Numeric:FFT inverse_real_fft doesn't trevni!

[Christian Tismer]

mhuster at hotmail.com wrote:
> 
> The FFT module with the LLNL Numeric stuff does not seem to perform an
> inverse on the result of real_fft. (At least not in an intuitive
> manner.) An N point real_fft returns N/2 +1 complex numbers. From
> experience, not documentation, I guess that element 0 is real and the DC
> component, and element N/2 is real and the Nyquist value.
> 
> But, trying inverse_real_fft does not seem to give the number of values
> or the values I expect. I thought that calling inverse_real_fft on the
> same array as real_fft produced would give back the original array,
> possibly scaled by N. Not so. I pasted in a sample of a session.

No, that's a misconception.
The real fft is just a speedup for cases where your data
has only real coefficients. Providing an inverse function
would be useless, since after any modifications of your
complex result, you will usually not expect a real result.
The inverse_real_fft is just a reverse FFT on real data,
likewise assuming that you are in the frequency domain
with real data. 
(sounds strange? Exactly as strange as for the time domain,
that's what causes all the Nyquist trouble :-)

ciao - chris

-- 
Christian Tismer             :^)   <mailto:tismer at appliedbiometrics.com>
Applied Biometrics GmbH      :     Have a break! Take a ride on Python's
Kaiserin-Augusta-Allee 101   :    *Starship* http://starship.python.net
10553 Berlin                 :     PGP key -> http://wwwkeys.pgp.net
PGP Fingerprint       E182 71C7 1A9D 66E9 9D15  D3CC D4D7 93E2 1FAE F6DF
     we're tired of banana software - shipped green, ripens at home