Everything you could possbile want to know about FFTs. [wink] Happy hunting!

Are you supposed to treat the bits of the byte as a signal (ie. pwm square wave), because you calculate 1-dimensional DFT (or any ft) for a signal, not scalar. Such a signal could be a time serie (such as audio signal), or serie of spatial samples. So you could form a serie of that random byte by taking the 8 bits and treating them as 8 samples (1 or 0).

Recall that sum of signals (ie series) is equal to sum of the Fourier transforms, so you can simplify your task by calculating the FT of a simple square pulse with width of one bit and start time of the pulse as a parameter and just summing those according to the bits.

with pulse I mean something like this: __---__ where the start time is the time when pulse goes up. Time could be expressed easily in samples (atleast in the case of DFT).

edit: the bits could also be defined as values -1 and 1.. this gives slightly different results (changes the DC component). Oh.. and of course the width of the bit you use could be wider than one sample, which again gives you different results. Perhaps you should parametrize the width also.

Then again I might have misunderstood the whole assignment :)