A "step" theremin? Also a few q's

"Watch the index finger of my left (volume control) hand."  - coalport

Peter, I watched, and I saw the finger - LOL - But I am none the wiser! ... What was that finger actually doing? is there some hidden (or at least not visible on the video) switch or actuator? - Or was the tiny change in capacitance caused by finger movement being picked up by the expression loop and converted to a MIDI "chord change" signal ?

-There I was thinking it was all from some box of electronic trickery, or a pre-programmed chord sequence - and thinking to myself "He doesnt need to do that - when he just plays the theremin and plays any other instrument to back it, he produces wonderful music .. why is he wasting his time on some horrible automated chorder" .. in fact, somehow, you were applying real-time control..!

Doesnt change my opinion on the piece / the "production" or whatever - But it does change my opinion on your performance of the piece!

but hey - LOL .. My opinion is just that, a musically illiterate geeks opinion ;-)

"*Sigh*  At what point did every geek with a soldering iron turn into Dirty Harry?" - Dewster

War has been a primary motivator for technological advancement - sad, but true I think.. I watched the whole video, and there were some almost funny things in it, like the warning that causing damage to the raw crystals was "sabotage"..

I wonder where technology would be today if there hadnt been huge diversion of public funds to R+D motivated by the war machine.. but I think the "At what point " has a simple answer - as soon as there were soldering irons! :-( - it may even be that the soldering iron was a direct result of demands from the "war machine".

And yes - those with technical skills are the ones needed to do this kind of work .. I dont think its so much that technical people are "inclined" to be "dirty harry's" as that they are almost forced into this role - registers are kept of people with technical skills, and when "needed" these people are drafted into the role demanded - or face the penalties.. On a day-to-day level, the coersion is slightly more subtle, but boils down to being "forcibly" drafted..... If you are technical - particularly electronics and related - more jobs are available in defence than anywhere else.. Refuse to work in this, and you can be unemployed even if highly skilled - The pressure to comply can be intense - particularly if your partner / family does not share your objections.

IMO, we would be far better off and far happier if technology had not advanced so quickly - I do not believe that the "advantage" that the war machine has brought in terms of technology has been worth it.. we would have advanced - but probably more slowly... And in the extra decades or centuries we might have matured enough that when we did have todays technology, we would be better placed to use it responsibly - rather than letting it make our world become unsuitable for humanity as we are doing.

Fred.

The MIDI Ethervox was designed with a memory for 100 user programs (nowhere near enough if you ask me). In CHROMATIC MODE you can program the instrument to send a MIDI stream to a receiver telling it which notes to play and in what register. Let's say I want to send an F major arpeggio spanning four octaves I can save that to program 001. That is followed by an A minor arpeggio with an added perfect 4th - I can save that to program 002, and so on....

I can do this until I have the complete accompaniment for whatever piece of music I want to play (up to 100 separate programs). The E'vox is provided with incremental +/- buttons that I can then use to scroll through my saved programs in real time, but I cannot spontaneously deviate from what I have put into the instrument. 

One of the design flaws in the E'vox, is that the increment +/- buttons are in the middle of the control panel, so it is necessary for either the pitch hand or the volume hand to leave its post in order to advance the chain of programs. These are the kinds of idiotic, totally impractical ideas that inventors come up with when they design instruments without consulting working musicians. They THINK they know what musicians want BUT THEY DON'T and there is nothing you can tell them to change their minds. The only thing to do is buy the instrument and then customize it so it can actually be used. Of course, this will invalidate your warranty but what the hell.

That's what I did.

I drilled the cabinet and installed a second set of increment buttons just next to the volume antenna, so I could change programs without having to compromise antenna control. What the index finger of my left hand is doing is simply advancing programs. I can decide at what moment I am going to play a particular MIDI stream, and how I am going to play it, but I cannot deviate from its preprogrammed parameters. So you are partly right - the instrument IS a "horrible automated chorder" but I can play those horribly automated chords in real time, and in any way I like. 

Before I stopped concertizing altogether I played the MIDI theremin with full symphony orchestra on more than one occasion, and if you have powerful, rock concert type bins on either side of the stage, the instrument is SPECTACULAR to say the least. There is only one problem. Your audience will have no idea what you are doing! You wave your magic hands in the air, and choruses of voices and harps and FX of all kinds fill the auditorium but most people in the audience will assume that they are hearing a CD or some sort of pre-recorded music, while you are making meaningless gestures in the air. 

It's bad enough with the traditional theremin. Even when you explain exactly what you are doing many people will not understand and will believe it is some kind of trick. With the MIDI theremin, things are complicated by the fact that you can produce multiple sounds from several MIDI receivers simultaneously and trigger them all with a simple gesture of your hand. 

