Moog Theremini!

Fred,

All really good points. At the heart of the matter is how you perceive the Theremini. Is it supposed to be a replacement for the Theremin (I don't think Moog is taking this position) or an adjunt to it and an acceptable replacement in certain venues (say a rock concert - the sound systems there are so poor by the time the mass of sound gets to your ears, I doubt most people could even tell the difference of a Theremini vs theremin (or care) things are so loud)? I really see the Theremini as a very viable tool for venues like this as it can produce a variety of other sounds. I don't at all dislike the Theremini (I just have issues with its current implementation - which Moog is working on - and which as you point out from just he theremini perspective can only go so far). 

My question - can I get an accpetable sound out of the Theremini by maybe adjusting the spectrum to better match the Etherwave (which I can certainly do with the tools at my disposal)  so I can use it for a practice instrument and maybe to easily plug in MIDI-controlled Theremin sounds in compositions I write for the instrument for mockups (my goal in addition to using it as a synthesider for Animoog sounds). Certainly MIDI keyboards have included a portamento wheel forever (so the general public is quite used to accepting pitch gliding on virtually any MIDI instrument - though it may not be an analog glide). If the theremini can be controlled through MIDI, this would add a very important use for it (as a tool for composers to test out how a piece they write might sounds - as I expect few of them would be good enough to play their own compositions - me included at this point).

The sinusoidal waveforms on the Etherwave are in fact constantly changing as you state as I can see from snapshoting at different times. I believe they do too on the Theremini (I need to check that out). Not sure about how the Theremini's spectral content changes with pitch. When I collect this data I'll throw it all into a paper I'll put on the web somewhere. 

Do you know if TW maintains a collection of .WAV files for different theremins? That would be a boon to looking at getting an optimal sound out of some instruments (for example what can I do to get a more RCA sound out of an Etherwave if that is my goal?). Ideally I'd like to see say plots of A220 and A440 at least (likely some dynamic examples would be good too). I wonder if people in the community would be willing to record sustained tones on different instruments and put it in a database (if one does not exist).

Rich

Hi Rich,

Yeah - I agree with your comments and it would be unfair to expect the theremini to be high end for its price.. My questions were more about the audio aspects of digital voicing for theremins in general, rather than specific to the theremini.

"Do you know if TW maintains a collection of .WAV files for different theremins? That would be a boon to looking at getting an optimal sound out of some instruments (for example what can I do to get a more RCA sound out of an Etherwave if that is my goal?). Ideally I'd like to see say plots of A220 and A440 at least (likely some dynamic examples would be good too). I wonder if people in the community would be willing to record sustained tones on different instruments and put it in a database (if one does not exist)." - Rich

IMO, this is one of the biggest scandals in the whole theremin community - Absolutely NO database of samples exists! - Classic instruments will be lost eventually because such a database has never been created let alone maintained, and there is no resource that developers can use, all we have is recordings (mostly with accompaniment) from which we can painstakingly extract samples for analysis without knowing the conditions under which the recordings were obtained.

I have begged and pleaded, contacted companies involved in putting sample collections together, and offered to collate and make samples freely available - but no one is interested.

We need samples from Lev's theremins (and those from other early pioneers) recorded from the instruments and, where possible, recorded to a high standard from quality microphones with the theremins played via their original amplifiers and speakers. We need sustained notes across the whole register, we need sweeps both fast and slow, we need volume modulations at different depths and speeds - in short, we need a systematic scientific catalog from which, at some future time, it is possible to recreate the instruments using digital sample engines..

Even if its not possible to produce such instruments yet, at some future time it will be possible to fully implement every nuance of the theremin sound with such technology - But the way things are going, there may be no original instruments left to provide the samples! - As equipment ages, the sound changes - Change a tube and the sound changes - We can never get exactly the same sound ever - but we need at least some samples close enough.

