Moog Theremini!

"Question: would a raspberry pi have the processing power to do the necessary audio synthesis etc? Just wondering." - GordonC

The ARM CPU on the Raspberry Pi is quite old and about as fast as a Pentium 3 with 300 MHz.

The Beaglebone Black mini-computer for $45 would be a better choice.

Inputs: The Raspberry has 0 analog inputs vs. the BBB with 7 analog inputs
CPU: The BBB's CPU is about 2.5x faster due to more modern ARM architecture
Storage: The BBB has 2 GB on board so you don't need to buy an SD card

GordonC wrote:  Here's another question - how much will it reduce etherwave sales? Probably this is why it won't have a heterodyne voice.

Well, it seems to me that since the Theremini is to cost less than the Etherwave, and since the Theremini a far more complex item, in every regard to its internal design/circuitry, that something has to give as to the price of the Etherwave or the Theremini. 

I think there are several factors which influence pricing of the EW and / vs the Theremini.. The order I give here is probably wrong, but I think the issues probably relevant:

1.) The EW is overpriced for what it is. Profit is high but sales quantity quite low.

2.) The Theremini is priced to a target market, and I strongly suspect its price is cut to the bone with as small a profit margin as  can be allowed.. with the plan being that the volume of sales will reduce the manufacturing costs and therebye increase this profit margin.

3.) The only way that the theremini target price could be achieved and a good profit made would be if large numbers were manufactured and sold - I suspect its reduction in cost due to (implemented or planned) far larger production runs than have ever been done for any theremin, and the resulting massive cost reduction which comes with this..

4.) If (as I suspect, and I think would be Moogs prediction) sales of theremini are large, then Moog will make a big profit on each, even at the low price - loss of EW sales will be irrelevant - in fact, if the theremini sells well, and moog follows business "dictates", it wouldnt surprise me if the EW was discontinued.

5.) If theremini sales are not as good as is needed to sustain the low price and get healthy profit, the theremini price will go up or it may be discontinued - But I personally think this is the least likely scenario.

None of the above should have been reason to not spend perhaps $5 more on essentials like an acceptable loop and pitch preview - $5 would perhaps have put $15 on the retail price, not enough to move it away from its target price significantly.

Price of electronic instruments is not a simple thing to compute or understand - it doesnt relate in any linear way to complexity.. The biggest costs are often the unseen ones - the time it takes to produce, the quality (and therebye expence) of the personell required to set it up, the test equipment required, the burn-in / conditioning / quality control time.. That sort of thing..

And it is these type of things, unseen by the purchaser, which will often determine the quality of what they are buying.. One can be SURE that a low cost instrument will have less time spent on insuring its quality than a GOOD expensive instrument, but sadly you cannot be sure that all expensive instruments are good instruments.

The other things which influence cost and quality are determined earlier, in the planning and design phase - the standards set then.. If, for example, it is deemed that 20ms latency is acceptable, this will impact greatly and reduce the cost of the product compared to one designed for maximum 5ms latency - it may be more costly to produce a "simple" EW than to produce a complex digital theremin which has long latency and/or poor linearity.

Fred.

"Here's another question - how much will it reduce etherwave sales? Probably this is why it won't have a heterodyne voice. " - GordonC

Massively, I suspect..

But I think thats not the reason for lack of heterodyning voice in the theremini.. I think that the digital topology of the theremini probably means that even though heterodyning is used as the "sensor" engine, what it produces (in terms of pitch) probably bears no direct relationship to the pitch coming from the digital voice.

When I made that suggestion I was thinking in terms of "my" way of doing things - Discovering a bit more about the theremini I no longer think a heterodyning voice was a real option - it could have been an option, but this would have needed to be part of the initial plan - they could have had both for minimal extra cost if they'd talked to me or read my posts and understood their implications! ;-) LOL

I believe internal competition can be a big driver as well.  In the digital piano market it seems manufacturers intentionally hobble new technology so that it won't blow their older technology higher end models out of the water.  We own the Yamaha P105, and another $5 or so spent on Flash would have made it really stand out (in comparison to other DPs, which are mostly horribly lame) but it's a low end model, can't have it out performing higher priced models.

I also think there is a "function follows form" thing currently going on, where clueless overpaid style people are called in for a day or so to spitball, overriding the engineers, and what you see is what you get.

