voice recognition using MFCC(feature extraction)

[shumaila.iqbal]

i have don with initial code of voice print. which can read voice and display its graph.now i want to impliment mfcc on this voice sample but hav problem in doing so..my codes are............
for voice print.......

function x;
%function FMatrix=mfccf(num,s,Fs)
Fs = 10000;   % Sampling Frequency (Hz)

Nseconds = 5; % Length of speech signal

fprintf('say a word immediately after hitting enter: ');

input('');

x  = wavrecord(Nseconds*Fs, Fs, 'int16'); % Recording the sound
 
wavplay(x,Fs);                             % Playing the recorded sound

ms20=Fs/50;                 % minimum speech Fx at 50Hz
ms1=Fs/1000;                 % maximum speech Fx at 1000Hz
t=(0:length(x)-1)/Fs;        % times of sampling instants
subplot(2,1,1);
plot(t,x);
legend('Waveform');
xlabel('Time (s)');
ylabel('Amplitude');
------------------------------------------------------ and for mfcc
function FMatrix=mfccf(num,x,Fs)
%Fs=1000;
n=512;              %Number of FFT points
Tf=0.025;           %Frame duration in seconds
N=Fs*Tf;            %Number of samples per frame
fn=24;              %Number of mel filters
l=length(x);        %total number of samples in speech
Ts=0.01;            %Frame step in seconds
FrameStep=Fs*Ts;    %Frame step in samples
a=1;
b=[1, -0.97];       %a and b are high pass filter coefficients

noFrames=floor(l/FrameStep);    %Maximum no of frames in speech sample
FMatrix=zeros(noFrames-2, num); %Matrix to hold cepstral coefficients
lifter=1:num;                   %Lifter vector index
lifter=1+floor((num)/2)*(sin(lifter*pi/num));%raised sine lifter version

if mean(abs(s)) > 0.01
    s=s/max(s);                     %Normalises to compensate for mic vol differences
end

%Segment the signal into overlapping frames and compute MFCC coefficients
for i=1:noFrames-2
    frame=x((i-1)*FrameStep+1:(i-1)*FrameStep+N);  %Holds individual frames
    Ce1=sum(frame.^2);          %Frame energy
    Ce2=max(Ce1,2e-22);         %floors to 2 X 10 raised to power -22
    Ce=log(Ce2);
    framef=filter(b,a,frame);   %High pass pre-emphasis filter
    F=framef.*hamming(N);       %multiplies each frame with hamming window
    FFTo=fft(F,N);              %computes the fft
    melf=melbankm(fn,n,Fs);     %creates 24 filter, mel filter bank
    halfn=1+floor(n/2);    
    spectr1=log10(melf*abs(FFTo(1:halfn)).^2);%result is mel-scale filtered
    spectr=max(spectr1(:),1e-22);
    c=dct(spectr);              %obtains DCT, changes to cepstral domain
    c(1)=Ce;                    %replaces first coefficient
    coeffs=c(1:num);            %retains first num coefficients
    ncoeffs=coeffs.*lifter';    %Multiplies coefficients by lifter value
    FMatrix(i, :)=ncoeffs';     %assigns mfcc coeffs to succesive rows i
end

%Call the deltacoeff function to compute derivatives of MFCC
%coefficients; add all together to yield a matrix with 3*num columns 
d=(deltacoeff(FMatrix)).*0.6;   %Computes delta-mfcc
d1=(deltacoeff(d)).*0.4;        %as above for delta-delta-mfcc
FMatrix=[FMatrix,d,d1];         %concatenates all together

-------------------------------------------------------
plzzzz help me how i can interrelate both of above codes

 

[mikkushukla]

Hi you can use jaudio for extracting MFCC feature..


But I also have another problem

I want make combination of extracted feature I am new in DSP

I want my feature should be

MFCC + derivative of MFCC + First Format Frequency + Fundamental Freuency.

I am done with the calculation of fundamental frequency but dont know how to proceed further :sad:

Kindly help me to proceed..

 

[moulalidas]

can anyone help me on how to extract speech features using mfcc???