and FFT-strech the frequencybands

smedge marius at teleweb.at
Tue Oct 31 09:19:37 CET 2000


hi list
i´m searching for a mathematical genius, who can help me with the
following problem.
i tried to build a patch, which can read out from fft~ results in
various speeds (for example to spread the frequencies 0-10000 over 
the whole frequency range of 22050 points). 
i took a squared phasor~ to read out the 2048 points of the
frequency-adresses (both real and imaginary part) of the fft.
but the sound-result was flickery and poor. now i think my
problem is the wrong handelling of the real- and imaginary-parts (or
their vector-sum) in the algorithm and re-fft. or would i have to change
the fft-process itself?
any ideas? 
marius.



More information about the Pd-list mailing list