I wonder if I could get sufficient crowd funding to fund a collection? - travel the world and employ thereminists to play the instruments in the required manner, and produce a studio collection.

Fred.

 http://www.thereminworld.com/Forums/T/28807/on-my-knees---begging-for-samples-of-levs-theremins---please

http://www.thereminworld.com/Forums/T/24552/sound-samples-in-theremin-gallery

Fred,

Thanks as always. That's too bad. Its such a simple thing to do today. I can't tell you the number of samples I have - almost every instrument imaginable - and more and more companies are coming out with sample sets each week. Hard to believe the theremin community is not on this bandwagon to preserve these wonderful instruments. I wonder if there are some copyright issues with this for current instruments (though those are not the ones that are really at the heart of the discussion).

Also my assumption I can play the Theremini through MIDI is apparently wrong (at least at this point). I was hoping I could use it as a composition tool (I guess I'll just have to do it the hard way and practice more!).

But the Theremini waveforms are dynamic (according to my scope plugin) and appear to be changing quite like the Etherwave (though I haven't done any exhaustive analysis there).

And in the end perhaps, assuming the Theremini improves and gets a more mainstream following due to it multi-functional identity, sample companies may see there is commercial value in producing a Theremin sample set using classic instruments.

Rich

Interesting - someone is developing a patch editor for the theremini ahead of Moog. This is an opportunity for thereminists to be involved in its development.

Some points - I'm pretty sure this will not address linearity or latency issues. It looks like it requires an iPad. If you do participate, please consider suggesting an arpeggiator, based on the midi transpose command.

http://www.synthtopia.com/content/2014/08/27/ipad-patch-editor-for-the-theremini-theremidi-now-in-beta/

wow, three days before it was under construction and now it's beta!  quick, someone more serious should contact kai aras  and explain the needs occured here.

dani

Back home, back to the Theremini.  I stuck my cardboard Theremin pitch field measuring stick on it and took some data:

The above sums up the Theremini pitch experience for me.  It isn't super linear anywhere, but is somewhat linear seeming in the near field (0 to 0.25 meters) with the far field pretty much going to hell.  Who wants to be chained to playing within 10" of the antenna?

Above: The sensitivity seems to drop off roughly linearly with distance - not good.

Above: The CV goes from 0 to max (5V or 10V settable via the menu).  It hits max at the nearest pitch calibration hand location, and hits zero at the farthest calibration hand location.  Curve appears the same as the log of the frequency.

Excel spreadsheet with the data and details of how I captured it: http://www.mediafire.com/download/dui8yr59ftdfjys/Theremini_linearity_cv_2014-08-27.xls

MP3 I used to get pitch data: http://www.mediafire.com/download/t81327qhbr9nll4/Theremini_2014-08-27.mp3

[EDIT 1] Just stuck a scope probe on the pitch antenna through a 1pF cap and see ~750mV p-p.  The scope probe is 8pF, so the voltage swing at the antenna is ~6.75V p-p.  Measuring directly I see 4.8V p-p, which I'm more inclined to believe as the oscillator logic is likely 5V.

[EDIT 2] Oddly, touching the pitch antenna makes the pitch drop to the minimum.  The latency associated with this action is quite obvious, with the settling time 1/2 second or more.

[EDIT 3] Subjective: The Theremini is fun to fart around with for maybe 15 minutes, but once you've exhausted the presets (count 'em: 32) you're left with the uninspiringly non-linear and lethargic pitch response.  The synth engine is the most interesting feature, it's a crying shame that there's no real-time control over any of the parameters.  The UI could have been much better thought out, the menus are clunky, and the whole thing has a dashed together, designed by clueless committee vibe.

Hi Dewster,

Re pitch CV - Can you confirm / refute that the CV actually tracks pitch?

As in, if you set a 5V output, and select 5 octaves coverage (field), do you get a change of 1V for every audio octave change?

