d.wav',g);
  x=wavread(fname);
  x=x(:,1);
  [c,l]=wavedec(x,3,'db4');
  ca=appcoef(c,l,'db4',3);%近似系数
  x=ca;
%规一化mel滤波器组系数
bank=melbankm(24,128,22050);
bank=full(bank);
bank=bank/max(bank(:));
%DCT系数，12×24
for k=1:12
    n=0:23;
    dctcoef(k,:)=cos((2*n+1)*k*pi/(2*24));
end
%规一化倒谱提升窗口
w=1+6*sin(pi*[1:12]./12);
w=w/max(w);
%预加重滤波器
xx=double(x);
xx=filter([1-0.9375],1,xx);
%音频信号分帧
xx=enframe(xx,128,64);%对信号x分帧处理，每帧有128个采样点，相邻帧之间有50%的叠加率
%计算每帧的MFCC参数
for i=1:size(xx,1)
    y=xx(i,:);
    s=y'.*hamming(128);
    t=abs(fft(s));
    t=t.^2;
    c1=dctcoef*log(bank*t(1:65));
    c2=c1.*w';
    m(i,:)=c2';
end
[row col]=size(m);
mfcc=zeros(1,12);
  for i=1:col
      for j=1:row  
        mfcc(1,i)=mean(m(j,i));
      end
  end
mfccmusic(g,:)=mfcc;
end
mfccmusic


clear all
for g=1:20
  fname=sprintf('D:\\study\\师姐论文\\邢峰师姐\\邢峰论文程序\\邢峰论文程序\\audio files\\speech with music\\%d.wav',g);
  x=wavread(fname);
  x=x(:,1);
  [c,l]=wavedec(x,3,'db4');
  ca=appcoef(c,l,'db4',3);%近似系数
  x=ca;
%规一化mel滤波器组系数
bank=melbankm(24,128,22050);
bank=full(bank);
bank=bank/max(bank(:));
%DCT系数，12×24
for k=1:12
    n=0:23;
    dctcoef(k,:)=cos((2*n+1)*k*pi/(2*24));
end
%规一化倒谱提升窗口
w=1+6*sin(pi*[1:12]./12);
w=w/max(w);
%预加重滤波器
xx=double(x);
xx=filter([1-0.9375],1,xx);
%音频信号分帧
xx=enframe(xx,128,64);%对信号x分帧处理，每帧有128个采样点，相邻帧之间有50%的叠加率
%计算每帧的MFCC参数
for i=1:size(xx,1)
    y=xx(i,:);
    s=y'.*hamming(128);
    t=abs(fft(s));
    t=t.^2;
    c1=dctcoef*log(bank*t(1:65));
    c2=c1.*w';
    m(i,:)=c2';
end
[row col]=size(m);
mfcc=zeros(1,12);
  for i=1:col
      for j=1:row  
        mfcc(1,i)=mean(m(j,i));
      end
  end
mfccswm(g,:)=mfcc;
end
mfccswm