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

Not to pile on you Touchless, but Theremin seems to have been aware of the downsides if his creations.  He slapped on switchable filters to alter the harmonics, and likely would have been thrilled to add arbitrary waveforms and dynamic filtering.  He also had a crude tuner IIRC. 

Literally anything implemented in older technology can be improved with more modern technology, but engineer-musicians are a vanishingly small lot, and the corporate versions of us aren't given the time to do it right as it wouldn't be profitable in the exceedingly short run on which all modern business seems to operate.  So the buying public gets one older design by Bob (not his best IMHO), a seemingly dashed out Theremini by his replacement(s), assorted hobbyist fare, and the output of talented amateurs / semi-emeritus with a bit of time and/or money on their hands.  There's no telling what you might get from this motley crew, or when you might get it (if at all).

The field of study is surprisingly broad and deep.  I've enjoyed it immensely, but it can be an intellectual quicksand for curious cats.  You have to understand the fundamentals as well as or better than most to find the value in the conventional wisdom (or to challenge it) and then make any headway at all.  Even the simple copying of existing designs is fraught with peril.

Thank god for the free-flow of ideas and info on the web and particularly here at TW, without which I would still be wrestling with the basics (and with the voodoo).  If anything, the biggest challenge to implementing a modern Theremin seems to be the DSP side of things, where the cards are played too close to the chest and the learning curve is quite steep.

Not to pile on you Touchless, but Theremin seems to have been aware of the downsides if his creations. - dewster

You can’t pile on me as I am Touchless. The original theremin was perfect by the dictionary definition of the word, it is the birth of electronics. If you have religious upbringing, as I see in your words below, it may have given you a misguided understanding of the meaning of perfect.

Thank god for the free-flow of ideas and info on the web and particularly here at TW, without which I would still be wrestling with the basics (and with the voodoo). - dewster

Really? I don't think your are any closer to the goal today than you were a year ago or even your birth. To try and imitate perfection may be your downfall, in your future I see you tossing your hands up over theremins and walking away. You will ask yourself over and over - What did I miss?  A Voodoo prophecy.

I have no doubt you have amassed more knowledge than Leon Theremin, but you do not have the Vision!

T

"I don't think your are any closer to the goal today than you were a year ago or even your birth."  - Touchless

It probably seems that way if you've been reading this thread for a couple of years.  But there have been a lot of things to investigate and nail down, and Hive - which is a critical component - has consumed at least half of that.  I don't plan on dropping it all tomorrow, but manufacture of anything is a major, major hurdle.

Dewster,

That troll isnt worth answering! - I know, I take his bait and give explanations, but he comes back with the same snipes and nasty subversive bullshit again and again! his only agenda is to cause unpleasantness and discouragement.. Low life scum! Tell him to stick his crystal balls up his RS!  - Please feel free to use my trash can!

Fred.

I don't know exactly why (other than to understand them better) but I've been playing around with MASH DACs in simulation:

http://www.mediafire.com/download/cpzoa7l088bechf/MASH_DACs_2014-09-01.xls

Unlike other modulators, they're super simple and guaranteed stable regardless of input value and order.  The biggest caveat is they produce 2^N bits of output for order N, so they generally require a small DAC at the output.

Here's a nice treatment of them, in the context of several extremely interesting projects (click around at the top):

http://www.aholme.co.uk/Frac2/Mash.htm

[EDIT] Found some errors in the spreadsheet when simulating in SystemVerilog, so I updated it and the link above. 

There appear to be diminishing returns associated with higher order MASH modulators.  4th order doesn't seem to utilize the output levels at the limits very much.  Also, the output averaged signal is constrained to [0,1] but the output range is 2^N, so more dynamic range at the output is required for higher orders.

Sorry if you got distracted by this post - it was BS and ive deleted it.. My sinulator screwed up its digital - analogue interface.. When I actrually ran it on the PSoC I got triangels not ramps.. Should have seen it without building the damn thing - what the simulation gave was an obvious impossibility!

If you never saw the deleted post, thats good ;-) Just dont want anyone wasting time or being misled by it..

:-(

Working on the verilog for the digital audio output.  SPDIF is a bit of a pig when it comes to bandwidth, it's encoded bi-phase mark so takes a ~2x clock right off the bat, and they've included so much other useless info in there that it takes another 2x to transfer 16 bit PCM.  No wonder it doesn't travel well over video coax, with a 6.144 MHz clock required for 48 kHz stereo PCM.

