Volume Antenna Linearity - Question..

Hi again.. I'm on the scrounge for advice - Again!

I have my first completed Theremin ASIC on a board, (ASIC = Application Specific Intergrated Circuit.. I wonder if this is the first ever Theremin ASIC?).. Anyway.. this ASIC is a complete volume control section for a Theremin, incorperating the oscillator, and outputting a voltage to drive a VCA, and also outputting a MIDI volume data stream..

It works well, but not exactly as I had planned .. :{ .. I made a miscalculation with regard to the capacitance / volume relationship (I got this 'upside down' and correcting this is going to be a major pain)

My question.. How important is the volume change to hand position relationship?

I was aiming for a logarithmic response, so that, over the selected distance (which is adjustable from 30mm to about 500mm) the hand position would adjust the volume as if one was moving a Log slider over that distance..

Looking at other designs, I have now realised that the volume response on most Theremins seems to be completely different to my 'ideal' - It seems that volume change at furthest from the antenna (high volume side) is quite insensitive, and that major changes in volume only occur quite close to the antenna or null point.. Which is exactly what my ASIC does (albeit my ASICs response is more extreme, because I have added extra maths to compensate the response, and this maths is working the opposite way to what it should)

I suppose I have 3 options really - Strip out the maths, and leave the response uncompensated and non-linear, like all other Theremins seem to be - Or do some really convoluted manipulations to give true 'Hi-Fi Volume control' linearity, or produce a linear position -> voltage output, which will sound more linear than most theremins, but not give truly linear audio response (would be like having a linear potentiometer in a volume control)..

All and any ideas / advice would be welcome.

As far as I'm aware, linear volume control is someething with which few have experimented. My first instinct is to say it wouldn't be especially helpful since it would remove the attack curve. On the other hand, you want as good a range of response as possible. It needs to be able to go from as loud as possible to as quiet as possible within the full range of your elbow. That was the principal problem with the Etherwave Pro - lifting your hand any higher than 30 cms appeared to make no difference to the overall volume, and the lift from silence to sound was bumpy.

That's the only thing that has to be perfect - the pull up from silence to sound. It has to be such that, with sufficient control, it can be made almost indiscernible.

Fred, you should find a way to make the volume response adjustable:

a) In a way that if the player wants to play staccato at high volume, he may get from 0 to 100 within a few cm.

b) In another way that the player will be able to make a subtile crescendo over a path of ca. 30 cm.

Carolina Eyck writes in her method that one should not adjust ones theremin's volume response for a greater distance than 30 cm (from "Elbow forms 90°" up to "hand in the shoulders height") because on risks to affect pitch with more movement.

Thank you Charlie + Thierry..

I think I am going to need to take a closer look at my options on this, and probably implement something a lot more complex than I have at present.. But this has been the case with every aspect of this design - I could easily implement something as 'good' as what the Theremin community has put up with to date .. But at this stage I probably have the chance to make something better, and I dont want to miss this chance - even if it takes longer to get the final products to market.

---
[b]a) In a way that if the player wants to play staccato at high volume, he may get from 0 to 100 within a few cm.

b) In another way that the player will be able to make a subtile crescendo over a path of ca. 30 cm.[/b]

[i]Both of the above are possible BUT it is the control 'law' over the active range which is the problem.. for a range of say 30cm, almost no change in volume occurs for the furthest 15cm (a 3db change, in fact) then 3db's for the next 7.5cm, then 3db for next 3cm, then 3db for next 1.5cm etc... What I think is needed is a 3db change for every 1cm ... If the sensitivity is increased to span 15cm, then a 3db change per 0.5cm would be required..
-88db would always occur at <= 1cm from the antenna (maximum attenuation).[/i]

Thierry - I like the idea of adjustable linearity.. At present my ASIC is an 8 pin part, with no spare pins to implement anything else.. A am looking now at porting the design to a 28 pin part which would give more ability to fit extra features (like adjustable linearity, and selection of MIDI Channel and controller type). I will email you soon, in the hope that you can help me with some maths.

