Sitars and Waveshaping

[media at mail1.nai.net]

>Anyone have any idea what type of waveshaper could be built to simulate a
>sitar? I can't even think of where to begin looking for clues.  The sitar
>seems pretty nonlinear (the buzzing sound is caused by a curved bridge);
>conventional synthesis doesn't seem like it would simulate it too well.  It
>would be great to build a box that would create sitar-like effects that
>would be volume-dependent, so as to work with guitar (or with synth, if the
>waveshaper is preceeded by a VCA).

I'm no sitar player, but I have tried to get sitar sounds by using um uh a
"popular self-sequenced monosynth" and a six-band digital resonator.  You
see, a sitar has several strings that serve as drones -- they resonate when
other notes are sounded, but aren't directly played.  What I found is that
you have to seriously detune the resonators so that they do not have the
relationship of a Western scale, but I still couldn't get it right.  Does
anyone know the specific frequencies of a sitar's drones??

So what I think you need to get the effect of these drones are number of
resonant filters in parallel.  You might also have to add in a way to mix
in feedback as in a real sitar all the parts can "hear" each other.

Beyond that you will need a contoller that can play Indian scales.  Since I
believe (but I am guessing based on vague memories) the drones are also
fretted, you will need a seperate controller for the filters as well.

Otoh, most Indian food markets sell cd's on the side and samplers are
getting cheaper everyday :)

>Also, anyone out there have much experience with cool waveshapers?  I would
>love to build circuits much like the Serge Wave Multiplier.  As far as I can
>tell, the Wave Multiplier works as follows (these are just guesses - I have
>never looked inside of one of these beasts):

I have no experience with waveshapers except for looking at this big pile
of xeroxed schematics I have to mail around the world :)

Anyway, this simplest way to change the shape of a wave is to clip it with
a diode -- you know, like a distortion pedal with two diodes going each way
from signal to ground.  For some reason using silicon diodes isn't
considered cool even though many guitar amps deliberately have them in
their signal path solely for that purpose.  "Germanium Fuzz" was rather
popular.  Other folks have tried zeners and borrowing the pn junctions from
numerous other semiconductors in order to get a unique sound.

I had read this guy had figured out with an RTA and a signal generator that
clipping with LED's was the closest thing to the distortion created by
overdriving a tube amplifier.  He based this on the fact that the LED's
created the most even harmonics.  Using LED's might make a great "tube
distortion" box, but I believe the output of a guitar and a sine wave are
too disimilar to make such a conclusion based on such a test.

But you weren't trying to run a veena through a Marshall stack were you?? :)

PEACE OUT :)
MARK