Pitch to CV, and related.. TECHNICAL! Part 2.

Continued from here (http://www.thereminworld.com/forum.asp?cmd=p&T=3939&F=3&p=3).

Fred - what does the filter do? Smooth out vibrato and minor fluctuations?

Also, how much will the S+S add/subtract, multiply/divide by?

[i]"Fred - what does the filter do? Smooth out vibrato and minor fluctuations?

Also, how much will the S+S add/subtract, multiply/divide by?"[/i]

Gordon, with the components shown, S+S will:
multiply by x1 to x2
Subtract 0 to 5V

But the above can be set to almost anything by selecting a few different component values.

The filter causes changes in CV to be exponentially smoothed - this will (again, with the components shown) really just remove (attenuate) the effect of any rapid CV glitching - Filter has adjustable time constant from 1.2ms to 12ms with components given, could be taken up to 5ms to 550ms with more expensive opamp and component changes - simply by changing the value of R10 and C4.

With a longer filter, the effect of vibrato would be reduced (output voltage would eventually average around the vibrato centre voltage) - one would have a half second portamento ( with filter TC set to max) .. As this filter is a simple R*C, its change would be exponential - Pitch (voltage) would change rapidly at first, rate of change slowing as the target Pitch(voltage) was approached.. this is a pleasant portamento sound.

I have NOT built THIS circuit - but have built many almost identical - The only bit which has any potential problem is when a bigger filter time is required, as this will load the opamp output.. If long filter times (>200ms) are needed I would probably do some minor redesign. If anyone is interested they can email me with specification for what they want, and I could certainly build a custom unit for under £75.

You have some interesting ideas about CV delay.. This is quite easy to implement, the most difficult / expensive bit being the A/D and D/A .. Once one has this, you probably only need to sample the CV at about 1kHz, which means that storing 1s of data will only require 1k words (assuming 16 bit resolution) .. one does not need a DSP or anything fancy, just two pointers to this data area, one for inserting new data, one for extracting delayed data.. The whole job (with exception of A/D and D/A) could be done with a low cost MCU.

I am almost tempted to divert and do a quick "CV only" Theremin incorperating all the stuff discussed in this thread - But I cannot - I keep doing this and end up with loads of 'suspended' projects and nothing getting to market - right or wrong, I am sticking to my present direction (yes - I have now chosen a direction! ;) .. CV might be next... However, if anyone is interested in taking my designs (and a couple of pre-programmed PSoC's) and turning these into a 'ready for manufacture' prototype, I would be extremely interested -

[i]I am almost tempted to divert[/i]

No! Don't! Be strong! I'm a very limited market and I can wait until you're good and ready - then we can talk about a generalised CV processor (at a sample rate of one per millisecond you could do a lot with a low cost MCU and some RAM - it could certainly make an interesting module for the analogue modular fiends.)

[i]"only need to sample the CV at about 1kHz, which means that storing 1s of data will only require 1k words "[/i]

A little bit more complex.. One also needs to delay any other control signals - volume CV and gating being the ones which spring to mind.. Still possible, but more parts!

a CV delay box does sound quite interesting - what if it also did looping? it would be an interesting way to add a quasi "sequencer" functionality to the theremin!

[i]"a CV delay box does sound quite interesting - what if it also did looping? it would be an interesting way to add a quasi "sequencer" functionality to the theremin!"[/i]

First, I think, we need the "REAL" Theremin which gives a CV output that REALLY tracks the audio over it whole possible range (20Hz to 15kHz ?) without any glitching or 'fiddle factors' .. And perhaps include a quite simple CV delay for this.. But anything beyond this would not, in my opinion, make any sense (and certainly would not make commercial sense unless some rich musician put a LOT of money up front - which aint going to happen!)

As soon as one gets any more complex, the user interface becomes complex, and complexity detracts from playability -

Working on my SkyWave Theremin, this has become the most difficult area to resolve - I have realised that there is good reason for the small number of knobs on Theremins - Unlike synths, where the player occasionally has a spare hand, one does not have this luxury..

It is an INSANE instrument - it seems to be locked in the 30's by technical factors one only sees when (as a designer) one cannot turn back.. Oh, I do wish I had never got into this.. I wish I had'nt built that first damn prototype and got the 'bug' watching my kids punch the field and make it squeek in time with "Bob the Builder"...

AAAaaarrrgggggghhh!

Hmm, looping. Interesting. No overdubbing, of course - CV is strictly monophonic but... you could have a bank of paired loops ceach driving an osc/opamp pair, and switch recording from one to the next.

And... you can change the speed of a loop without affecting the pitch - even have the volume loop going at a different speed to the pitch loop.

And pitch shift is easy and inexpensive - it's just adding a voltage.

I think it could be possible to create new melodic lines from the original line that are non-obvious but musically related. (For instance by subtracting the pitch from a constant to invert the melody.) Algorithmic accompaniment.


[edit]Fred posted while I was typing.[/edit]

These ideas reflect some of the ideas I have used in my recordings by other means. It seems to me that, for some purposes, there are advantages to processing sounds encoded as control voltage pairs and translated them into an audio signal afterwards, than dealing with an audio signal directly. I agree that, as with any modular system excessive complexity is possible and to be avoided - we have all seen those photos of monstrous modular systems with patch cords everywhere, but in practice I feel that the restrained use of effects can be very effective, and the theremin is well suited to delays and pitch shifting - two things that are good in CV.

Spooky..
I posted a LONG sad letter, edited it to delete about 60% - And somehow you got your posting in-between my edits..

Gordon - Sometimes I feel like im being stalked :-{} But, as Slartybartfast said "Thats just normal paranoia - everyone in the galaxy suffers from that...!"

" you could have a bank of paired loops, each driving an osc/opamp pair and switch recording from one to the next."

"And... you can change the speed of a loop without affecting the pitch - [b]even have the volume loop going at a different speed to the pitch loop.[/b]
And pitch shift is easy and inexpensive - it's just adding a voltage.
I think it could be possible to create new melodic lines from the original line that are non-obvious but musically related. Algorithmic accompaniment."

NO! STOP! "Theyr coming to take me away ha ha he he ho ho - to the funny farm....."

I have that effect on people. :-)




[i]"They're trying to kill me," Yossarian told him calmly.
"No one's trying to kill you," Clevinger cried.
"Then why are they shooting at me?" Yossarian asked.
"They're shooting at everyone," Clevinger answered. "They're trying to kill everyone."
"And what difference does that make?"[/i]

On a more serious 'note' :)

One can buy D/A A/D interfaces for PC's .. I think they are even USB.. One could do all the delay / looping / sequencing in a laptop, assign hotkeys for presets etc..

D/A A/D spec would include the fact that they must be DC coupled, and probably 14 bit minimum resolution..

I do not think it would be commercially viable to design / build a device like this.. but using A/D D/A and a laptop, one can tinker about with the (fairly simple) software.