The more sophisticated the technology, the less people are going to relate what they're hearing to what they're seeing.

It is worth mentioning that there is a fundamental mismatch between the theremin and MIDI. 

MIDI is based around the concept of a note event. For a keyboard this is easily defined - when a key is pressed or released, a note event is generated. With a theremin it is not so easily defined - a lot of it comes down to the performer's intent - was supposed to be a vibrato or a trill? a glissando or a portamento? a crescendo or a repeated note? And so on. And can we determine intent by mathematical analysis of the information available from the pitch and volume circuits by considering not only position but velocity and even acceleration?

So far we have only looked at simple pitch to MIDI conversion based on distance of the hand from the antennas. These are very limiting - they take away much of the expressive capability of the instrument.

There have been two serious examinations of how to impose MIDI on a theremin that go beyond simple "stepping" - the Moog Ethervox and the MIDI-Wave (aka the MCV1a) theremin to MIDI/CV interface. Both tell us that there is no single best answer to these questions, as they provide a variety of ways of determining note events and pitch bend for the performer to choose from according to which best fits the requirements of a particular performance. That's making the best of an imperfect solution.

My preference would be not for MIDI but for the more recent OSC protocol that more closely fits non-keyboard-based musical instruments.

Gordon wrote: there is a fundamental mismatch between the theremin and MIDI. .......MIDI is based around the concept of a note event.

The "mismatch" is not between the theremin and MIDI.  

The MIDI theremin transmits volume and pitchbend commands, and it does this very efficiently and precisely. MIDI receiver modules and keyboards, on the other hand, although they COULD be capable of reading and obeying these commands, are unable to execute them because the manufacturers of the modules have decided that this would entail an unjustifiable investment in a level of sophistication that would be of no use to the vast majority of consumers. 

It is what manufacturers are doing with MIDI technology that is based around the concept of a note event, not the MIDI technology itself.

Modules are designed to receive note on/note off commands because that's what people want. It is fashionable these days to denigrate MIDI but the fact is few users of the technology understand what it can really do, or know how to use it to full advantage. 

You can record your Ethervox MIDI theremin performance (in PITCH MOVE MODE) to a MIDI sequencer but, up until recently, the only device that could play it back was the Ethervox itself (via its MIDI IN port)! The advantage is that you can record a sloppy theremin performance to a sequencer, tidy it up on-screen with the software, and then play back a flawless "corrected" version of the piece. 

Just for the record, my Haken Continuum can play the MIDI Ethervox, and vice versa. 

In CHROMATIC MODE, the E'vox sends conventional note on/note off commands as well as volume control info, and you can use it, in real time, to play any commercial MIDI keyboard or module. 

"Just for the record, my Haken Continuum can play the MIDI Ethervox, and vice versa." - Coalport

I am interested - When a MIDI stream (pitch move mode - presumably also with streamed expression data from the loop) is recorded / played back, is there no discernable difference ? ..

To me, whilst you are correct that " It is fashionable these days to denigrate MIDI but the fact is few users of the technology understand what it can really do, or know how to use it to full advantage." - MIDI was (and still is) an elegant simple invention, and the MIDI protocol was IMO utterly brilliant.. BUT..

The MIDI hardware is based on data transfer rate of 31.25k baud - thats 31250 bits per second.. 10 bits are required for each "packet" and the maximum data each packet can carry is 7 bits.. So we have a maximum data stream which can provide a string of numbers between 0 and 128 (extremely low resolution, only good enough for note events and commands) at a rate of 3.25kHz... And 3.25kHz is well in the audio spectrum..

But for high resolution data (required for pitch) without any interleaved commands or volume data etc, two packets are required - reducing the effective data update rate to 1.56kHz

In reality, with volume data and command packets, even using sysx, you are not going to get any faster than about 800Hz.

Yes - the above data when integrated to smooth out the steps, is I think just fast enough - But only on a system specifically optimised for this purpose.

There is good justification from an engineering perspective, to improve MIDI .. At the most basic level, a simple increase in the data transfer rate while retaining the protocol would be enough to get rid of the potential headaches with streaming mode.. An upgrade of MIDI to say 300kb could easily be implemented with minor hardware change - and if the MIDI protocol was retained, the firmware developed for modules and controllers would not need to be reworked - it could simply drop in to new faster products.

I have not looked much at the MIDI replacements like OSC.. From what I have seen it is centred on ethernet type protocols etc which are a field I know almost nothing about - they have the potential to transfer vast quantities of data and stream audio etc... But we could have had a simple MIDI speed upgrade 10 years ago easily, and IMO this is all that is really needed.

