Moog Theremini!

"Thus, non heterodyning instruments may have earned a bad reputation, not for being non heterodyning, but for their embarrassing uniformity of the timbre(s). So, I came to the conclusion that there is a lot to do for "us theremin designers" to give the instruments a truly musical sound like a human voice which changes completely its character when moving from the chest voice to the head voice. And simple wave table synthesis is IMHO just the wrong way."  - Thierry

Yes.  Heterodyning + 100% analog processing can give all kinds of variation, the trick I suppose is corralling the design to somehow generate mostly pleasing variation, and to do this economically.

Analog synthesis, and worse digital synthesis, need to have this variation introduced, which isn't always straightforward.  I'm thinking here of guitar multi-effects pedals, where only recently IMO have the processors had the power to do wave distortion and complex filtering tasks necessary to simulate a guitar amp & speaker cabinet.  But with digital, once you've got it, you've got it!  And so you can crank limitless copies out fairly painlessly. 

Clara in a can, as it were.

OT - Not about the Theremini..

"to give the instruments a truly musical sound like a human voice which changes completely its character when moving from the chest voice to the head voice. " - Thierry

The peculiar irony is that you are completely right on this IMO, but also (IMO) a little wrong! ;-)

When it comes to vocal charactaristics, if singing "ahh" in a bass register, then moving up and singing "ahh" in a higher register, the "ahh" formants remain nearly constant (and this is true for whatever one is singing - each vowel or whatever has its own formant pattern) - But it is because they stay constant that the waveform (harmonic mix) changes as the pitch changes..

To achieve a human voice sound, one doesnt really need to alter the excitation waveform much, if at all - but to make a realistic voice which isnt just doing an unnaturally constant vowel, one needs to change the vowel filters.

I see voice simulation as an entirely seperate process to waveform generation etc -I think that dynamic change of the waveform is needed more for other sounds than for voacl sounds - but that one can have both simultaneously.. You can change the waveform and feed this to a formant filter, the output waveform can then have mixable degrees of "direct" or "formant processed" qualities. biggest problem is the number of knobs required to enable all the mixing possibilities! ;-)

 I suspect the reasons why Lev's theremins had more vocal quality than modern theremins comes down to the technology used - in particular the audio transformers - This is where, IMO, you got it right when we were discussing the Lev mixer - I think I grossly underestimated their importance -

The mixer was also hugely important in producing an excitation signal that is extremely close to the larynx waveforms - Trying to work out what was changing the raw waveform from the mixer into the rich "vocal" waveform, I am now almost sure that it can only be the transformers and related resonances that are doing this..And thats good news if its true - because its quite easy to implement formant filters using gyrators etc.. and simulation is starting to yield audio results that I am starting to like ;-)

Fred.

 I dont yet know the mechanism which moves the formants (they only move a little - and the formants are a bit strange - not exactly human at all .. this may be the key - or it may just be an analysis error due to the low quality samples)

And oh yes - I completely agree that wavetable is the wrong way to go - but I suspect you know this by now! ;-)

Dewster wrote: "A full-blown OS makes all kinds of wonderful things possible and greatly eases development, but it seems you're always fighting them for access to real time."

Linux does seem like overkill. Something more old-school seems appropriate for such a tiny little computer.

I found this curiosity. If my programming skills didn't have thirty years of rust on them I'd have ordered one just for the fun of it. :-)

https://sites.google.com/site/libby8dev/fignition

OT - Not about the Theremini..

"PS-I don't doubt Fred's suggestions except that I had cobbled my setup together before I knew what I was doing/had Fred's insight!" - RoyP

LOL ;-) Thanks Roy, but doubting me wouldnt be unwise even if you did! - Question everything, Doubt everyone, believe nothing, be prepared - in this damn world, that seems like the best course!

It does surprise me though, the wiring used on some theremins - One theremin I repaired (this was a special - I aint getting into Thierrys territory ;-) had a thick cable to the antenna running over a steel plate - the bottom of the theremin was more sensitive than its antenna! .. It had no EQ, and had awful sensitivity, and tuning was maxed out to compensate for the capacitance... I was able to replace the wire with a thin one and fit a neon, and still needed to bend the wire towards the plate so that I got the tuning to the centre slug position ;-) And yeah - it got better!

I have also modified a couple of cheap DIY theremins to have the option of swapping antennas over from LH to RH - and can do this simply (no switching required) by connecting them with thin wire I use to wind coils on IFTs - these wires are running horizontal the whole length of the theremin, but because they are so thin they cause no coupling problems (I put heat shrink tubing over the wires so they can be seen and to re-inforce them)

Fred.

Also, this guy says that what you've got on the Open Theremin is not Wavetable Synthesis in anything but an academic sense, and then goes on to describe WS as something that addresses the concerns raised here as it consists of using multiple related waveforms for a single sound, and morphing through them in response to changes in pitch, or volume, or over time.

http://www.propellerheads.se/blog/tutorials/thor-demystified-11-the-wavetable-oscillator-part-1/

(scoot down to sound #9 at the bottom of the page)

That is the difference between the theremini and the Open Theremin.

