Sir, I tried for the feature extraction of a speech using mel frequency cepstral coefficient (mfcc) but the code now showing error. I dont know how to rectify this error. So, Sir can you please help me to rectify this error.

The code is given below,

[audio, fs1] = audioread('cryrumble.wav');
%sound(x,fs1);
   ts1=1/fs1;
    N1=length(audio);
    Tmax1=(N1-1)*ts1;
    t1=(0:ts1:Tmax1);
    figure;
  plot(t1,audio),xlabel('Time'),title('Original audio');
%   fs2 = (20/441)*fs1;
%   y=resample(audio,2000,44100);
% %sound(y,fs2);
%    ts2=1/fs2;
%    N2=length(y);
%    Tmax2=(N2-1)*ts2;
%    t2=(0:ts2:Tmax2);
%    figure;
%    plot(t2,y),xlabel('Time'),title('resampled audio');



 %Step 1: Pre-Emphasis
 
 a=[1];
     b=[1 -0.95];
    z=filter(b,a,audio);
    subplot(413),plot(t1,z),xlabel('Time'),title('Signal After High Pass Filter - Time Domain');
     subplot(414),plot(fs1,fftshift(abs(fft(z)))),xlabel('Freq (Hz)'),title('Signal After High Pass Filter - Frequency Spectrum');
       nchan = size(audio,2);
for chan = 1 : nchan
  %subplot(1, nchan, chan)
  spectrogram(y(:,chan), 256, [], 25, 2000, 'yaxis');
  title( sprintf('spectrogram of resampled audio ' ) );
end

% Step 2: Frame Blocking

     frameSize=1000;
%     frameOverlap=128;
%     frames=enframe(y,frameSize,frameOverlap);
%     NumFrames=size(frames,1);
frame_duration=0.03;
frame_len = frame_duration*fs1;
framestep=0.01;
framestep_len=framestep*fs1;
% N = length (x);
num_frames =floor(N2/frame_len);
% new_sig =zeros(N,1);
% count=0;
% frame1 =x(1:frame_len);
% frame2 =x(frame_len+1:frame_len*2);
% frame3 =x(frame_len*2+1:frame_len*3);
frames=[];
for j=1:num_frames
     frame=z((j-1)*framestep_len + 1: ((j-1)*framestep_len)+frame_len);
%     frame=x((j-1)*frame_len +1 :frame_len*j);
%     identify the silence by finding frames with max amplitude less than
%     0.025
max_val=max(frame);
   if (max_val>0.025)
%     count = count+1;
%     new_sig((count-1)*frame_len+1:frame_len*count)=frames;
    frames=[frames;frame];
   end
   end
   % Step 3: Hamming Windowing

NumFrames=size(frames,1);
hamm=hamming(1000)';
windowed = bsxfun(@times, frames, hamm);

      
     % Step 4: FFT 
% Taking only the positive values in the FFT that is the first half of the frame after being computed. 


       ft = abs( fft(windowed,500, 2) );
       plot(ft);
   
% Step 5: Mel Filterbanks
Lower_Frequency = 100;
Upper_Frequency = fs1/2;
% With a total of 22 points we can create 20 filters.
    Nofilters=20;
    lowhigh=[300 fs/2];
    %Here logarithm is of base 'e'
    lh_mel=1125*(log(1+lowhigh/700));
    mel=linspace(lh_mel(1),lh_mel(2),Nofilters+2);
    figure;
    plot(mel);
    xlabel('frequency in Hertz');ylabel('mels');
    title('melscale');
    melinhz=700*(exp(mel/1125)-1);
    %Converting to frequency resolution
    fres=floor(((frameSize)+1)*melinhz/fs2); 
    %Creating the filters
    for m =2:length(mel)-1
        for k=1:frameSize/2
     if k<fres(m-1)
        H(m-1,k) = 0;
    elseif (k>=fres(m-1)&&k<=fres(m))
        H(m-1,k)= (k-fres(m-1))/(fres(m)-fres(m-1));
    elseif (k>=fres(m)&&k<=fres(m+1))
       H(m-1,k)= (fres(m+1)-k)/(fres(m+1)-fres(m));
    elseif k>fres(m+1)
        H(m-1,k) = 0;    
     end 
        end
    end
        %H contains the 20 filterbanks, we now apply it to the processed signal.
    for i=1:NumFrames
    for j=1:Nofilters
        bankans(i,j)=sum((ft(i,:).*H(j,:)).^2);
    end
    end
    figure;
    plot(bankans(i,j));
    figure;
    plot(H);
    xlabel('Frequency');ylabel('Magnitude');
    title('Mel-Frequency Filter bank');
    % Step 6: Nautral Log and DCT

%     pkg load signal
    %Here logarithm is of base '10'
    logged=log10(bankans);
    for i=1:NumFrames

        mfcc(i,:)=dct2(logged(i,:));
    end
    %plotting the MFCC
    figure 
    hold on
    for i=1:NumFrames
        plot(mfcc(i,1:13));
        title('mfcc');
    end
    hold off
% save c5 mfcc

i= mfcc;
save i i
load i.mat
X=i;
k=1;
[IDXi,ci] = kmeans(X,k);
save c41i ci
 

The error is showing like this:

>> mfccfinal
Error using bsxfun
Non-singleton dimensions of the two input arrays must match each other.

Error in mfccfinal (line 70)
windowed = bsxfun(@times, frames, hamm);
 

[KALYAN ACHARJYA]

It have multiples error, define fs1 and for debug put the plots statements in comment section, and see is there any error.

Answer 1

[Walter Roberson]

You failed to set the hamming window size to either the frame size or the number of frames .

Also your frames variable is probably a column vector. you construct frame by indexing a column vector with a row vector. When you index a vector with a vector the result has the same orientation as the vector being indexed which is column vector in this case. Therefore frame is a column vector and you vertcat those together which gives you a column vector result .

I recommend that you use buffer() instead of breaking up the array yourself .

[Suchithra K S]

Sir, then how can I correct the code. Can you please tell me the corrected code.

[Walter Roberson]

You have a column vector frames of unknown size. You try to multiply it by a column vector hamm of fixed size. What are your options?

1. Truncate the column vector frames to be the same length as hamm so that you are multiplying two column vectors of the same size ?
2. Replicate hamm to be the same size column vector as frames is ?
3. Use a hamming window size that is the same length as your column vector frames so that again they are the same size ?
4. Apply some kind of mathematical transformation to frames (perhaps interpolation) to reduce it down to the same size as the column vector hamm ?
5. Apply some kind of mathematical transformation to hamm (perhaps interpolation) to expand it to be the same length as the column vector frames ?
6. Find some multiplying matrices to pre and post multiply frames by so that it becomes the same length as hamm ?
7. Find some multiplying matrices to pre and post multiply hamm by so that it becomes the same length as frames ?
8. Something else?

Which of these makes the most sense to you?

[Romody Momoto Sogavo]

Sir, regarding this code I tried running different audio.WAV files ( less then 5 seconds long each) but frames came up empty. As can be seen from my workspace "frame [ ]" but hamm = 1x1000 double. I'm thinking the problem is with frames not the hamm. if so what should i do to rectify this?

[Walter Roberson]

Which of the 8 choices I listed did you choose?

https://en.wikipedia.org/wiki/Window_function