How can I show that the matlab IFFT is accurate enough?

I have no experience with fft and ifft and I just started using them. First I was asked to simulate a forced damped harmonic system in matlab (with code of course) and then do the fft analysis. Then I was asked to basically reconstruct the original wave from what I got back from fft, I thought of using ifft but shall I do it by hand? how hard ? shall I just use ifft? how accurate is it? Thanks everyone

Answer 1

[Wayne King]

   x = randn(1000,1);
   xdft = fft(x);
   y = ifft(xdft);
   max(abs(x-y))

As long as you don't modify the Fourier transform coefficients before you take the inverse, I think you'll find it's very accurate.

[Nina]

Thanks a lot, I will discuss this with my professor.