MOTM

As I dust my brain from time to time I sweep out things cluttering my 
neurons, to make room for newer ideas.  Here are few ideas for modules that 
I'll probably never have time to follow up with, but still think have value:


1) Vocal-Tract Filter: (Was reminded of this with the recent post of the 
vocal-formants reference.)  The gist of this idea would have an input 
connected to a set of VC filters comprising a responding filter bank.  There 
would be a single knob (and CV input) to morph the filter bank response 
between the various vocal formants. 1-2 on the knob might be "uuu", 3-4 might 
be "aahh", 5-6 would be "oooh", and so on.  There could be another knob (and 
attending CV input) to select the frequency base for the formants: baritone 
versus soprano versus tenor, and so on.  Then a third knob (and attending CV 
input) would be an emphasis control that varied the intensity of the effect.  
And of course there would a mix knob that adjusted the wet/dry ratios.  I 
think that would be a useful module.

This module could be realized in a variety of ways.  If I were doing it, I 
would use something like 4, MOTM-420-like filter sections reconfigured to 
provide a bandpass response and voltage controlled "Q".  All 4 filters would 
be parallel connected on the input side and mixed on the outputs through a 
quad VCA chip.  The "magic" would be performed with a PIC processor that has 
A/D inputs.  By some empirical study, one could probably come up with a state 
table that depicts the required Q, resonant frequency, and mix level for each 
filter section, to yield the desired formant at the selected 
bass/tenor/soprano settings.  The emphasis control would cause the PIC to 
increase the "Q" of certain filter sections to highlight that portion of the 
formant.  By putting a PIC in-between the control inputs and the filters 
themselves, many desired responses could be synthesized: All filter functions 
would be under the auspices of the PIC.  Add a few toggle switches and some 
additional code in the PIC, and one could select between a "vocal-tract 
filter" and a number of other filter arrangements.  This could be built 
rather inexpensively, too.  The truly difficult portions would be lines of 
code inside a PIC.


2) JH Scanner Module: The "scanner" described on Juergen Haible's website 
would make a dandy module by itself.  There would be four or eight inputs 
that can be "morphed" between, under voltage control.  Maybe input 1 is your 
buzzy, unfiltered sawtooth from the primary 300 VCO.  Then input 2 might be a 
hard-sync'd second 300 VCO pulse output.  The third input might be from a 440 
tracking filter output that is swept in a standard "beeoooowww" arrangement.  
The fourth input might be from your 110 ring modulator.  With a single 
control voltage, the lone output (or output pair) could morph between the 
various combinations of mixed inputs at the mercy of an envelope generator.  
I think that, that would be a killer module and Juergen has already done it 
(more or less)!  Using a quad VCA chip coupled with a clever control circuit 
is nothing short of brilliant.  Sounds like a 500-series module to me.


3) Universal Synthesis Module: This module could process audio in unlimited 
ways... literally.  Base a MOTM-format module on a DSP "Evaluation Board" 
from TI or National Semiconductor.  (All these boards can have DSP algorithms 
downloaded from a PC.)  Since a DSP is the ultimate in sound mangling 
devices, it is a natural.  The idea is that a cottage business could be 
created developing downloadable functional bits for this module.  Maybe even 
create a subscription service.  Initially, the module would be equipped with 
various reverbs, phasers, flangers, Leslie speaker simulators, and so on, as 
the standard fare.  Downloaded code chunks could increase the repertoire of 
effects possible.  The additional algorithms would be saved in flash memory 
(very cheap these days).  There could even be a front panel accessible socket 
to allow the insertion of memory cards like those used on digital cameras.  
They could store a ton of algorithms.  Maybe add a MIDI interface and have a 
playback-sampler kind of oscillator - sort of a 300/MiniWave on steroids.  
Add a jack for footswitch control and an LED display, and you'd have 
something useful in a live situation.

Food for thought or ramblings of a madman?  What's the difference?

Cheers!
Tom Farrand