Hi everyone,
  I am getting the following error when I am implementing the code written below:
%Read the original signal
[y,fs]=wavread('angry');
>> a=1;
>> samples=fs*window_d; %Number of samples per window
rects=rectwin(samples); %rectangular window of duration 20ms
>> frame=y(a:a+samples-1)*rects';
>> acf=xcorr(frame);

error:?? Error using ==> ifft
Out of memory. Type HELP MEMORY for your options.

Error in ==> xcorr>matrixCorr at 94
c = ifft(C);

Error in ==> xcorr at 52
	[c,M,N] = matrixCorr(x);
 What would be the alternative way to implement the above code.Pls help and thanx in advance.

Maybe you want to specifically use the xcorr function but in order to use xcorr as an autocorrelation the option 'coeff' would need to be used. I have attempted using xcorr for some time however the output is the xcorr out combined with the autocorrelation out which becomes messy to handle when processing the data. I found a much faster, more straight forward, script at the follow link titled auto.m which was a huge help. Sorry i can't offer more assistance for your application. http://www.atmos.washington.edu/~dennis/552_Software.html

"astro mmi" <pyarsa_madhu@yahoo.co.in> wrote in message <hf6o4o$ho8$1@fred.mathworks.com>...
> Hi everyone,
>   I am getting the following error when I am implementing the code written below:
> %Read the original signal
> [y,fs]=wavread('angry');
> >> a=1;
> >> samples=fs*window_d; %Number of samples per window
> rects=rectwin(samples); %rectangular window of duration 20ms
> >> frame=y(a:a+samples-1)*rects';
> >> acf=xcorr(frame);
> 
> error:?? Error using ==> ifft
> Out of memory. Type HELP MEMORY for your options.
> 
> Error in ==> xcorr>matrixCorr at 94
> c = ifft(C);
> 
> Error in ==> xcorr at 52
> 	[c,M,N] = matrixCorr(x);
>  What would be the alternative way to implement the above code.Pls help and thanx in advance.

Can you tell us what size frame is? Is it a matrix and if it is, do you really want to input a matrix to xcorr()? I can see that rects' is a row vector (1xN), but if you are multiplying it by y a Mx1 column vector, than you're getting a MxN matrix as an input to xcorr() and that can get computationally expensive very quickly.
wayne

Thank you so much.The frame size is 320*320 i.e, after multiplication with the rectangular window.

"Wayne King" <wmkingty@gmail.com> wrote in message <hf6pjn$gus$1@fred.mathworks.com>...
> "astro mmi" <pyarsa_madhu@yahoo.co.in> wrote in message <hf6o4o$ho8$1@fred.mathworks.com>...
> > Hi everyone,
> >   I am getting the following error when I am implementing the code written below:
> > %Read the original signal
> > [y,fs]=wavread('angry');
> > >> a=1;
> > >> samples=fs*window_d; %Number of samples per window
> > rects=rectwin(samples); %rectangular window of duration 20ms
> > >> frame=y(a:a+samples-1)*rects';
> > >> acf=xcorr(frame);
> > 
> > error:?? Error using ==> ifft
> > Out of memory. Type HELP MEMORY for your options.
> > 
> > Error in ==> xcorr>matrixCorr at 94
> > c = ifft(C);
> > 
> > Error in ==> xcorr at 52
> > 	[c,M,N] = matrixCorr(x);
> >  What would be the alternative way to implement the above code.Pls help and thanx in advance.
> 
> Can you tell us what size frame is? Is it a matrix and if it is, do you really want to input a matrix to xcorr()? I can see that rects' is a row vector (1xN), but if you are multiplying it by y a Mx1 column vector, than you're getting a MxN matrix as an input to xcorr() and that can get computationally expensive very quickly.
> wayne