It does seem that volume response is a neglected area in Theremin design.. I am astounded that the Pro suffers from insensitivity at the further from antenna area, and that it is over-sensitive close to the antenna - this is exactly what I am getting from my ASIC. I have also found that the volume response from other Theremins I have played annoyed me - but I put this down to inexpierience.

It is FAR easier to implement a volume control following the standard Theremin design - The fact is that change in capacitance is greatest for given movement close to the antenna, so having more extreme volume change in this area is a natural outcome.. However, this seems to be the opposite of the ideal response. To get an 'ideal' response, however, requires (I think)converting an inverse-square output to a Logarythmic output - And I am having great difficulty doing this.

It is a real shame that one seeks silence close to the antenna, and loudest furthest from the antenna - The other way round is automatically implemented almost perfectly by conventional theremin volume circuitry.. Ironic!

Consideration should also be given to the shape of the response field. I play "Pamelia Kurstin style" - taking advantage of the smaller field at the side of the volume loop for staccato and playing legato above the loop.

"Consideration should also be given to the shape of the response field. I play "Pamelia Kurstin style" - taking advantage of the smaller field at the side of the volume loop for staccato and playing legato above the loop."

Thank you, Gordon.. This is extremely interesting! I have been experimenting with field shaping, using bifilar 'wound' antennas - but never thought about using this for the volume antenna - this could be a neat answer to my problems!

Hello,
Why wouldn't linear response be a good thing for the volume loop? Thereminists look for the best linearity in the pitch rod, why not have that predictability in the volume loop?
Just a couple questions from an uninformed hand- waver.

Good Luck!

teslatheremin

Hi Tesla -
I think it is fair to say that volume linearity is far less important than pitch linearity - mainly because we are incredibly sensitive to pitch variation, and much less sensitive to volume variation..

I had almost completely overlooked the complexity of the questions which arise when one evaluates the effect of the volume control profile.. And, as a result, am now again in this state of not really knowing what is 'best'..

I was aiming for a linear audio response to hand position - and my preliminary tests confirmed that I had achieved this.. But I had made a stupid basic error.. During development / testing (and, I must confess, often my preference when playing) I had the volume set up so that it increased as the antenna was approached.. Jumping about between my desk and workbench and instrument rack, making adjustments, it is annoying to have the oscillator continuously whining in the background (I drive the volume circuit from a signal generator when testing) .. And I assumed that simply swapping the control signal polarity would give what I wanted... (give the same linear response with volume decreasing as the antenna was approached) OOPS, bad miscalculation!

I personally like the linear volume response, and do not like the 'cramping' of control down at the antenna, caused by the 'natural' Theremin response.. It is FAR easier to get a smooth change from silent to soft with a linear response, and for my SLOW playing style, I have nothing I dislike about playing.. BUT, I am primarily a designer, not a player - Fast, extreme volume changes would probably be more difficult with Linear control than with 'natural' control..

Which is why I think, If I can achieve it, the ability to select different responses may be worth putting into the ASIC..

But correcting my mistake is not going to be easy - effectively I need to do quite a lot of heavy digital computation within the ASIC to get close to the performance I was getting from a few minor corrections which I had implemented in analogue..

Perhaps I could convince everyone that getting louder as you approach the antenna is more sensible ?!?

I have just done some simulation 'testing' of conventional Theremin volume circuits.. These are my preliminary conclusions:
1.) Circuits which use beat-frequency to derive volume control signal (not many are like this) give a response which is severely non-linear.. BUT the linearity is a function of the same mechanisms as employed in pitch circuits - so tricks with the inductors etc can improve the performance.
2.) Circuits consisting of an oscillator driving into a tuned LC filter (as per EtherWave) can give better linearity than I expected - The linearity and performance is directly related to the Q of the filter - and it works the right way round if oscillator / filter tuning is correct. It seems I was wrong about the response being grossly non-linear.. in fact, the response is not too bad at all!
3.) RC circuits are completely hopeless!

As people seem (in general) to like the volume response of the EtherWave, and as this CAN give reasonably close to linear volume control, I suspect that Linear control is what is desired, and what I need to implement.

Fred, if you need help with the maths, don't hesitate to contact me!