Let's Design and Build a (mostly) Digital Theremin!

Designing a good digital Theremin can be a tricky business.  With all the pitch stepping / sticking digital Theremins can display, it would be nice to have some definitive way to test them, rather than the current imprecise "hand waving" exercises.  If I were a Theremin reviewer I'd probably take the time to build a motorized "hand on a stick":

Increase the speed of the motor (reddish blob) until the response drops 3dB and you know the bandwidth of the pitch side.  Slow the motor way down and listen / look for non-smooth response. 

something like this? : http://www.synthgear.com/2009/diy/lev-the-theremin-playing-robot

"something like this?"  - xtheremin8

No, you want smooth, cyclic movement, the more sine-like the better.

<PS - Real sorry! - I have just realized you are looking for "sticky pitch" and not looking to plot linearity! - So all that follows is OT - Yeah, a sine would be fine for what you are seeking! ... My head is full of linearity issues at the moment, LOL  ;->

"No, you want smooth, cyclic movement, the more sine-like the better." - Dewster

Not sure I understand or agree..

IMO, one wants a linear decrease in distance of "hand" to antenna, and a linear increase in distance on the 'pitch falling' cycle - A triangle wave.

To emulate something like 'correct' playing, one wants the 'arm' to outstretch (like punching straight) not swinging, and one wants a conductive 'arm' not an isolated hand.

But this is one of the problems - there is no "standard form" - Some people play with a swinging action, and there is a great difference in linearity profile for different playing styles ('non linear' theremins can be played more easily, IMO, and behave more linearly, if the player adopts more arm swinging and uses arm-antenna angles to 'dampen' the capacitive change)

In many ways, I think an android of approximate human proportions, would give the most realistic response.. the "Lev" robot is utterly useless for this task though. (it primarily uses 'arm' antenna angle, which is probably the most linear playing mode with a non-linear theremin)

In order to assess a theremins "linearity" one must assess based on the most common configuration / style that includes the player - take a real player out, and one must replace them with an accurate capacitive model which changes its capacitive relationship to the antenna / theremin / ground in the same way that the player would.

And that isn't easy! ;-)

Fred.

one idea:

sorry about my silly comment. but maybe that would make a good, real expensive solution for linearity test-sites:

http://www.techno-isel.com/LMS/ZF1BeltActuatorVideo.htm 

wouldn't that also reveal the digital jumps? 

i think, to get a good eliptic movement you need some cams, cranks and shafts : http://www.robives.com/mechs

(now that link wasn't entirely meant to be serious.) but explains mechanisms.

"If I were a Theremin reviewer I'd probably take the time to build a motorized "hand on a stick":"  i love that one. 

Hey, Xtheremin - that ZF1 looks great! ... I wonder if they do a non-metal / non-conductive version ;-)

But the trouble is.... Spend that money on equipment for a test lab where there are enough people willing to pay to get product tested, and you may stay solvent.. But there aren't enough people with theremins or long range capacitive sensors they want tested to even touch your interest payments!

I actually think that a basic body / shoulder / upper arm / elbow / arm / hand construction which could be locked into a fixed 'punching' movement and have one motor at the shoulder would do the job and probably cost about £400 and take a week to cobble together..

But how do you make that money back? Compare theremins on the market and show that yours is more linear and responsive than the competition? - That's the only way I can see.... But even that probably wont boost your sales enough to justify the expense and effort, because most folks aren't bothered or wont understand, and because the testing wouldn't be 'independent' so wouldn't be trusted.

What one really needs is an android clone of Lydia Kavina ;-) .. One needs the feedback every step of the way really, and cant really be knocking on Lydia's door every week to get her to test every tweak you make - one reserves such visits for when you are sure you've got it right!

Oh hell - the toy theremin market is so much easier! - Why the hell didn't I go with the flow...

Fred.

hi fred,

shure, that's far too much hitech stuff for a small market. or even moog. but it shure looks cool and comes in different sizes.  but only to  proof top linarity? or to explain the price.  but something that does the sweeps of the hand movement for you, so you have your hands free for other work, could be a cool simple thing. there are some projects with laser-cutted wood gears and cranks, that could do the movement. but if one can weld...what do you need as a simple "dummy-load", to compensate a body/arm/hand?  a earthed limb of metal? we experimented once with a minimum theremin (pitch only) tuned for it's maximum range and tried all different things approaching the antanna to see what influences : a pluged in guitarpickup on a stick.r but it was a miracoulous device back then anyway, (and was surprised that it worked without problems.) there is that cool sheet on this site somewhere, with the different linearities of different theremins. i think for a elliptic movement of the fake hand you need a groove in the desired form and size,made of non conductiv material or a cutout into thick plywood.(you won't use it for ever, do you?) then something that can follow that groove, maybe with a ballbearing or wheels,  a transmission arm, (like a pantograph) and a excentric disc in the same shape as the groove, but proporzionally smaller, mounted on the motor axis. so the pantographic arm would transmit the small ellipse on the motor side and a big eliptic move on the antenna side would be the result. how stable that should be built depends on how small one can construct the dummy.  but i still think that a continuous back and forth would do the trick and could be done easier than ellipses. i like dewsters inspiring idea a lot and wonder: what would jean tinguely have done with that?

"sorry about my silly comment. but maybe that would make a good, real expensive solution for linearity test-sites..."  - xtheremin8

The stepper motor might possibly confound things.  I think linearity is something a reviewer can test given their own human arm and hand and a frequency indicator of some sort.  To get numbers for latency requires some kind of rapid cyclic movement with adjustable and smooth speed.

Hi Xtheremin,

To me, mechanical construction is a dark art! But this is the kind of thing that could probably do the job: (if correctly drawn / sized and built ;-)

One would want the 'body' to be at the centre of the motor, and the 'hand' to go from this position forwards about 60cm, some sort of mechanism to restrict the arm to horizontal axis, and be able to just rotate the motor and vary its speed.

getting more sophisticated, one would want an output monitoring horizontal position quite accurately,  and ideally feed this back to an analogue controller so that the motor drive would be in a closed loop to keep motion linear.

But this is becoming a hijack! ;-) and I am utterly incompetent on these matters, so this is where I stop on this matter! ;-)

Any such test kit, IMO, needs to cover 3 things - Linearity, latency, and field irregularities / "sticky" pitches - The real use would be for development and comparison purposes - players can assess the linearity approximately, and developers can get a 'feel' for this without mechanical aids, but to do actual comparisons one would need a  repeatable process in a fixed environment (one would need a 'clean room' for the kit, a room where nothing changed the capacitive environment, and where humans were excluded or shielded while tests were running ;-)

Fred

 "what do you need as a simple "dummy-load", to compensate a body/arm/hand?" - Xteremin

LOL ;-) .. For a "dummy load" all you need is a variable capacitor! .. Forget mechanics, slap a bunch of varicap diodes and simple circuit where the antenna should be, drive this electronically .. Not simple, but a jell of a lot simpler than driving some mechanical monstrosity!

The only reason I can see for going mechanical is to simulate how a player interacts with the antenna - I see absolutely no point in robots or mechanical junk being controlled by electronics to 'play' the theremin, to me this is utterly pointless.

lol, maybe  i worked for too long with stonecutting machineries. ( loved these  monsters  to cut 4 ton blocks in slices )

and i also don't need no "lev". i would rather go for a kitten . ;-)

mechanics are not so dark art, its more geometric art.

sorry for the o.t.-ness. and the best for the digital thermin.

dani