Please bear with me on a long ramble (or you can just delete this Email now,
Mr. Phelps---it's always your choice).
I've been thinking about an idea for a module that I'd like to throw out to
the group for merciless comment. This discussion probably wont interest 95%
of you, and wont even make _sense_ to anyone that does not understand the
relationships of partials (harmonics) and how they mix to create a sound in
additive/'Fourier-fashion'... but if created, I think many people would see
the value in it once they heard it.
Anyway:
This idea sprang out of Mr. Hendry's recently posted concept for an octave
switch a week or so ago. To refresh your memories, his idea was something
like this:
The MOTM VCO does not have an octave switch, so whenever you want to tune
multiple VCOs to octaves-apart, it must be done carefully with the
continuous pots. Now, if one wanted to add this feature, modifying the VCO
is a bad idea for several reasons. First of all, it ruins your warranty <g>
and collectibility, but also, having the VCO as it currently exists is very
often desirable. So you don't want to mess with it; a better solution is to
have an off-board module for the times when you need an octave switching
function. By creating a bias voltage source with a rotary switch to
switch-in trimmer resistors, and then feed this bias voltage into the VCOs'
FM inputs, you can make this happen. Just need some trim pots on the
switching module to get the precise octave intervals on the VCOs.
Okay, that got me thinking.
Tuning VCOs to octaves is certainly something that is done often, especially
for that big fat analog sound. But what if you wanted to find ∗other∗ useful
intervals. What would these be? The Partial relationships between the
fundamental tone (VCO #1, let's say) and "harmonic #n" (played by VCO #2),
etc.! Reinforcing harmonically-related partials is very cool. After all,
that's why tuning to ∗octaves∗ is so cool---the octave is a strong harmonic
relationship. But it's not the ∗only∗ one. There are very sophisticated
harmonic tunings possible. A musical 'twelfth,' for example, because the
twelfth reinforces the octave+fifth which is the third Partial (second
harmonic). [See, I knew this would bore those not 'into' musical acoustic
physics. But therein lies the basis of all we do as synthesists, consciously
or not.]
AAAs some of you may have noticed from all my bitching about "partials" on
the list, > I'm really interested in working with the harmonic series in
pseudo additive synthesis fashion. With a module like this, these very
important relationships could be easily dialed in. So first you would tune
your VCOs to unison (for simplicity let's say there are only two VCOs) with
this Harmonic Biasing module set at the null position, #1. > At position 1,
there is no bias, so VCO #2 > is at unison with VCO #1. Switching to
position #2 would put VCO#2 up one > octave (the first harmonic). The next
position is octave-fifth (2nd > harmonic), then 2-octaves, 2-octaves-third,
two-octaves-fifth.... etc, up to > as many harmonics as you have switch
positions and trim pots for.
[Using terminology like "twelfth" and "two-octaves-fifth" is technically
erroneous, but it gets the idea across. Due to compromises made in Equal
Temperament, as you go up the harmonic series you start to drift from the
"notes on the piano." This module should be tuned to the true partial
relationships, not their Equal Tempered counterparts. The intervals need to
be f/2, f/3, f/4...&c.]
(Please hold your criticism until the end... I'm building this idea in your
minds one step at a time.)
> This ability to easily dial in exact partials on secondary VCOs, to me,
would be awesome. Most people detune oscillators by octaves to > get that
big fat "organ" sound, but the really cool way to tune VCOs is to >
emphasize the real positions of other "natural" harmonics.
>
(> In fact, there is a little known property of acoustic physics called >
"sub-harmonics." I don't mean sub-octaves. But a sound can induce a larger >
object to resonate at frequencies LOWER than its fundamental by exciting a >
higher harmonic of the resonant properties of the second object. It is very
> subtle. Henry Cowell wrote about this in his book "New Musical Resources"
back in 1911. I think Walter Piston even addresses it in one of is famous
books on harmony and/or orchestration. The first sub-harmonic of C is a
lower F. The sounding board of a > piano actually adds these sub-harmonics
in VERY subtle amounts. It would be > neat to be able to play with these
easily.
>
>
> Now, to make this ∗more∗ of a nightmare project...