We haven't any proof yet that the theremini does it "the better way"... I suspect the bean counters at Moog music to have chosen a not too complex solution to maintain the benefits...

"I suspect the reasons why Lev's theremins had more vocal quality than modern theremins comes down to the technology used - in particular the audio transformers - This is where, IMO, you got it right when we were discussing the Lev mixer - I think I grossly underestimated their importance -"

It feels good to read that, Fred, thank you! In the time your reaction was completely different. If I remember well you even made some comments which I perceived as rather hostile and which were considerably reducing my motivation to write anything here in the forums during a long time.

"and then goes on to describe WS as something that addresses the concerns raised here as it consists of using multiple related waveforms for a single sound, and morphing through them in response to changes in pitch, or volume, or over time." - Gordon

Yeah - the more layers there are, and the cleverer the interpolation between waves as a function of pitch or amplitude or some other controller is, the greater the possibility for an interesting sound or accurate emulation of the target instrument.

I havent got into the OT, and exactly what the theremini will give in terms of sound is a but unknown right now - but its certainly true that with wavetable synthesis one could have a different set of harmonics tor every frequency point in the field..

Theres so much non theremini clutter on this thread that its difficult to know who / what you are "replying" to (think I may go back and delete some of my droppings ;-) - But I certainly never meant to imply that WT synthesis is crap for everything.. I love the sound from my Yamaha EP and other WT modules

My only "objection" to WT synthesis is that I dont think its good for sounds where the frequency is doing a lot of gliding - I cannot emulate a good portamentoed synth sound with WT, and (although I havent extensively tried by any stretch) cannot get a good theremin sound using it.. I dont think the "problem" is primarily the lack of frequency dependant harmonic change, it think its the lack of pitch movement related change - that the cycle-to-cycle "conformity" is inferior to a waveform which alters continuously, including continuously within a cycle as a function of changing pitch when synthesising instruments whose character depends on this.

As to whether the theremins "character" does depend on this - well, I am inclined to think it does, or that is a really important component anyway - but I am also fully aware that I could be completely wrong! ;-) .. If I am "right", then WT will never be able to convincingly emulate a "real" theremin - other digital synthesis probably could, but not WT.. WT Synthesis would be the worst possible synthesis scheme to use for a theremin.. (unless there was a way to dynamically distort the cycle period to emulate the way heterodyning or analogue waveforms are distorted)

If I am wrong then WT is the right route for the theremin, and also the right way to get rid of all the other inferior old technologies like analogue subtractive synthesis - because in theory, with complex enough WT's and fast / sophisticated enough processing, WT can do everything!

Fred.

OT - Not about theremini - personal exchange which shouldnt be here..

" In the time your reaction was completely different. If I remember well you even made some comments which I perceived as rather hostile and which were considerably reducing my motivation to write anything here in the forums during a long time" - Thierry

Sorry Thierry.

I think we do this to each other a lot - Not sure why it happens, whether its all just misunderstanding or communications or .. who the fuck knows..  No one has made me feel like storming out of TW (as I did from Levnet - but it wasnt you who delivered the deciding punch there) never to return, more than you! LOL - We seem to be able to deliver near mortal blows to each other quite frequently... Some of these sting or ache even now...

Fred.

"No one has made me feel like storming out of TW (as I did from Levnet - but it wasnt you who delivered the deciding punch there) never to return, more than you! LOL - We seem to be able to deliver near mortal blows to each other quite frequently... Some of these sting or ache even now..."

I'm really sorry to read this, Fred. It clearly demonstrates my obvious lack of what they call "social intelligence"... I have to go to work now, but I'll try to write you an email during the day.

OT - Not about theremini .. personal exchange with moderator ;-)

"I'm really sorry to read this, Fred. It clearly demonstrates my obvious lack of what they call "social intelligence"... I have to go to work now, but I'll try to write you an email during the day." - Thierry

LOL ;-) .. Lets just leave it here, Thierry .. Its nothing to do with "social intelligence" or any psychodynamic shit IMO  - it about being human, possibly a little about being male (LOL - I think women are even worse - often when they make "mistakes" any fool can see that they were contrived "mistakes ;-) About having high technical standards, too much pride, probably a little too much arrogance - Oh - Just in case you misunderstand, im talking about myself here!! - but feel free to help yourself to anything from that (greatly abreviated / reduced)  list which you feel applies to you! ;-) 

We are human! - The only way we have any hope of being perfect is after were dead, when we cant say or do stupid things!  ;-)  ..  While I can still say stupid things, that means I can also say sorry for stupid things Ive said, and thats all any of us can do .. Admit weve been wrong, make whatever reparations we can, and move on..

.. Fred.