"astro mmi" <pyarsa_madhu@yahoo.co.in> wrote in message <hf6s68$m35$1@fred.mathworks.com>...
> Thank you so much.The frame size is 320*320 i.e, after multiplication with the rectangular window.
> 
> "Wayne King" <wmkingty@gmail.com> wrote in message <hf6pjn$gus$1@fred.mathworks.com>...
> > "astro mmi" <pyarsa_madhu@yahoo.co.in> wrote in message <hf6o4o$ho8$1@fred.mathworks.com>...
> > > Hi everyone,
> > >   I am getting the following error when I am implementing the code written below:
> > > %Read the original signal
> > > [y,fs]=wavread('angry');
> > > >> a=1;
> > > >> samples=fs*window_d; %Number of samples per window
> > > rects=rectwin(samples); %rectangular window of duration 20ms
> > > >> frame=y(a:a+samples-1)*rects';
> > > >> acf=xcorr(frame);
> > > 
> > > error:?? Error using ==> ifft
> > > Out of memory. Type HELP MEMORY for your options.
> > > 
> > > Error in ==> xcorr>matrixCorr at 94
> > > c = ifft(C);
> > > 
> > > Error in ==> xcorr at 52
> > > 	[c,M,N] = matrixCorr(x);
> > >  What would be the alternative way to implement the above code.Pls help and thanx in advance.
> > 
> > Can you tell us what size frame is? Is it a matrix and if it is, do you really want to input a matrix to xcorr()? I can see that rects' is a row vector (1xN), but if you are multiplying it by y a Mx1 column vector, than you're getting a MxN matrix as an input to xcorr() and that can get computationally expensive very quickly.
> > wayne

Do you really want to compute the autocorrelation on a matrix? or do you simply wish to compute the autocorrelation sequence for a 320 x1 vector? And why rectwin(), that's just a vector of ones.

But if I simply used the definition of the acf,still I'd be having  a problem with the index reference starting with a zero.


"Wayne King" <wmkingty@gmail.com> wrote in message <hf6tfg$9od$1@fred.mathworks.com>...
> "astro mmi" <pyarsa_madhu@yahoo.co.in> wrote in message <hf6s68$m35$1@fred.mathworks.com>...
> > Thank you so much.The frame size is 320*320 i.e, after multiplication with the rectangular window.
> > 
> > "Wayne King" <wmkingty@gmail.com> wrote in message <hf6pjn$gus$1@fred.mathworks.com>...
> > > "astro mmi" <pyarsa_madhu@yahoo.co.in> wrote in message <hf6o4o$ho8$1@fred.mathworks.com>...
> > > > Hi everyone,
> > > >   I am getting the following error when I am implementing the code written below:
> > > > %Read the original signal
> > > > [y,fs]=wavread('angry');
> > > > >> a=1;
> > > > >> samples=fs*window_d; %Number of samples per window
> > > > rects=rectwin(samples); %rectangular window of duration 20ms
> > > > >> frame=y(a:a+samples-1)*rects';
> > > > >> acf=xcorr(frame);
> > > > 
> > > > error:?? Error using ==> ifft
> > > > Out of memory. Type HELP MEMORY for your options.
> > > > 
> > > > Error in ==> xcorr>matrixCorr at 94
> > > > c = ifft(C);
> > > > 
> > > > Error in ==> xcorr at 52
> > > > 	[c,M,N] = matrixCorr(x);
> > > >  What would be the alternative way to implement the above code.Pls help and thanx in advance.
> > > 
> > > Can you tell us what size frame is? Is it a matrix and if it is, do you really want to input a matrix to xcorr()? I can see that rects' is a row vector (1xN), but if you are multiplying it by y a Mx1 column vector, than you're getting a MxN matrix as an input to xcorr() and that can get computationally expensive very quickly.
> > > wayne
> 
> Do you really want to compute the autocorrelation on a matrix? or do you simply wish to compute the autocorrelation sequence for a 320 x1 vector? And why rectwin(), that's just a vector of ones.