But Fred nailed it with the latency observation.  Literally anyone can build a low-cost Theremin given lots of latency.  I guess we'll see.

Fred wrote: The Theremini is a digital synthesiser connected to a capacitive sensing front-end which outputs MIDI and adjustable pitch corrected data.. Thats what it is, and thats all it is! - Many times here we have "debated" what a theremin "is" - And on most counts, based on these discussions, the theremini is not a theremin..

Fred, I think you'll find that the Theremini is a true heterodyne theremin (Steve Dunnington says so in a couple of the introductory NAMM videos) as well as a digital synthesizer (like the Ethervox). 

As for the question of what is and what is not a theremin, it's a theremin if you say it's a theremin. Most people these days seem to feel that the word "theremin" applies more to a sound than to a specific device. It might be a 1929 RCA or a touch screen app, if you think sounds like a theremin, IT'S A THEREMIN.

"Fred, I think you'll find that the Theremini is a true heterodyne theremin (Steve Dunnington says so in a couple of the introductory NAMM videos) as well as a digital synthesizer (like the Ethervox)."  - coalport

These videos show Steve talking about it and demoing it:

http://www.youtube.com/watch?v=ClGnPikoeSY

1:30 - "This is an analog Theremin controlling a digital sound engine..."

2:50 - "There are presets, and because this is a sound engine based on Animoog, we can have different sounds, classic Theremin sounds like we were listening to..."

http://www.youtube.com/watch?v=Ke5nIq4qwok

1:10 - "So what you hear is actually a wavetable..."

My take-away on what Steve says is that the "classic Theremin sounds" are synthesized.  I take them at their word that heterodyning is going on, but I think it's only for pitch data capture, not voice generation.  The tendency I think is to push anything as "analog" because it implies old-school, retro, organic, etc. even if it is just feeding a frequency counter.

The latency looks pretty good in that first video!

I'm wondering how linear it is, but the variable sensitivity is a huge deal IMO.  I'd like to see the setup steps demoed.

Re: Peter's Turing Test for theremins (if it sounds like a theremin...) On most of the presets this does not sound much like a theremin. It sounds like a synthesiser. (If that is even a meaningful statement - synthesisers can sound like pretty much anything, including theremins.)

Gibarian - so the Beaglebone Black has the same sort of wallop as a top of the range desktop from 1999, or a three year old smartphone. Wow! It would certainly be neat if there was one inside the theremini, what with it being open source and all.

Yes, I know, wishful thinking.

"Fred, I think you'll find that the Theremini is a true heterodyne theremin (Steve Dunnington says so in a couple of the introductory NAMM videos) as well as a digital synthesizer (like the Ethervox). " - Coalport

I interpret what Steve Dunnington says quite differently - I think he says that the voice is not analogue or heterodyning.

I think reference to "heterodyning" at the sensor side is a detail about technology which is completely irrelevant to the matter, and perhaps misleading.. If the heterodyning is used simply to produce a digital number for digital computing / synthesis to use (which I am 100% sure it is) then, as I said before, it doesnt pass the criterion of a "theremin" that was generally "agreed" here before. (not that this means anything)

With the E-Vox, there was a single heterodyning voice, which was output as audio AND used as the basis for deriving "numbers" - both for a voltage controlled synthesiser "theremin voice"  (which also took incoming MIDI) and for the conversion to digital "numbers" (for MIDI output) {I never got as deeply into the E-Vox as I had hoped I would, and never fully confirmed anything - but it seemed that CV generation was central to its operation.. In fact, I cannot even confirm that there was any "normal" heterodyning voice - to me it looked like all the voices were possibly voltage controlled synthesis - but as I say, I really dont know}

All I do know with almost absolute certainty is that the theremini isnt in any way the same as the E-Vox - It doesnt have a heterodyning voice, and it doesnt have an analogue voice, it only has a digital wavetable voice of some kind.

None of this matters at all IMO - I have always felt that the primary requirement for an instrument to qualify as a "theremin" are capacitive sensing employed to control pitch and volume from two sensors.. In this regard the theremini certainly is a theremin.

If its a good theremin then it may clarify or redefine what a "theremin" is - and may answer the question about whether heterodyning (for the audio) has any special place.

