Build Project: Dewster's D-Lev Digital Theremin

What a beautiful work!!
Absolutely great!

Thank you, gerd.  I appreciate it!

Here the front panel which supports most of the D-lev's electronics is separated from the antennas and AFE drivers by about 50cm, and it seems to be working fine with no twitchy spots.  I haven't tested how long the cables can be before problems start showing up, but even this distance is adequate for testing without having to reassemble the panel to the case after every change.

I think it would be interesting to have separate stands for the antennas and drivers with the main electronics located off to the side or behind the player.  I'm still thinking of a setup where I can position the antennas on each side of my recliner.

The fit and finish is just stunning. I'm so anxious to hear more of it being played as well.  If it sounds as good as it looks (no doubt it will) this will become extremely sought after.
Congratulations!

I don't think I would be putting this effort into the the project if I wasn't 100% behind Eric's design.  He has put a good number of years into this and knows his stuff when it comes to the hardware and software. Whenever I have requested a feature he has in nearly all cases come back with new software within days, so I keep asking and he keeps producing.  And all the while he keeps adding capabilities while streamlining the user interface.

The first thing I notice about the D-Lev is how open and comfortable it is to play.  The sensitivity, linearity, and bias adjustments of the pitch and volume fields allow you to personalize the theremin to your playing style.  The pitch field can be so large that is possible to sit comfortably back from the theremin and reach even the highest octaves without having to extend your arm close to the antenna where you lose stability and where the pitch field tends to get compressed on many theremins.  If you prefer to have octave intervals compressed at high or low pitches because you have played for years that way and have learned to compensate, you can still have that. But if you want nearly perfectly spaced octave intervals, you can have that too.  The same goes for the volume side; you can shape the response to your preference.

The second major part of the D-Lev is the synthesizer, and while this is not as feature laden as you would find in a modern keyboard or rack synthesizer, it has many of the same parts and probably exceeds anything that has ever been offered on a theremin. It has the capacity for 127 user presets and another 127 "factory" presets.

Here are the screens of the current user interface (note that these may change):

(Eric:  please comment if you have additions or corrections - I'm winging it on some of these features)

Main screen that appears on power up:

Bass and treble controls, plus control of levels for master volume, pitch preview, oscillator, and noise:

The D-Lev has a sophisticated pitch preview that controls harmonic content (can make low pitches easier to hear), octave, pmod (can raise or lower volume with pitch) and vmod (allows pitch preview volume to vary with main volume):

Pitch correction, if desired:

Envelope generator can alter the volume response spatially and over time:

Noise generator for special effects and for adding realism such as bowing noise with stringed instruments or breath noise in vocals:

Noise filter with control of filter type (LPF, BP, etc), corner or center frequency for the type of filter chosen, resonance (Q) at frequency, pitch and/or volume modulation of the filter frequency, and mix:

The next two screens control the source oscillators.  "hmult" is 0 for a sine wave, 1 for a sawtooth-ish spectrum, 2 for odd harmonics, and higher numbers produce various other waveforms. Spread and offset between the oscillators can produce phasing or beat effects:

If other than sine waves are selected in the previous screen, harmonic content can be adjusted here:

The oscillator filter has the same features as the previous noise filter:

This is new to me but I believe that this allows configuration of the pivot points for all the the pmod and vmod modifiers:
[Edit] Applies to all pmod and vmod modifiers for frequency but not volume... see followup post for details.

The next 4 screens "0_FORM" through "3_FORM" control parameters for 8 formant filters.  These are very important elements for synthesizing realistic instrument and vocal sounds.  The formant frequencies are modifiable by both volume and pitch using pmod and vmod.  The resonance of the formants can be adjusted in pairs:

The inharmonic resonator can generate frequencies that are not harmonically related to the sources.  This can be used for many instrument and vocal sounds:

And now we get to the system parameters.  These are global settings and are stored in preset 0.  This screen controls the volume linearity, sensitivity, and bias:

and for the pitch side:

Here are the visual tuner controls:

and finally a few utility settings.  PV allows reversal of the pitch and volume sides for left handed players:

I should add that the "pmod" and "vmod" parameters that show up on so many of the screens above are two of the most important and powerful modifiers offered.  They allow pitch or volume to modify other parameters such as formant center frequencies or pitch preview volume.

