Volume Antenna Linearity - Question..

Thank you, Thierry -
I will be contacting you soon - but just getting to the point of formulating sensible questions is taking ages!

Effectively, I need to take the numbers which follow an inverse square curve, and convert these to a logarithmic curve - without using a lookup table, using minimum processor overhead etc..

I am running my emulator on the existing ASIC, so that I can determine the numbers which I am getting at present.. I will then plot these, and generate a plot of what I actually need.. At this point I will email it all to you.. It may be that there is some simple algorythm which will do the job - but I am awkward dealing with large floating point results and then converting these to smaller usable integer results without causing distortion or losing too much resolution (the main reason I stay with analogue circuits wherever possible)

I am also (since examining the way other theremin volume circuits operate) thinking that perhaps I could implement a similar scheme in the ASICs Analogue blocks, using an auto-tuned bandpass filter..

So, at the moment, everything is up in the air...

There is some sense in a non-linear volume response.

Small, delicate sounds are played with small, delicate motions, and big, bold sounds are played with big, bold motions.

Like a conductor.

(The sort that wields a baton.)

Fred, there is no problem. Almost every behavior my be approximated by simple functions and integer numbers. Each polynom may be transformed into the so called Horner scheme, which allows to calculate function values by a series of simple additions and multiplications. And, as I know from my ancient audio design times, we may allow an error of +/- 2 db without audible consequences, as long as the resulting response curve will keep monotonic increasing or decreasing.

Gordon, thank you for reminding the conductor's behavior. This gave me the idea that in a further step, we could optimize Fred's volume response with a dynamic factor. That means that the same movement would lead to a more important volume change when carried out rapidly.

Fred, just for my personal curiosity: Why do you need to get logarithmic output values? There are some logarithmic VCA designs whose response is between -6 and -12dB/V between 0,5 and 4 V. Two of them in series would give you the desired range. I'll look if I can find that if you are interested.

In general: I like this kind of discussions but I'm often deceived when I have to wait more than 24 hours for a reply...
What do you all think about a fixed time each week, where Fred and the pool of his "muses" could do a text or audio based just-in-time conference, i.e. with Skype or MSN?
The only problem would be that my spoken English is still worse than the written version, since I may not have the time to look up missing vocabulary on dict.leo.org... ;-)

Not for me, I prefer the thinking time that talking on the forum permits.

The problem I'm having is that no-one, with perhaps the exception of Fred, has used a theremin with a linear volume response. Maybe it has advantages, maybe it's too different to what players are used to. I do know that I have never heard the volume linearity questioned before. Rather, people criticise volume loops for being too sluggish or too snappy. This is a musician's rather than an electrician's assessment, and it is not clear to me what that means in terms of the geometry of the response field.



Fred, tangentially, had you noticed that there's going to be an accumulation of top-notch players and a broad selection of theremins in nearby Germany come September?

http://www.edn.com/contents/images/120601di.pdf

Page 2 of 7 shows the vca design.

Hi Gordon + Thierry..

1.) Much as it is a great idea, I cannot do any fixed-time conferencing.. My life is much too chaotic! It is quite likely (in fact it almost always happens) that the moment I start something which has been planned, one of the children will fall out of bed, or one of them will lock themselves in the loo, or I will have a hypoglycaemic episode and be unable to speak either because I an shoveling glucose into my mouth, or am incoherent because my brain is sugar starved (and when that happens, I can also be quite outrageous! ;-)

No - I have found that planning anything is almost a certain recipit for disaster!

Sorry about delays - but I cannot escape these.. I do the best I can to keep things rolling - but life has a different rythm to the tempo I would choose... The last 2 weeks, for example, I have been looking after my wife + children following her emergency apendectomy.. Somehow, the gods seem to have fun 'planning' this sort of event to make them always happen at the worst possible moment.

2).. Following from the above - Gordon, I have certainly got plans to attend the European Theremin event.. I would seriosly like to have a complete Theremin ready, and hope to perhaps get this done in time to book some slot at the event so that I could demo it.

3.) Thierry - Yes, I could use non-linear VCA's or other non-linear analogue external components to implement linear volume control - BUT - I am trying to produce a ASIC which can be used by other developers / manufacturers, and do not want to make external component selection too restrictive or expensive.

4.) Actually, I think much of what I said on this thread regarding volume non-linearity was wrong, or at least close to wrong.. I think that a correctly tuned Etherwave will give a position->volume response which is quite close to linear.. I made some big mistakes in my initial assessment of how the volume circuits would perform. I am quite sure that I am not the only person who has played a Theremin with Linear volume response.

"people criticise volume loops for being too sluggish or too snappy." - I now think there are 2 reasons for this - one is the speed of response on the volume circuit, the other is its degree of linearity - A non-linear response will tend (I think) to make timing delays 'feel' far worse, as the dynamics will appear to change more rapidly in one part of the curve than at other parts.. A slow response and non-linearity will, for example, make it extremely difficult to go from silent to soft volume smoothly.

Thierry - I am most interested in your mathematical suggestions.. but it will be a few days at least before I can send you anything to look at.. All the relevant numbers prior to mathematical manipulations are generated from quite a complex sequence which I have had to change (The capacitance sensed is auto-ranged, this result produces a multiplier which is applied to the instantanious value read, this result is multiplied by the sensitivity setting - only after all this is final scaling performed to get number to feed to linearising maths manipulations) - I want to be sure I have these stages working perfectly before I send you the data.

I am extremely grateful for your help and support - It is good to know that there are people I can turn to when I get a bit stuck!

Fred wrote: [i]I wonder if this is the first ever Theremin ASIC?[/i]

And how long before you start work on the first memristor (http://www.spectrum.ieee.org/may08/6207)(*) based theremin?

:-)



(*) [i]Memristor[/i] a.k.a. Flux Capacitor (http://en.wikipedia.org/wiki/De_Lorean_time_machine#Equipment). See this (http://en.wikipedia.org/wiki/Talk:Memristor#Is_it_seriously_also_known_as_a_flux_capacitor.3F_lol) talk page on wikipedia.

I can't wait until memristors are readily available. I'm thinking a memristor theremin would be much more stable than a normal theremin and would not have a start up or shut down sound.

Anyway, staying on topic, I would imagine a linear volume antenna would be great. You would be able to smoothly transition from high to low volume with little effort. Keep up the good work!

Memsistors, if and when they do become something we can actually buy and have stay in production for a bit, might be very interesting.

I can envision it as the gain setting element in sort of an analog sample-and-hold circuit for those moments when you might like to pre-set the level of your theremin for a passage without having to hold your hand still the whole time.

Don

"And how long before you start work on the first memristor(*) based theremin?"

LOL! - I had never heard of these damn things till following your link!.. Have heard of the flux capacitor though - Back to the Future??

Well - Lets see.. If I can use the memresistor as a flux capacitor to build a time machine, I could start work on the memresistor Theremin a week ago!