>I'll have a look at that page to see if I can understand what's happening.
>There's also a lot of information on the Doepfer site
yep, that's how it all started for me :-) i got there by accident and
realized that it shouldn't be too difficult to dp it in software.
>IIRC, the division is
>done 'digitally', and the saw waveform reconstucted from this.
>OTOH, if authenticity is not the major concern, it would be easier
>to start with the fundamental saw waveform: this will give you the
>required slope ('delta' in you message) directly from the start.
>Note that if you take the original saw, and add a 'staircase' from
>the ouput of a counter, the result will be a new saw at the
My original patch simply used the 4 osc organ module from ams for this.
But it just does not sound like the original.
OK, the "right" choice of filters and their respective settings, and of
course that special "band manual" propably make the monster part of
this unique sound, but then again i thought that the sub osc generators
are propably the easiest thing to start with. Or put in other words, i
simply needed an excuse for starting to write my first self-made ladspa
In the mean time i had a friend of mine (who did study physics) look at
the schematics, and that brought a little bit of light into some of the
issues. It seems that in that circuit there is a capacitor whose
"unloading curve" is used for generating the saw wave. So i changed the
behavour of my plugin in two aspects:
- i don't use a linear ramp for the saw wave but rather something like
exp( -t * 1/capacity ) (I hope that comes close enough to what happens
in a capacitor)
- i don't use the fall off speed (1/capacity) to compensate for
increased master freq, but rather make capacity a fixed value, try to
get rid of the resulting dc offset (not optimal yet, how would this be
done in hardware? ) and finaly multiply with exp(pitch) (where pitch is
that V/Octave thing, not the actual freq.) (also not optimal yet. I was
to tired last night to completely think it over)
The result does not sound that bad at all, although there are a some
artifacts which i can't explain. Also the problem with the (desired)
high master freq remains to be solved. Something is still rotten there,
i don't know what, if master freq exeeds a certain point, the results
become chaotic. Maybe also an numerical instability in my code.
But all in all i'm slowly getting somewhere. Another friend of mine
seems to become involved enough into the whole thing, so he might
consider to build a real "band manual" which i can then connect to the
joystick port and use as an input device. This would be realy something
to show of with at the next lad meeting :-). But, heck, i still need
LOTS of more midi sliders to actualy play that thing.
One thing that i am badly missing though is knowledge and experience. I
try to listen carefully to the records of Sala, and try to make some
conclusions on what happens there technicaly. This leaves a lot of
things obscure to me. How does he create that incredible deep and
voluminous bass sounds? How does he setup the freq shifter, so that the
shifted signal sounds in perfect harmony to the original sound? I think
i've got the wiring correct, but the shifted signal turns into a noisy
mess pretty soon. So it is usable as an effect, but not exactly as a
means of adding "another voice to the choir".
I think this evening, I will upload a alpha-state version of my current
patch and the plugin, so anyone interested can play arround a bit.