Building a new kind of theremin...

OregonJim wrote: [i]I'm curious, Gordon - what is it that you believe makes the task difficult? What is lacking in existing designs that makes them not good? By "good", do you mean somehow easier to play? [/i]

As I said previously, I am not an electronics expert. However, I have been following the theremin engineering threads here for five years and the general consensus amongst the skilled electronics guys is that it's darned hard to make a commercial quality theremin. This is reinforced by the number of designs that do not produce playable instruments.

An ideal instrument would have a linear pitch response over a useful number of octaves. By this I mean that if the player makes the same hand movement in any part of the field, it will produce the same interval change in pitch - for instance from closed fist to fingertips stretched towards the pitch rod might be an octave - a doubling in frequency. I believe the natural tendency of capacitative fields is to obey an inverse square law which is in contradiction of this.

From my background in computer programming I note that the theremin pitch circuit combines two different functions - those of frequency determination and waveform synthesis - in a single functional unit - which sets of a little alarm bell - my supposition is that it will be hard to adjust one functionality without affecting the other.

I also think that the various sounds of theremins and ondes martenots - both heterodyne synthesisers, are distinct from those synthesised by other technologies, and are a large part of the attraction of the theremin to me. I think part of it - for the theremin - is to do with FM synthesis - the waveform of a theremin is not fixed, but varies as the player's pitch hand modulates its frequency. To do this it samples the player's position very rapidly - at radio frequencies. I think a lot of information is lost when a theremin's audio is converted to a representation of pitch and volume alone, discarding timbral changes, making this blended functionality an essential aspect of theremin design.

[i]I cannot stretch my brain far enough to be able to call an electronic oscillator "organic"[/i]

I can. :-) Well, not quite. But I can call it cybernetic. Correct me if I am wrong, but I understand that the player's hands are acting as capacitor plates, making them - literally - part of the circuit - like the cybernetic organisms of science fiction - and also in the Norbert Weiner sense of being part of a feedback loop which includes the theremin, the loudspeaker and the thereminist who is constantly correcting pitch and volume based on audio feedback.

Well said, Gordon. I understand now what you mean by a "good" theremin. I also understand that it is not a theremin you are after at all, but an instrument which does not yet exist!

In order to fully overcome the physical and aural limitations that you describe would require modification of the "accepted" design to an extent that most people here would no longer classify it as a true theremin. But, I agree with you - it would be a nice instrument, indeed.

>But I can call it cybernetic

Yes, I do believe that is a better analogy than organic. I still don't understand the concept of "fully matured" in this context, though...

It's good to aim for an ideal.

Yes, I do understand that real world considerations demand a certain amount of compromise in a design.

You'll find we're open to new ideas - but not afraid to point out anticipated problems with them either. :-)


Here's the other thing. The theremin bridges the classical and electronic worlds - some of us here make electronic music and come from an electronic music background. And others come from a classical music background and think in terms of music usually intended for acoustic instruments. (This is not so silly, the theremin in classical music has all the expressiveness you would want from an orchestral lead instrument - in this sense it is as near acoustic as you can get using electronics.)

Our pitch field response requirements are not so different, one from the other. In other areas, we differ. I would like to see more development incorporating ideas from other synthesisers, whereas others just want the one good sound.

@Gordon: We managed at least to open somewhat the interface door between both worlds when adding the FX mod to your Etherwave...

The Plus Kit is another step to make the link between both worlds.

[i]"I also think that the various sounds of theremins and ondes martenots - both heterodyne synthesisers, are distinct from those synthesised by other technologies, and are a large part of the attraction of the theremin to me. I think part of it - for the theremin - is to do with FM synthesis - the waveform of a theremin is not fixed, but varies as the player's pitch hand modulates its frequency. To do this it samples the player's position very rapidly - at radio frequencies. I think a lot of information is lost when a theremin's audio is converted to a representation of pitch and volume alone, discarding timbral changes, making this blended functionality an essential aspect of the theremin." - Gordon[/i]

I agree - I believe that the single most important feature of theremin tone is the 'distortion' which occurs to the waveform as a function of frequency change.

