Theremin Development - (more) questions for thereminists.

Volume control:   (if not technical, ignore the blue italic text )

Something I played with a little in the past, and quite liked at the time, I am going to test again and perhaps include -

Normal theremin operation is that the volume gets softer as the hand approaches the loop - I find this difficult (and the reverse equally difficult) - one is controlling volume as a function of position - I am (with hand drums and velocity sensitive keyboards) much more in tune with controlling volume as a function of velocity - to me its just a lot more natural.

Taking the voltage from the volume 'antenna' circuit, and passing this into a 'change' detector (HPF) so that the faster the signal changes, the higher the output voltage / volume, one gets a velocity / volume relationship - feeding this voltage into a RC lowpass with a TC of perhaps 5ms (or perhaps adjustable, say 1ms to 25ms.. its not pitch latency), one could sustain an output by continuously moving the volume hand - perhaps similar to the behavior of Bowed string instruments. where the sound is only produced while the bow is moving.

I can see that for sustained notes, the above might be a pain - but for short or staccato notes, one could play the note by moving the hand towards the loop, and the following note by moving the hand away from the loop - so one could effectively beat ones hand in time to the music, and control each notes loudness by the speed of that movement, and have a single switch on the theremin to select between this 'delta mode' and normal operation.

What I DON'T know, is whether such a scheme would actually be playable - whether the arm movements would interfere with the players ability to maintain pitch... But seeing some thereminists (one at least ;-) managing to do great fret-less bass simulations on a theremin, I think it may be possible - and it may also make cello / violin 'emulation' more natural.

Again, although not strictly framed as a question, I would really love to hear thoughts on the above from any musician / thereminist ;-)

Fred.

 "perhaps similar to the behavior of Bowed string instruments. where the sound is only produced while the bow is moving". -- FredM

I tried to implement it in digital domain,  corny producing the differentiation operation on the data from the volume resonance circuit  (altitude  values) . The results is still unsatisfactory, because the adjacent (and even not so adjacent) values differed little from each other and the output signal turns out too quantized.

The averaging slightly improves the situation, but not so much (besides, it increases the response time).  Hope you will obtain the better results in analog  domain.

And, how about “hither-thither” mode? (the "volume off" position is somewhere near the antenna, and any moving the hand towards the antenna or opposite increases the volume). 

"I tried to implement it in digital domain" - Ilya

Interesting! - I had actually thought that digital implementation might be better (past experiments were analogue) - My new design has the volume 'antenna' voltage going into a fast 12 bit ADC.. Not sure if I will use this (the voltage going to the ADC could just as easily be used directly to control a VCA - and I may well remove the ADC and IDAC and free up some oins for other uses)

One problem I had with the analogue experiments was related to linearity - one needs the volume to follow a log-like law, and I was implementing this at the front-end to save the hassle of needing a log VCA. This log law however works against the "delta" mode, because the delta is different when the hand approaches the loop as when it moves away from the loop.. So I was thinking in terms of the front-end loop voltage being linear, and implementing the log function in the processor, so that a linear delta could be obtained when in delta mode, and this could then be applied to the log processing in the same way that the normal mode signal was... This should mean that movement in either direction (towards / away from the loop) should give the same velocity proportional loudness.

"And, how about “hither-thither” mode?"

Personally, I would hate this! (if by this you mean that going on the "wrong side" of the loop "null" position increased the volume) - To me, when one gets 'below' the loop 'null' there should be silence.

I am not going to spend a lot of time on this delta mode - I have implemented the 'hooks' for it in the PSoC, and will try it - but unless its significantly useful will probably discard it and use the pins for something else.

UPDATE ->

In putting the blocks together, and particularly looking at shifting some simple functions out of the PSoC to external (low cost) CMOS ICs to free up resources (looking at possibly implementing a low cost preset storage which reads the potentiometers - but this never worked out..) I saw a way of implementing a really simple register switching theremin - this would have simplified linearity control for pitch only, no span control (stuck at a span configured when manufactured, 4 to 6 octaves), and hugely simplified sound pallette - it would not need a PSoC, would be nearly conventional topology (not employ phase-locked-loops) and suffer from the same problems all non-PLL theremins suffer (I am still looking at using PLL on the volume oscillator / detector, because IMO the AM detection method is hugely inferior - but PLL adds to the cost).

It is real tempting to put a simple single-board "base-line" theremin together while doing the main job - as minimalistic as I can manage.. This would be for less demanding players:

 Oh - for anyone wondering what "HF Cut" is all about - Its a cheap way to get arround an expensive problem - This is actually a low-pass filter, and one is changing the cut-off frequency... But the trouble is that if one wants the cut-off frequency to increase as one turns the knob clockwise (the "natural" way) one needs reverse log potentiometers, and these are a pain to get and expensive - oh, Alpha will do them as a custom order but you need to buy 1000+ to get the price down to the same as you can buy standard log pots for in quantity of 10 or even less.... So turning the log pot CW actually lowers the cut-off frequency, and the only way I could think of to make this rational was to call it "HF Cut".

Digital implementations wont have this problem because reversing the rotation can be managed easily by the firmware.

Fred.