Fred.

Sure you can fudge it by playing a single note of infinite duration and seven octaves of pitch bend and ignoring all off MIDI's note related facilities and accepting that in the real world this comes with a slew of technical difficulties, but isn't claiming that that proves there is no mismatch rather like claiming that a flamingo makes a good croquet mallet?

P.S. I note that all of your MIDI theremin videos use note events, rather than just pitch bend. I would be interested to hear how it sounds played through your Continuum just using pitch bend and volume.

(IMO) MIDI came out when electronic music was exploding and more egalitarian.  But it tends to commodify products, something manufacturers pretty much hate these days.  If they weren't in a sense required to implement it on certain products due to legacy I don't believe they would.

I agree with Fred that a speedup would be beneficial, particularly for non-piano type controllers and sound modules.  Though I would imagine the asthmatic processors they tend to use in electronic musical instruments would have difficulty keeping up.

Something bidirectional with at least the bandwidth of a stereo digital audio stream is required IMO (e.g. to stimulate synthetic resonant structures with real audio) but in this secretive and proprietary climate I can't imagine anyone implementing it and it really catching on.

(The creators of MIDI made a mistake with that control bit.  MIDI has multi-byte messages and operating modes, so that bit is just sucking up precious bandwidth and giving us oddly sized 7 bit info.  They also should have used common RS232 baud rates, and perhaps should have made the rate variable to somewhat future proof it.  Ideally they would have made the transmission AC and balanced in order to reduce impulse noise bleeding into the audio and to allow simple capacitive coupling rather than expensive optical isolators.)

"(The creators of MIDI made a mistake with that control bit.  MIDI has multi-byte messages and operating modes, so that bit is just sucking up precious bandwidth and giving us oddly sized 7 bit info.  They also should have used common RS232 baud rates, and perhaps should have made the rate variable to somewhat future proof it.  Ideally they would have made the transmission AC and balanced in order to reduce impulse noise bleeding into the audio and to allow simple capacitive coupling rather than expensive optical isolators.)" - Dewster

If MIDI had come to be in the last decade, I would agree with the above - But the complete draft of the MIDI specification certainly appeared before 1985 - I think it came at about 1980.. That was way back before anyone imagined anything like the technology we have available today.

taking the points one by one - (1) The command bit - yes, its a pain and limits BW.. But it is also a foolproof means of implementing assynchronous data transfer and ensuring a clear delineation of command bytes and data bytes.. (2) - Yes - standard RS232 with the ability to select the baud rate would have been better, and could have been done then - except I think it wasnt done because of speed limitations on available opto-isolators.. I think 31.25k is about as fast as one could realiably drive a standard opto back then - and even today, opto-isolators are quite slow (3) Opto-isolator.. I remember back at the time engineers, including myself, thinking that the current-loop / opto-isolator scheme was a poor choice - BUT - It did make the interconnection completely safe and nearly completely reliable - it eliminated any ground-loop problems - and retrospectively, I think it was a wise choice.

That was back in the days when the PCs available for music were things like the Yamaha MSX or Commodore 64, when, in order to save your data or load programs you connected a mono audio cassette player / recorder to your PC, and a simple program could take 5 minutes to load! - When even if you were lucky enough to own a recycled data recorder, you had to write your own drivers for it and implement the interface with RS232... Some years later things improved a bit with "cheap" PC clones - I remember buying an Amstrad 1512 for about £600, a 20M hard drive for about £200, and a Roland MPU401 MIDI interface for about £150, and this was high-end stuff! (I made more than enough money to pay for all that within a year by copying the MPU401 and selling boards to musicians).

It is also interesting to remember that MIDI was not the first digital interface for instruments - There was a parallel DCB from (I think it was) Roland and I think Yamaha had some experimental interfaces. Sequential got MIDI "out there" because, I think, it was truly open-source.. unlike the others who kept their interfaces for their own kit.

If you look at the full MIDI specification, MIDI 1.0, it is, in my view, a work of art. Everything relevant to control of musical instruments which was possible at that time, is catered for - The mere fact that MIDI is still being fitted to equipment made today is testimony IMO to its cleverness... How many interfaces / standards from the early 80's are still fitted as standard to anything? - PC's dont even have RS232 as standard these days, let alone a Centronics interface ;-)

And what motivation is there for manufacturers to accept a new "MIDI" scheme ? The vast majority of control is still keyboard based - And I doubt that there are enough thereminists or Continuum users to make incorporation of OSC or anything like that into standard synth modules worth doing..

