Lindy Mayfield wrote:

> Or take it a step further.  If you are using a microcontroller 
in the equation then why not let it do some of the work for you.

Uhh, that's exactly what I was suggesting.

Measure hand distance. You want it linear since most
instruments have linear spacing of their keys.

Compute note number.

Do math. N * 1.059463094....(12th root of 2) * base frequency
= note frequency.
Simplest is to compute offline in  spreadsheet and use a lookup table
with some interpolation to get smooth note gliding.

> I cannot remember but I think the distances between the antenna and the hand relative to the pitch is ... I don't know what the word is or how to explain what I mean... logarithmic is it?  

Logbut with a fairly low slope (log vs linear graph paper).

> 
> For example, from 110Hz to 220Hz, A to A, is one octave or 12 steps on a piano.  From 220Hz to 440Hz is still another A to A octave and 12 steps on a piano or any other instrument.  But the difference between the two has gone from 220-110=110 to 220-440=220.  Logarithmic is that?
> 
> But the difference between the steps on a piano are based on distance and therefore equal.  Every key is the same width, of course.  (-:
>

> I'm not sure how the pitch goes on a Theremin but I think it is logarithmic rather than linear.  

It's actually 1/X because of the inverse square law. Close enough since 
the human corrects the pitch.

> My idea is what would a Theremin be like to play if it were linear like a piano with each step being x number of millimeters distance to the antenna?  

Sure. See above comments about how to compute the needed frequency.

> A PIC could do the math to make that happen.  

No hard math needed. Just use a lookup table and interpolate.

> I think for me at least I would have to try each one to see which is easier.  

Easier how? To play or to code?

>But this is a cool topic, for me at least. (-:

> What do you think?

Do it step by step.
Since measuring the hand position is going to be the hardest part
(IMO), start with that.  Use it to directly compute a pitch value for
the PIC timers. Then change the shape of the transfer function (log vs 
linear).

Have fun.

But somehow I think it's more 'real' to use RF sensing.

Robert

  > -Lindy
> 
> -----Original Message-----
> From: piclist-bounces@mit.edu [mailto:piclist-bounces@mit.edu] On Behalf Of Robert Rolf
> Sent: 2. joulukuuta 2005 0:05
> To: Microcontroller discussion list - Public.
> Subject: Re: [OT] Theremins
> 
> 
> I rather like the idea of an optical thermin.
> They'd be just as hard to play, but would be easier for the
> digital geek to build.
> 
> Ranging could be as simple as a modulated IR source with
> a tuned amplifier on the photodetector so that you get a linearish
> response to intensity vs hand distance. (1/R^2). Then simple math
> to scale to a linear note number times 12th root of 2 note intervals.
> 
> You could even cheat and have the PIC 'fix' your bad pitch just
> as modern DSPs do for today's not so good singers.
> 
> 
> Robert

-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist