Interpolating scanner again (was: AW: Ideas for a new circuit)

[Haible Juergen]

	> What do you think? Is this the way your circuit work too?

Yes. It's almost an exact description of my "Interpolating Scanner".
I use 8 stages. The VCAs are SSM2024's. 
The control section is one inverting opamp (level shift) plus one PNP 
waveshaper for the otas, plus a second PNP stage for driving a LED
array. (Today I would use discrete OTAs that can handle a few mA's
of bias current, and include the LEDs directly into the OTA's control
path ...)
Each input channel has a 1/4" Jack, a level potentiometer, and then
it goes into the ota. All ota outputs are directly conected together for 
current summing, and share one opamp buffer.
Each input jack is normalized to the previous one, so with less than
8 input signals there are no "dead" input regions. The first jack is
normalized to a 5V DC source (over a switch). So i can either use
the whole thing to scan separate audio inputs (It's located directly 
below the EMS Octave Filter Bank clone in my modular, for obvious
reasons ...), the same works for different  modulation inputs
of course. The other main (basic) aplication is not to plug anything
in and switch on the DC source instead, which is then passed from
input to input voia the normalized jacks. Then I can "draw" piecewise
linear waveforms (similar to the Matrix synths.)
But my favorite applications go a little further.
I can start with an audio signal from a VCO (preferably triangle) into
the "scan" input (the one that creates the movement thru the various
input channels). That's static, piecewise linear (PWL) waveshaping (setting
the
levels at the breakpoints manually). Even this can produce a lot of
sonic variety if I only change the input level of the scanning triangle
wave,
as only parts of the PWL courve are driven then. Or I can add a variable
DC shift over a second "scan" input, to change the active branch of the
PWL courve - that's like PWM, only with complex waves.
Then I can modulate some of the 8 signal inputs with envelopes, LFOs, and
so on. This results in a dynamical change of the PWL courve.
Then I can put another audio rate signal into one of the 8 inputs. Whenever
the "scanning" (first) oscillator signal is large enough (or dc shifted
enough)
to include one of these audio rate inputs in the scan, there are some more
or
less ringmodulation-like effects.
Now the best thing is that you can combine all these methods. Have a static
area in the middle (ch. 2 to 6), two different envelopes on the right (ch. 7
and 8)
and a second audio range signal on the left (ch.1) of the whole array. Use
the
triangle of the main oscillator for scanning, at a level that only comprises
scanning
of the middle channels. Then either perform a DC shift to go into the outer
regions,
or increase the triangle level to comprise all input sections with one scan.
There is no limit. You'll be surprised what strange "digital" timbres you
can 
produce with this totally analogue circuit.
This is one of the main reasons why most digital synths don't impress me at
all.
(;->)

JH.