But - there is USB .. Lots of recent keyboards have added a USB interface, down which they TX/RX MIDI data.. There is (AFAICS) no reason why this needs to be limited to 31kb - In fact, USB can operate as a multi-function interface - it can carry MIDI and Audio and other data simultaneously -

I suspect that there wont be any 'deliberate' move towards a new "MIDI replacement" - I suspect that there will be a messy hodge-podge of USB/MIDI implementations which may eventually evolve into a new standard.. I hope I am wrong!

"but isn't claiming that that proves there is no mismatch rather like claiming that a flamingo makes a good croquet mallet?" - GordonC

LOL ;-) .. IMO, its more like claiming that a croquet mallet makes a good flamingo!

Its not, IMO, that there is a "mismatch" - its more that something to fully transfer the data required for theremin interfacing really needs to fly - flamingo's are better at flying than croquet mallet's are!  ;-)

IMO, there is an interface almost perfectly suited to theremins - Its pre-MIDI, and available, and there has been a recent resurgence in its use, with many new instruments incorporating this interface.. Its analogue Voltage control! ;-)

Fred.

Fred asked: I am interested - When a MIDI stream (pitch move mode - presumably also with streamed expression data from the loop) is recorded / played back, is there no discernable difference ? ..

No, there is no discernible difference between the original Ethervox performance recorded to a MIDI sequencer, and the same performance played back through the Ethervox, provided you do not alter the original timbre and tone control settings. Like a MIDI synthesizer, you can play back your MIDI performance and change the sound as much as you want without altering the original recording in any way. All that is in the sequencer is pitchbend and volume information. 

The Ethervox has two voices. One of them is a true heterodyne sound (VOICE ONE) and the other is a MIDI controllable VCO (VOICE TWO). Dr. Bob used a similar voice for his SERIES 91 theremins but without the MIDI capability. When playing back from a MIDI sequencer via the MIDI IN port, the Ethervox uses only VOICE TWO. For all intents and purposes the two voices are indistinguishable.

"The Ethervox has two voices. One of them is a true heterodyne sound (VOICE ONE) and the other is a MIDI controllable VCO (VOICE TWO). Dr. Bob used a similar voice for his SERIES 91 theremins but without the MIDI capability. When playing back from a MIDI sequencer via the MIDI IN port, the Ethervox uses only VOICE TWO. For all intents and purposes the two voices are indistinguishable." - Coalport

Thanks for that information.. Interesting that "For all intents and purposes the two voices are indistinguishable." and that the MIDI resolution is inaudible.

There are two voices in the E-Vox, this I have seen from my brief encounter with it and study of photos the boards... As far as I can tell, both voices certainly have identical post mixer processing, which could account for what makes them "indistinguishable" - but, to be honest, I actually suspect that either both voices are heterodyning, or that neither are.. From my observations, it appears that the front-end circuitry outputs a CV with no audio generated - this CV drives the "real-time" voice (Voice 1) and MIDI interface, and the identical 2nd voice is driven from CV decoded from incoming MIDI.

The thing which makes me suspect the above most was the fault on the E-Vox I looked at - this was a fault on Voice 1, where there was severe drift - and it turned out AFAIK to be a fault on an exponential converter driving this voice... Now, it is just possible that this exponential converter was performin some other function related to linearization, or that I completely misunderstood everything (I was in extremely bad shape at the time, and never concluded the job).. I had the client test my hypothesis by cooling the exponential converters transistors, and advised him to contact the well-known theremin repair shop in France, as I was not able to have the theremin in my lab, and could not undertake any more off-site examination or repairs at the time - I never heard any more.. But exponential converter circuitry is usually employed to convert a linear voltage (1V/Octave usually) to an exponental current or voltage.

I suspect that there is no heterodyning voice in the E-Vox, or that both voices are voltage-controlled heterodyning (which is the technology I was developing at the time - so it may be that I am "superimposing" my ideas on the situation, or it may be that Bob, once again, was way ahead of me.. ;-)

Whatever - Voltage controlled heterodyning voices are, IMO, the best topology for theremins - BUT - If the E-Vox "playback" voice is not heterodyning, and its real-time voice is heterodyning - and you cannot tell the difference between the two..... Then I wonder if there is really any point in the extra effort and expense required to make a Voltage controlled heterodyning theremin rather than just using a standard audio VCO..

To be honest, I believe that I can hear a difference - to me, my voltage controlled heterodyning voice sounds better than a standard voltage controlled subtractive synthesis voice - But Bob was the master of voltage controlled subtractive synthesis, and probably crafted the voice to perfection.. We know that the 91 series had no heterodyning voice, but I dont think there are many complaints about its sound..

Fred.