If one then changes to 10V output, do you get 2V change for every octave change? and if one has 5V output and spans say 7 octaves, do you get 5/7V / Octave ?

In just curios if the relationship 10V/[octave Span] and 5V/[Octave span] is true, and whether this fraction actually tracks the audio from the theremini.

Fred.

Hi Fred,

The way it seems to work is that frequency and CV are driven by a single number that is scaled and offset separately for both.  The number is also either operated on by exp2 for the frequency or log2 for the CV (but not both).  The calibration routine sets the physical limits of the pitch field, while separate menu options set the upper pitch point, lower pitch point, and CV max (the lower point of which is always 0V).

To get 1V/octave (5V / 60 notes) I set the pitch limits to C1 and C6, and the CV max to 5V.  With this setup I'm seeing 5V@C6, 4V@C5, 3V@C4, 2V@C3, 1V@C2, and 0V@C1.

To get 1.2V/octave (5V / 50 notes) I set the pitch limits to C1 and D5.  The pitch and CV seem to track.

It's interesting to look at the waveforms for the various notes in the MP3 I posted.  Highest frequencies are a sine, which segues through lopsided sine, double hump, and an EKG kind of spike at the lowest frequencies.  I suppose this could be filtering of a single wavetable at work.

===========

The user is prodded at power-up to perform the calibration routine, though this reminder goes away after a few seconds and the Theremini appears to then use the operating points from the last calibration.  (Boot time is also only a few seconds, thank god.)  There seems to be some thermal drift, and as usual it shows up most in the pitch far field. 

Thanks Dewster,

So yes, my hypothesis was correct, and my shifting / scaling circuit should work.. Ok, I still think its the most asinine implementation of CV ever dreamed up, but at least it should be possible to fix - perhaps with software if the numbers for the CV can be independently multiplied and offset prior to the D/A. (1V/Octave would require multiplying the number by a constant determined by the note coverage, and then have a number added corresponding to the lowest note - as the numbers going to the D/A are linear (have already been exponentiated) this should be quite simple to scale on the 10V output setting)

Fred.

Futzing around with the pitch correction while watching the tuner, the algorithm seems to be based on variable thresholding from the center of the current note.  Setting the knob to the center or 1/2 way point gives thresholds located 1/4 from the left and right edges of the tuner (limits of the current note).  Playing within the thresholds (i.e. around the note center) the pitch responds normally and you can apply vibrato.  Exceed the threshold and the pitch jumps to the next note at the opposite threshold (though there is some smoothing involved).

For example, if the "Scale" is set to "Chromatic" and the knob is set to the center and playing a C, when the hand moves closer to the pitch antenna, the pitch tracks the hand until the tuner hits the 3/4 point, whereupon the pitch jumps to C# with the tuner at the 1/4 point.  Higher settings of the knob give more narrowly spaced thresholds with lower "gain" when between the thresholds to space things out.

Moog Inc. seems to be selling the pitch correction as a useful aid to stay on pitch.  But due to the way it is implemented it strikes me as much more of an effect (and a useful one) but ironically rather counter-productive when it comes to staying on pitch.  Higher setting of pitch correction feel less controllable and less comfortable, as you can't tell how near a threshold you are until you hit one.  Lower settings (<1/3) change the behavior so slightly the value seems dubious.  And to have an entire knob and two switches devoted exclusively to it with an otherwise rudimentary / spartan UI seems wrong.

===========

Lowering the pitch range does seem to increase playability, but good luck developing an aerial fingering technique within the highly non-linear pitch field.  This isn't the precision Theremin you're looking for. </jedi_mind_trick>

===========

I hope no one thinks I'm disparaging the Theremini in order to pump up my own (as yet nonexistent) solutions.  With slightly different oscillators, the simple linearization techniques from the UNO, and deeper menus / UI, the Theremini could have totally dominated everything else out there.  It's kind of depressing to see such a lack of imagination / engineering coming from Moog.