I am really interested in this - in particular, I think that heterodyning produces musical distortion of the waveform as pitch changes**, and that analogue engines in general are capable of this - But I am not sure that wavetable / digital voices are capable of this unless specifically designed to replicate it. I strongly doubt that this distortion will be present / heard from the theremini - Will discerning people notice? (as in, will they prefer the sound from an analogue / heterodyning voice, without knowing why ?)

Or has it all just been "cork sniffing" ?

Fred.

**It is my suspicion that this distortion COULD be a big key to why some people (myself included) prefer musical sounds produced by analogue oscillators, than the sonds produced by most digital synths..

If we take a constant pitch, we can produce any waveform digitally that we can create with analogue - one cannot hear any difference, because there is no difference.. It makes no matter whether the  sound is produced by heterodyning, or analogue subtractive synthesis, or additive synthesis, or wavetable, or FM..

But as soon as we vary the frequency of the waveform (particularly if we vary it rapidly - as with fast glides or portamento or vibrato) the technology used to produce the sound can give different results to some other method of producing the sounds..

Wavetable engines tend to update their pitch on a cycle-by-cycle basis - so if one had a wavetable sine wave, shifting from say 440Hz to 440.5Hz over say 3ms, one would get a complete sine cycle at 440Hz (~2.3ms) followed by another complete sine cycle with a frequency somewhere between 440 and 440.5Hz, then (if pitch is held at 440.5Hz) one would get complete sines continuing at 440.5Hz... Each waveform cycle will be exactly the same shape, a sine! there will be no cycle-to-cycle distortion, but there will be an unnatural sequence of tidy sine "samples" packed next to each other following the pitch data..

Analogue (and heterodyning) engines do things differently - If one has a 440Hz sine from an oscillator, it will only be a pure sine while the frequency is constant - as soon as the frequency is altered, the cycle in action will immediately respond, so the waveform will distort in order to shift the frequency to its new value - while gliding between two pitches, there will never be any time that the waveform will be a pure sine, and the faster the glide (and/or greater the shift), the less pure the waveform will be.

The above applies to all waveforms, not just sines - and the more complex the waveform, the more complex the distortion.

I think its this mechanism which makes analogue synths and analogue theremins much more interesting to listen to probably more than anything else.. I have one digital synth which does distort in the same way, but many which dont.. Wavetable synths I have come across (and I havent played with any synths made in the last 10 years, or perhaps longer - The last digital synth I aquired was the Virus B rack -  so I may be well out of date! - most are Roland and Yamaha WT stuff) are all horrible when it comes to pitch shifting, some taking several cycles to update..

It should be noted that its not the type of component ("digital" "logic" "mixed signal" "analogue") which determines whether pitch 'glide' introduces the distortion I am talking about, its the way they are configured.. The E-Pro uses a mixed signal (logic level) topology / mixer that generates a triangle wave, but produces this distortion just as an analogue heterodyning mixer does - The mixer I am most proud of does the same - "heterodynes" and shapes the waveforms using logic constructs, but all these waveforms (ramp, triangle, pulse) undergo the same pitch shift related distortion that any purely analogue heterodyning mixer does..

 I may be wrong, but in my view this is essential to producing a sound that is anything like a "real" theremin (or for that matter a "real" emulation of an analogue synthesiser when applying pitc bend, pitch modulation, or portamento)

I think we have evolved in a world where, when pitches change continuously from one to another, nature dictates that the waveform MUST distort - this is true for all natural / acoustic sound - stretch a string, and the waveform distorts.. Until electronic (and particularly sampling) technology, there was no possibility of things being otherwise.. Rapidly sequenced samples can fool us into hearing a continuous pitch change - but the auditary cues are wrong - I suspect that our musically sophisticated "inner animal" notices, even if "we" dont!

;-)

*

(to me, "pure analogue" is where there are no constraints on the signal, either in amplitude or time.. "mixed signal" is where only input amplitude is constrained to logic levels, time is not constrained or quantized, "digital" is when both amplitude and time are constrained, amplitude constrained to logic levels, time constrained to quantized intervals. In my view there mixed signal is, for almost all synthesis applications, as good as / equivalent to analogue - there are some cases, like when oscillators in a theremin are distorting due to "pulling" close to null, that analogue will give marginally different mixer output harmonics to mixed signal - but this is easy to replicate and even enhance with mixed signal topology.)

Fred: ".....I think reference to "heterodyning" at the sensor side is a detail about technology which is completely irrelevant to the matter, and perhaps misleading.. "

I think you're right, Fred. I was misled!