AnsweredAssumed Answered

ifft size

Question asked by VRFuser on Apr 28, 1998
  • Comment • 0
Here's a relatively arcane issue that I hope someone has the answer for.

I understand that given the vagarities of FFT's if I start with
a waveform of say 256 _or_ 257 points I will end up with an fft
of 129 points. That's fine.

Conversely, if I take an ifft of the spectrum (129) points I can
get either a 256 or 257 point waveform.

The question is- how do I control which waveform to get?

By experiment, if I take a waveform and then take an fft and an ifft
I get exactly the original waveform

ifft(fft(256 point waveform))  gives a 256 point waveform
ifft(fft(257 point waveform))  gives a 257 point waveform

This is great.

HOWEVER, it seems that if I do anything to the spectrum in between
the fft and ifft, then I don't know what I will get back

for example:

ifft(randomize(fft(256 point waveform)))  gives a 257 point waveform
ifft(sin(fft(256 point waveform)))  gives a 256 point waveform

and it seems that any time I create a spectrum using the BUILD SPECTRUM
object I get  the 257 point waveform

ifft(129 point spectrum built from BUILD SPECTRUM)   gives a 257 point waveform


***************

BOTTOM LINE:  I need to be able to control this so that there is a consistent
result!

Any hints would be appreciated.

thanks

Stan Bischof


----------------------------------------------------------------------
Stan Bischof   Hewlett Packard Company   707-577-3994    rsb@sr.hp.com
----------------------------------------------------------------------

No one else has this question

Outcomes