Voltage controlled arpeggiator (was Re: [sdiy] Design Process)

[David G. Dixon]

> You've just described a multibit ADC

Yes.  In fact, what I'm thinking of is a kind of quantizer, but one where
the binary number is further processed through logic to add the intervals.
For instance, say you wanted to make a full dominant 13 chord.  The notes
(relative to major, and numbered from the root) are 1, 3, 5, b7, 9, #11, 13,
1....  The intervals (in half-steps) are 4-3-3-4-4-3-3-....  Now, a typical
quantizer will produce the following intervals: 1-1-1-1-1-1-1....  Hence,
the logical operation here involves adding the correct binary numbers to
each output in order to obtain each successive output.  In this way, any
number of intervals can be activated or deactivated over a given range.  For
example, if one only wanted the sus 4 chord, one would require 1, 4, 5, 1,
4, 5, 1....  The intervals are 5-2-5-5-2-5-....  Hence, one would have to
add 4-1-4-4-1-4 to the output of the ADC before latching back into the DAC.
In this way, it doesn't matter to the quantizer how many intervals there are
or how widely spaced apart the notes of the arpeggio are.  It is simply
going up and down a certain number of steps as determined by the CV
attenuator.  It also doesn't particularly care on which step it starts the
pattern.

Now all I have to figure out is how to implement that logic efficiently.

For the more musically inclined among you, the way I would actually use this
thing is to have two oscillators; one playing a bass note, and the other
playing the arpeggio.  In this way, the bass note (which is optional) would
not have to be part of the arpeggio.  This frees the arpeggiator up to play
shells and upper structures over a two octave range, giving very rich
harmonic textures.  For example, if another oscillator is playing 1 (an
octave down), then the arpeggiator could be playing 3, b7, b9, #9, #11.  The
overall range of the chord is more than two octaves, but separating out the
root (or deleting it) gives more flexibility.  It also allows the bass note
to be changed (in this case, perhaps, to #4 in order to obtain the tritone
substitution) without reprogramming the arpeggio.

Now, of course, what you'd really want is to be able to program the arpeggio
in digitally and change it in real time, or even from the keyboard.  One
could even have a dedicated mini-keyboard which acted as the switch bank for
this device.  With that, one could play broken chords almost as if one had a
polyphonic synth.  With a couple of analog shift registers (particularly my
new and improved ones, which are damn-near perfect) operating at high clock
speed and driving several oscillators, it would almost be like having a
polyphonic synth.  Hmmm... better take another shower...