>
>
> What if, instead of using rotary switches for selecting the offset
relationship, you use pots with smooth, internal > electronic switching? As
you turned the pot, the voltage would jump the VCO's pitch in > quantized,
harmonic steps. The first benefit is that you don't need a 64 position >
rotary to be able to get 64 harmonics on the dial, but the ∗BEST∗ reason for
> this is... <tah-dah!> ∗Voltage control!!∗ A CV could change the quantized
> biasing as well as the front panel control. What could you do with this?
You could have a CV, like an LFO, force a VCO to > sweep the harmonic
series! This is an effect people try to get with Serge > waveshapers,
high-resonance bandpass filters, etc. but you could have two > VCOs tracking
each other with the second one sweeping its frequency, but > always its
pitch is a true "natural" harmonic of the first VCO! Or an ∗EG∗ could sweep
it---on a percussive envelope, the EG would push the second VCO up to some
higher harmonic of the first VCO, and then VCO #2 would fall to some other,
still harmonic relationship. This is similar to what happens in nature with
pitched percussive instruments, but can be just different enough to be
really wild.
>
This to me is a VERY exciting idea. With the second VCO producing a sine
wave and the above effect being patched up, I can only imagine that it would
sound like the first VCO's signal was being split and the side-chain was
going through an impossibly tight bandpass filter that only emphasized one
single harmonic at a time. Of course, no real bandpass filter could be so
flexible and selective as to single out individual harmonics like this,
especially relatively high in the harmonic series where they get very tight.
And, of course, it need not be used just with VCOs. If instead of producing
a bias voltage for the VCO FM input it actually altered a 1V/OCT voltage
passing through it, it could be used with the VCLFO, the filters, anything
that works in the pitch domain. How would a filter sound with a peaky
resonance tracking the keyboard but also being swept with
harmonically-related control voltages? I bet it would be very cool, and at
the same time subtle because you are dealing with the natural harmonic
series, not just crude 'organ stop' relationships. The difference would be
the difference between equal temperament and just intonation, which is
simultaneously subtle and yet very dramatic!
(Wendy Carlos comments on "Secrets of Synthesis" that analog tends to
deteriorate "into that quasi-organ sound that most synthesizer work
eventually degenerates into." That's because most people tune square waves
to simple octaves and fifths to build a sound, which does start to sound the
same after a while. This need not be the case, and this module works
∗against∗ that organ-ish tendency.)
>
> I can imagine a 2u module. The four pots on the left side set the harmonic
> offsets for four VCOs. The four pots on the right are attenuators for the
incoming CVs to modulate them. Four > jacks are CV ins, and 4 are the biased
outs.
> >
Okay, next.
The weak link is the bank of trim pots, one per harmonic relationship. Now,
I know NOTHING about > PIC processors, but I'm told they are cheap and
really useful. I'm wondering > if the DACs in these are good enough so that
the harmonic relationships > could be programmed into a PIC chip. That way
they can be exactly specified > and theoretically would not drift. And, of
course, no calibration would be needed---you could lose the 64 trim pots per
channel, a huge expense and pain in the ass. > > Several people on this list
are PIC gurus who may be able to answer this question. >
As the above idea evolved, it resembled less of the original octave switch
and more of a really unique, specific kind of quantizer. Maybe this function
can work its way into any quantizer designs that might be on the drawing
board out there. As a stand-alone module, it might be esoteric, but if this
could be a feature of another module, it would be more economical. It would
be neat if there was some kind of switch: in one position, it quantizes in
1/12 volt steps, like you would ∗expect∗ a quantizer to, and in other switch
positions, an effect like the above could be implemented. Because really,
what I'm describing is a quantizer for Just-Intoned tuning, if used that
way, which is an important compliment to equal-tempered tuning. If the
module were _uP_ based, I would think this would be a low-part
implementation. Just a little extra programming and a switch. (Am I
wrong...?)
I'm raving about this out loud because I'm not an engineer, and can't
implement this nut-bag idea on my own, otherwise I would just build it and
demonstrate it. I understand acoustic physics a hell of a lot better than I
understand the nitty-gritty of the electronics. I just wanted to see if
there was anyone out there who thinks this could be as cool as I do. Can I
get anyone to chant along with me...?
As you can tell, I don't use my modular just to do two-VCO leads over guitar
pads... I'm trying to use it to create worlds.... I have weird needs... :)
Thanks for indulging this rave, folks. My vocal chords hurt now, and I
wasn't even talking....