[sdiy] Re: granular synthesis (long + may be OT)

[jbv]

Thank you Fernando for this very dense introduction on
granular synthesis.

Having made quite a few experiments on this topic mydelf,
I can only second Fernando and recommend gran. synth. ,
may be not as a composition technique, but as a way to
achieve very rich and unpredictable sound textures.
You can even use these textures (far less boring than the
usual synth chords) in your sampler, or (if you're really
perverse) as sound sources for further granular experiments...

However, I'd like to discuss Fernando's distinction between
gran. synthesis and processing.
I'm not sure about what B. Truax and other gurus have to
say on that topic, but I consider that using "natural sound
sources" as just another form of synthesis, not really a
"processing"... Or if we put it another way, just a first
(and shy ) footstep in the field of gran. processing...

Indeed the "granular technique" can be extended, and various
processings can be made :
- granular FFT/iFFT  (through envelopes convolution)
- granular filtering
- granular reverb (to achieve very rich and cahotic textures
from simple sound sources like a singing human voice -
quite different from using voice samples for regular gran.
synthesis - and AFAIK granular reverb has been used as
a technique to emulate natural reverbs)
- granular phasing (adding grains of phasing organised in
random harmonic series on top of a simple synth drone for
instance)
- granular distorsion (adding grains of distorsion on top of
a simple synth drone for instance)
- etc etc

Of course, combining these techniques is also possible...
I tried to send a granular FFT/iFFT into a gran. reverb,
and well...

I think it's time for this technique to enter the world of
modular synths (via events generator modules for instance).
Of course, the question of realtime controls of all parameters
involved is still raised...
And I must confess that during my realtime experiments with
MAX, it was quite difficult to keep everything under control :
pretty soon the sound texture can become so dense and
unpredictable (especially with gran. reverbs) that it's hard
to figure out what effect will have a slight move of a slider...

May be this 3D interface controller idea is an option, since it
could give a realtime graphic transposition of the sound
processings, and thus help keeping things under control...
But when so much complexity is added to a synthesis or
processing technique, is it still interesting to keep things
under control ? May be just to make sure parameters
variations remain within some pre-defined boundaries...

JB