Page 1 of 1

Fourier synthesis

Posted: Thu Mar 26, 2009 5:02 am
by eric
Hey,

Does anyone know the feasibility of Fourier synthesis with SFP or Xcite or if it has been done.

Eric.

Re: Fourier synthesis

Posted: Thu Mar 26, 2009 2:16 pm
by at0m
IMHO, and in short: with the current SDK, FFT will not happen either on Scope or XITE.

Re: Fourier synthesis

Posted: Thu Mar 26, 2009 2:37 pm
by Shroomz~>
at0m wrote:IMHO, and in short: with the current SDK, FFT will not happen either on Scope or XITE.
I think that maybe the Harmonic Effects Bundle made by G.O.S.T involves FFT done in Scope's sdk. I could be wrong in assuming that though, so apologies in advance if I'm spreading misinformation. Maybe tgstgs or someone else can clarify. :wink:

Re: Fourier synthesis

Posted: Thu Mar 26, 2009 10:50 pm
by tgstgs
you need a sin a cos a mul and a div;
if one gets access to a pi somehow one may do the butterflywalz;
---------

we experimented a lot with ffts/dfts;
analyzer filter pitchshifter wavetables who knows ?
----------

thers no classical fft/dft inside the harmonic bundle;
fast fourier isnt fast enough for dynomic!
we found a much smarter way;


transformin vibes

Re: Fourier synthesis

Posted: Fri Mar 27, 2009 1:06 am
by tgstgs
when surching for a source i realized its written waltz in english!!

Re: Fourier synthesis

Posted: Fri Mar 27, 2009 4:40 am
by Shroomz~>
tgstgs wrote:thers no classical fft/dft inside the harmonic bundle;
fast fourier isnt fast enough for dynomic!
we found a much smarter way;
Sorry, my mistake! :oops:

Re: Fourier synthesis

Posted: Fri Mar 27, 2009 6:10 am
by Neutron
its easy to do Fourier synthesis, just not analysis

Re: Fourier synthesis

Posted: Fri Mar 27, 2009 4:09 pm
by valis
Depends on the analysis....STFT could be done on Scope quite easily I bet. You may have to resort to using math modules directly as RedMuze does and deal with the dsp-overhead this creates, but if one were to move beyond the typical FFT eq's & FIR filters I'll bet some interesting Fourier transforms are still out there :P

Re: Fourier synthesis

Posted: Fri Mar 27, 2009 5:06 pm
by Neutron
valis wrote:Depends on the analysis....STFT could be done on Scope quite easily I bet. You may have to resort to using math modules directly as RedMuze does and deal with the dsp-overhead this creates, but if one were to move beyond the typical FFT eq's & FIR filters I'll bet some interesting Fourier transforms are still out there :P
yes possible but not "easy" :D

and all the best hackers are figuring out ways to steal your identity or make your computer in to a zombie. not working in sdk like they should be.

Re: Fourier synthesis

Posted: Fri Mar 27, 2009 7:52 pm
by dawman
Since we can download devices for demos now, anyone who wants to take a leap into the future should check out the Harmonic Bundle from G.O.S.T.
There's nothing in the world like it, and if you are looking for a signature sound it can do amazing things to keyboards and audio in general.

http://forums.planetz.com/download/file.php?id=4026
Hendrix Synth...

http://forums.planetz.com/download/file.php?id=3915
Weird Synth Shit....

Re: Fourier synthesis

Posted: Fri Mar 27, 2009 7:56 pm
by Neutron
yet :D

Re: Fourier synthesis

Posted: Fri Mar 27, 2009 8:25 pm
by valis
Neutron wrote: yes possible but not "easy" :D

and all the best hackers are figuring out ways to steal your identity or make your computer in to a zombie. not working in sdk like they should be.
Right when I said 'easy' I guess I meant that STFT analysis & resynthesis can be less computationally demanding as doing fft based resynthesis, at least according to my experiences over a decade ago using csound. Not having an SDK license here, I'll be quite honest that I've no clue what it would take in scope. :D

Also there has been at least one device on Scope that integrated host-based FFT analysis with Scope, so that's another avenue for doing at least the analysis portion for more complex modelling.

Re: Fourier synthesis

Posted: Fri Mar 27, 2009 8:30 pm
by dawman
I like the word " yet ".
Much more than the word " soon. "
We have a Clintonesque definition of that word these days......
He did teach the world how to talk yourself into passing a lie detector test.... :lol:

" Real Soon Now " is most refreshing.... :D

Re: Fourier synthesis

Posted: Sat Mar 28, 2009 4:09 am
by tgstgs
transforming from time to frequency domain and back again is definitive possible;
---------

the told stft in spezial i never tried;
on the first look its probl. i would see in Mr. Gauß if my read was o.k.;
but thers always a workaround!
--------

if im allowed to give a hint for one developing without takin your fun:

build a workin butterflynetwork first;
then you are done for 90% of a basic fftstuff;

and alsways think about Mr. Fourier himself!
He had no calculater to press some buttons to get a result;
and he had no sdk with ready made atoms;
so one would say: he wasnt able to do a fft at all; NO?

------
its up to the developer how to implement which algo depending on the job to do;
they bsaicaly deliver all the same results _ more or less;
the quality on the job in detail is a direct indicator of the developers time spent on optimizing;


have a nice weekend and good vibes from vienna

Re: Fourier synthesis

Posted: Mon Mar 30, 2009 11:07 pm
by eric
So the answer is - yes maybe, and it's kinda been done? :lol:

I'll check out the G.O.S.T harmonic bundle, anyway.

Eric.