Search All of the Math Forum:

Views expressed in these public forums are not endorsed by NCTM or The Math Forum.

Notice: We are no longer accepting new posts, but the forums will continue to be readable.

Topic: frequency to time domain using IFFT and FFT for a real-valued function
Replies: 2   Last Post: Nov 24, 2012 6:41 PM

 Messages: [ Previous | Next ]
 Greg Heath Posts: 6,387 Registered: 12/7/04
Re: frequency to time domain using IFFT and FFT for a real-valued function
Posted: Nov 24, 2012 6:41 PM

"nerdynerd" wrote in message <k8rero\$b4i\$1@newscl01ah.mathworks.com>...
> Some suggest that it might be a zero-padding issue; for FFT we can pad the end of the vector with zeros. but for IFFT we should add zeros in the middle of the matrix. I tried that as well but doesn't seem to work for me..
>
> "nerdynerd" wrote in message <k8rcmc\$4gc\$1@newscl01ah.mathworks.com>...

> > Hi, I have a real-valued function in frequency domain and I'm trying to IFFT to get the signal in time domain and then FFT back to frequency domain. However, I'm unable to get back the original signal. The first step IFFT gives me the correct function but FFT'ing it does not give the original signal. Any ideas what I'm doing wrong?
> > here's my code:

Debugging Tips

1.When you have problems with a large data set, ALWAYS debug with a small,
but representative set.
2. Remove semicolons from statements to check the result

>fsam=-10:0.1:10; % frequency vector

3. ==> Overkill. -7:0.5:7 is good enough for this example

>est_func=exp(-fsam.^2); % test-function

>fsamp=fsam(1)-fsam(2); % \Delta_frequency

4. ==>fsamp is negative

>nfft=2^12;

>r0=length(test_func);
>dt=1/fsamp;

6. ==>What is dt supposed to be?

>x = ifft(ifftshift(test_func),nfft); % DFT of signal
>g = (-nfft/2:nfft/2-1)*(1/(fsamp*nfft)); % Time range

7. ==> negative fsamp

8. Didn't check below here

> > q = ifftshift(x);
> >
> > dg=g(1)-g(2);
> > y = fft(fftshift((q)),nfft);
> > w = (-nfft/2:nfft/2-1)*(2*dg); %Frequency range; also tried "fsamp" instead of "2*dg"
> > s = fftshift(y);
> >
> > figure
> > subplot(2,1,1)
> > plot(fsam,testfunc) %original signal
> > hold on
> > plot(w,real(s),'r') %signal after IFFT and FFT
> > hold off
> >
> > subplot(2,1,2)
> > plot(g,real(q)) %real part of the signal after IFFT
> > hold on
> > plot(g,imag(q),'r') %imaginary part of the signal after IFFT
> > hold off

Hope this helps.

Greg

Date Subject Author
11/24/12 nerdynerd
11/24/12 nerdynerd
11/24/12 Greg Heath