However, this form of distortion also occurs if the theremin is used to generate control voltages which then drive a 'normal' non-heterodyning oscillator (VCO) - It should also be noted that there are highly regarded theremins which use VCO's, including the Moog 91 series AND the Ethervox!

As for how "simple" or "difficult" it is to design a high-end theremin.. Well, I have (to put it mildly!) found that designing a theremin which meets my standards and [b] can be put into production at a sensible price [/b] extremely challenging.. nay, liquidating! - However, making a one-off where one is able to tweek the circuits to taste and fitting a myriad of presets is not a problem, then the difficulty is greatly reduced... One can spend a month trimming a one-off.. but if one is spending a week trimming a theremin one wants to sell, then the cost of this time will price your theremin way out of the market.

Having played with both heterodyning and non-heterodyning theremins, I believe there is a quality one only gets from the heterodyning variants - non-heterodyning can sound close to "the real thing" - and a good player can make their performance adapt so that one does not notice any difference (When Peter plays the E-Vox, it sounds like a heterodyning theremin.. unless he is using MIDI mode) but I think that for less skillful playing (or much less - like mine), heterodyning sounds better than non-heterodyning.

It is possible to produce an additive heterodyning theremin voice as proposed in the first posting on this thread - It is complex, but, given enough tweeking time, does work.. Simply run phased-locked-loops on both the variable and reference oscillators, multiplying both of these frequencies by the same constant up to as high a frequency as you can manage.. then devide these frequencies digitally down to give the required intervals, then XOR these outputs and filter them to produce triangle/sine outputs at audio frequency.. I built an additive voice this way using Cypress CY8C29466 PSoC parts to provide the multipliers and dividers, and had 6 harmonics whos levels were adjustable - But the cost / complexity was not worth it - In order to get anything aproaching theremin emulation, I had to dynamically alter the mixture as a function of frequency and volume (one loses the natural variation in harmonic content as soon as one uses XOR's for heterodyning) - the complexity of doing this (having VCA's for each harmonic and control paths to these driven from a frequency - voltage converter and the volume CV) created a monster which I abandoned - I could see that I could never put this into production, and that it was not worth doing even if I could - Far better to distort the waveforms before mixing, and keep everything analogue, even if one has less control over the tones.

@FredM: Don't you think that PLL's would be a kind of overkill? Don't you think that classical frequency doubler or tripler stages (nonlinear amp stages with tuned load impedance) would do the job in this case? Just a thought from an old analogical brain...

FredM wrote: "(When Peter plays the E-Vox, it sounds like a heterodyning theremin.. unless he is using MIDI mode) but I think that for less skillful playing (or much less - like mine), heterodyning sounds better than non-heterodyning."

Fred, besides its MIDI functions, the Moog MIDI Ethervox theremin has two voices. VOICE ONE is produced through the traditional heterodyne process, while VOICE TWO is a VCO that is designed to be used for playback (through the MIDI IN port) after the Ethervox has been recorded to a MIDI sequencer.

VOICE TWO responds to all the same timbre and register settings as VOICE ONE, and for all intents & purposes it is identical. It is designed for playback only, although it can be played the same way heterodyne VOICE ONE is played. Since a theremin recorded to a MIDI sequencer sends only volume and pitchbend information (there are no "note on", "note off" commands), the recorded data cannot be played back on any device except the Ethervox (VOICE TWO) itself.

There is one exception - VOICE TWO can be triggered by the Haken Continuum fingerboard but it's a tricky setup and it can be unpredictable.

I don't use VOICE TWO on the Ethervox, and I do not record to a sequencer. Everything I do is done live in real time directly to a digital recorder. When you hear/see me playing the Ethervox theremin, it is always heterodyne VOICE ONE (unless I am using the Ethervox in its MIDI mode to trigger some other device).

The SERIES 91 theremins are VCO devices, not heterodyne instruments. As you withdraw your hand from the pitch antenna, they never go to zero beats. There is always a sound...putt....putt....putt......

This discussion continues here (http://www.thereminworld.com/forum.asp?cmd=p&T=5097&F=3)