Guys,

Taking a break from practicing my Etherwave (after a while my ears get saturated) and am trying to go through some of this post (now my brain is saturated). Being a systems/software engineer (some may consider that an oxymoron) a considerable amount of this is beyond me but perhaps you can summarize how this thing is going to actually digitally model theremin sounds/waveforms? Is the goal to have a dial that selects different theremins or have some kind of software interface that lets you build a theremin sound from some building blocks you might provide and maybe shape it with some finely banded EQ or filtering? You going to use some kind of virtual modeling, sampling & filtering? From what I have seen, I have no doubt that Dr. Dewster can get the thing to work - but how are you going to make it sound great? What's the user interface gonna look like? This thing gonna have a MIDI port if it needs loading or configuration? Do you have a rough diagram of what you want it to look like?

Hope these are not stupid questions.

Rich

rkram53 said: "From what I have seen, I have no doubt that Dr. Dewster can get the thing to work - but how are you going to make it sound great? "

Now that is the million dollar question, after 10 years. Yes our first theremin contact was early 2005. I think his sound is wishful thinking inside his head. If the sound is not recognized as unique to the theremin, what is the point. If it sounds like a modern day synth we might as well play a keyboard. Interestingly pianos are where dew cut his teeth, does this cause bias?

Every successful theremin designer knows the theremin is about discovering what does not work when it comes to hands on practical construction. That should be many days of rude awakenings.

I like dew but I am fearful of being in the grave when he demonstrates his first results.

The authentic theremin sound is complex, a digitally generated ramp wave is fine for keyboard but a theremin?

Christopher

Rich,

The user interface I'm currently planning:

Tuner: 13 LEDs arranged in a 6 point star for notes, 8 LEDs arranged in a rough bar graph for octaves.

Display: Two 8 character LED displays.  One or both of the displays may be 14 segment to accommodate alpha.  Otherwise will be 7 segment.  The displays will be arranged thusly:

  8.8.8.8.8.8.8.8.

  8.8.8.8.8.8.8.8.

Control: Two rotary encoders (with integrated push button switch).  I'm debating whether to have one encoder on either side of the display, or both on one side, and whether or not to positionally associate the encoders with the displays:

  8.8.8.8.8.8.8.8.  O
  8.8.8.8.8.8.8.8.  O

  O  8.8.8.8.8.8.8.8.
  O  8.8.8.8.8.8.8.8.

  O  8.8.8.8.8.8.8.8.
     8.8.8.8.8.8.8.8.  O

      8.8.8.8.8.8.8.8.
  O                     O
      8.8.8.8.8.8.8.8.

etc.  Placing the encoders on the left would allow one to twiddle the knobs with the left hand while playing with the right, but it might feel strange for right handed users.  Placing them on either side would require both hands to manipulate things - it is an ambidextrous layout that could speed up the input of data.  Placing both on the right would likely feel most comfortable.  An extra button or two might be necessary for whatever, but I'm trying to keep the controls to a minimum.

The synthesis engine is kind of down the pike, but I primarily want it to do human vocal sounds well, so I'm thinking bright excitation with formant filtering.  Other types of sounds could certainly be generated from such a setup, but my goal isn't to replicate a DSP-based Moog type subtractive synth (this has been overdone IMO).

I'm not thinking MIDI at this point.  MIDI out would be an easy thing to add, but it doesn't interest me and doesn't seem to be a good fit with Theremins.  CV out would be a bit more complex, but that doesn't interest me much either.  Perhaps I'll change my mind on these things at some point.

=========

"Now that is the million dollar question, after 10 years. Yes our first theremin contact was early 2005."  - RS Theremin

I joined TW in February 2012.  Looking in my notebooks, my Theremin research started in earnest at that point.  Hive development has taken a huge chunk of that time (maybe 1/2) but I believe that's mostly settled for now.  I could have dashed something out years ago, but if I had I'd be spending my time supporting something mediocre rather than developing something more deserving of my time.  Not that I'm anywhere near the artisan he was, but I wonder if Stradivarius had people hectoring him to "just slap some varnish on that hunk of carved wood and get it out the door!"  The research I'm doing now is broadly applicable and so has much reuse ability for future electronic musical instruments of mine, so I'm treading slowly (and hopefully surely).  Does no one read Aesop anymore?