As one example, a series of three or four formant filters can create a realistic vocal sound over a limited pitch range, but if you go too far with fixed formants the voice can sound like singing mice.  Applying "pmod" can allow the formant frequencies to change with pitch and more realistically emulate the natural formant changes that occur in a human voice.  Likewise, applying "vmod" can change the formant frequencies slightly with volume, again adding subtle realism.

Another example of how "pmod" and"vmod" can come into play is for pitch preview.  Identifying pitch through an earpiece can difficult at low frequencies if the preview audio has low harmonic content, yet setting the harmonic content high enough to identify low pitches will make the higher pitches excessively buzzy.  "pmod" can increase the harmonic content at low pitches and roll it off for higher pitches.  Similarly, "vmod" can roll off the volume with pitch to level out the volume perception over the theremin's full pitch range.  "vmod" can also be used to reverse the sensing of the pitch preview audio, allowing the player to hear it only when the main audio volume drops out.

Yes, I have been following Eric's thread for years. Unfortunately most of its way over my head. I think it imust be very rewarding to see his years of dedication realized in such a beautiful and relatable way.

Thanks to both of you for making this happen. 

Thanks for your comments.  Eric and I have had a few discussions about what would be the best way to get this into the hands of some other players.  It is important to have a certain degree of stability in the hardware and software and some documentation to put in the hands of new users.   It is equally important to have a decent array of preset voices available to the player to be able to experience what the D-Lev offers without having to delve into the more sophisticated configuration capabilities. Eric has spent a ton of time supporting me on this project, and it is impractical for him to do that for more than one or two users, hence the need for user-friendly documentation and a library of presets.

Another thing that we have touched on is the possibility of having a loaner model to offer.  I don't want to go too far here because we haven't had any recent discussions on the subject, but I could see building up a model for players like yourself that might like to try it without any real commitments (except perhaps a deposit while on loan).  If you are in Minnesota this might be even easier.  But I'm getting ahead of where I should be.  Stay tuned on this, though.  I have always planned to build more than one of these for myself, in more original and creative enclosures than my copycat EWPro cabinet.  I could see putting something together that would allow players to try it out without the expense of any fancy wood enclosure.

"(Eric:  please comment if you have additions or corrections - I'm winging it on some of these features)"  = pitts8rh

Nice pictures and explanations! 

Not a correction, more of an expansion: the "PV_FMOD" UI page knobs only apply to the pmod and vmod settings associated with filter frequencies (i.e. the oscillator filter, the noise filter, and the formant filters).  "*_lo" sets the onset threshold (hand location), "*_hi" sets the operating range (so the modulation happens from lo to lo+hi and is constant beyond these limits), and the three "*_en" knobs are used to individually enable this limit processing for the oscillator filter, noise filter, and formant filters.  If not enabled, the modulation pivot point is -24dB on the volume axis and  262Hz on the pitch axis (and these are the pivot points for all non-filter pmod/vmod knobs).  The main use here is to have human vocal vowels change from one to another over a limited range of volume or pitch range, without sounding weird outside of that range.

Also: Roger, the D-Lev wouldn't be as far along at this point without your creative input, and all the effort you've put into this project, I really appreciate it!

"Thanks to both of you for making this happen."  - Johnthom

Thanks for your long-term interest!

Very nice and professionally looking.

Is it possible to publish some performance audio samples?

Thanks, Buggins.  The only recordings I have at this time are these two impromptu videos that I made with the prototype in the shop on New Year's Eve while I was looking at waveforms.  I was just playing around and decided to upload Auld Lang Syne in time before midnight, and as usual I regretted it after listening again.  With that pre-apology, here they are (the videos are also embedded in this thread a few pages back):

Violin:      https://www.youtube.com/watch?v=Sg1wpmmOokE

Male Voice:  https://www.youtube.com/watch?v=Xs8qnl7dWhM

The newly built "Pro" cabinet is still a few weeks from being finished (more boards to do), so I won't be making any recordings with that for a while.  The prototype is functionally the same though.