Doepfer

Weeeelllll.....

there is no reason a DSP-based granular module should not be able to  
output n grains to n outputs. Grains are computed as separate  
entities and get summed just before the output. So skipping the  
summing stage and replacing it with n outputs would not be such an  
issue. It would raise the price of a module because you'd use more D/ 
A's, though. So filter, amp and panning could all be done outside  
such a module, provided the module outputs trigger data etc for each  
grain.

What would really make the whole thing rely on a high-powered DSP (or  
more than one) is if you want to be able to modulate transposition/ 
position etc at audio rate....IMHO a must in a modular (can you spell  
"control rate vs audio rate" in reaktor? ugh)

I personally think that an 8 or 16 grain system would be sufficient,  
provided there was a way to link more than one in an "intelligent"  
manner.

Now what i think is the main issue with a module like this is: what  
control philosophy does it use? You can either have dedicated control  
of every grain or use some sort of "structural logic" that the module  
itself applies (aka "grain density" instead of separate grain  
triggers, "pitch jitter" instead of CV control of every grain pitch,  
"density evolution" instead of variable rate grain trigger signal). I  
personally think that in a modular system you'd want the discrete  
version, if i want meta-control only i can use reaktor or anything  
else.....the cool thing is to have the grain stuff *integrate*  
perfectly into the system, and for that you definitely need access to  
the individual grain's parameters. Now the downside of this is that  
it would  be a large and expensive module, and that doing some of the  
standard meta-attribute stuff would take a lot of patching. Maybe  
here there would be a good opportunity to create some structured data  
modules.....;-)


Now imagine what you could do with a grain thang at audio rate....use  
an audio rate pulse for the triggering of a grain, clock a sequencer  
with the same pulse and have that select playback position and grain  
transposition (or grain duration) of the grain thang, effectively  
getting a waveform that is F*N in length ( F being pulse frequency  
expressed in seconds, N being number of steps in the sequence). And  
we're just talking about grain #1 here :-)


my 2 ct

d



On 05. Mar 2008, at 9:12 AM, Korhan Erel wrote:

> Hi Carlos,
>
> As I said in my earlier mail that even a 4-grain system would cost a
> fortune. On the other hand, a module that includes a DSP programmed  
> to do
> granular synthesis would also cost a lot, given that it will have  
> to be a
> powerful DSP to handle a high number of grains. It would also have to
> include features like slew, filter, pitch shifting, amp, panning,  
> etc, as
> you would not be able to process grains separately with modules.
>
> Korhan
>
> On 3/5/08, Carlos <bushwick@gmail.com> wrote:
> >
> > Granular synthesis uses anywhere from a few dozen grains to  
> thousands of
> > them.being able to change the number of grains is essential to  
> get a good
> > variety of sounds, I don't think a few fixed grains would do the  
> trick
> >
> > --- In Doepfer_a100@yahoogroups.com <Doepfer_a100% 
> 40yahoogroups.com>,
> > "Korhan Erel" <listekutusu@...> wrote:
> > >
> > > Since the A-112s cannot toggle between record and play, my idea  
> of the
> > > granular synthesis can work only on prerecorded samples  
> (assuming it was
> > a
> > > good/feasible idea to start with)
> > >
> > > On 3/3/08, Korhan Erel <listekutusu@...> wrote:
> > > >
> > > > Can several A-112s be used to construct a simple granular  
> synth, with
> > the
> > > > A-112s in wavetable mode? Each A-112 would represent a grain  
> going
> > into a
> > > > filter and then into a VCA, shaped by a ADSR? I guess a 4  
> grain synth
> > would
> > > > cost a small fortune, but it's probably worth it.
> > > >
> > > > On 2/17/08, gasp_uleg <uleg2@...> wrote:
> > > > >
> > > > > i am a fanatic of modular concept but i never had the money  
> to build
> > > > > the gigantic modular of my dreams so i went into virtual and i
> > > > > specialized myself into Reaktor, where i've found  
> everythink i ever
> > > > > needed.
> > > > >
> > > > > but the thing is that i deeply hate to use computer in  
> concerts so
> > > > > now i'm trying to rebuild my reaktor setups into a totally  
> physical
> > > > > way. with A100 and the other modulars everything i need can
> > perfectly
> > > > > be done EXCEPT granulation which has become my favourite sound
> > > > > generator and modifier latelly.
> > > > >
> > > > > i think not only granular cloud-delay processing is  
> interesting.
> > also
> > > > > a granular sample player would be an amazing module.
> > > > >
> > > > > i don't know much about DSP but if people like clavia have  
> been able
> > > > > to fit all what is inside a micromodular or if pedal  
> manufacturers
> > > > > are now using DSP for their new guitar pedals, it must be  
> possible
> > to
> > > > > find some cheap DSP technology to be fit into a vc modular  
> format.
> > > > >
> > > > > of course i understand development and technical expertise  
> is very
> > > > > diferent for digital stuff than
> > > > > for purelly analog electronics design but i'm sure ther are  
> many
> > > > > young geniuses running out there with their nice diplomas  
> in "very
> > > > > really complicated informatics" just willing to be emploid to
> > program
> > > > > the complicated DSP algorithms for our modulars.
> > > > >
> > > > >
> > > > >
> > > >
> > > >
> > >
> > >
> > > [Non-text portions of this message have been removed]
> > >
> >
> >
> >
>
> [Non-text portions of this message have been removed]
>
>
> 



[Non-text portions of this message have been removed]