Fingerboard design / construction ideas

This thread has been started because: (This stuff is purely hypothetical at this stage !)

a.) In discussion over on the Theremini thread, I made an OT posting about my capacitive ribbon instrument.

b.) Peter then said it seemed to have some similarity to the Continuum fingerbaord, and gave some details about this instrument.

c.) Dewster and I then investigated the Continuum and started to discuss possible technology by which to implement an alternative embodiment of something similar -

If discussion on this continued at the Theremini thread, this would represent a hijack of that thread, and would also be another subject likely to get lost as its not related at all to the theremini.

As to whether this thread will develop into anything, who knows... ;-)

=============================

Coalports original description of the Continuum:

The continuum interface is a neoprene pad that has a satiny-smooth woven upper surface applied to a thin rubber sheet (same stuff as a mouse pad). It is sensitive to the slightest motion of the hand. The "keyboard" that you see printed on the fabric is strictly a visual aid for placement of the finger (or fingers) and if you want to hit a note precisely, you must be in the exact center of the printed "key". If you are off by the slightest bit, you will be flat or sharp (similar to the theremin). Yes, you can quantize your sound if you want to, but I do not.

 
When I want to play the pad like a keyboard, I hit the notes precisely and then apply pressure for volume. If I want to "gliss" like a theremin, I apply pressure and then slide to where I want to go. There is no "toggling" as there is on an ondes - I was landing on precisely the right points to get the correct pitches.

 
Underneath the neoprene pad there is a mind-boggling row of thin metal bars that "float" on hundreds of delicate springs made of steel as fine as hair. Woe betide those who attempt to look underneath the neoprene without knowing exactly what they're doing. I had my own baptism of fire, thank you very much!

 

On the other hand, I now know how the damn thing works.

 

==============================================

 

FredM's Reply to  Dewsters comment:

"That thing's crying out for capacitive sensing via FPGA." - Dewster

Having now seen pictures and explanations etc of how the Continuum fingerboard is constructed and operates, I agree with you Dewster -

This instrument is just crying out for simplification, and this simplification could probably be achieved with capacitive sensing and FPGA better than by any other means.

The the magnetic fields from the fine magnetic rods in the fingerboard are sensed by hall devices at each end of each rod, so that finger pressure and placement can be determined and processed - if these rods were non-magnetic, grounded, and capacitive sensing pads placed beneath them, exactly the same data could be derived with even higher resolution than the hall sensors are capable of.

The mechanics could be hugely simplified, the grounded rods would be between the player and the sensors, so there woulf be no capacitive interference problems, and my guess is that the whole fingerboard could be produced at a fraction of its current price.

All the existing algorythms and firmware (which is the clever bit of the Continuum IMO) should work with a capacitive fingerboard, and the 2 hall sensors required for each rod replaced with simple copper sensors (probably just PCB tracks).

======================================

Dewsters reply to above:

"The mechanics could be hugely simplified, the grounded rods would be between the player and the sensors, so there woulf be no capacitive interference problems, and my guess is that the whole fingerboard could be produced at a fraction of its current price."  - FredM

I'd try to leverage the huge I/O count in even a low-end FPGA to simplify the hardware even further.  From top to bottom:

1. The top neoprene mouse pad stuff.

2. A flexible ground layer.

3. A flexible insulating layer.

4. A flexible layer of short narrow conductive parallel strips, maybe 5mm wide spaced 5mm apart, running front to back.

5. A springy insulating layer of some sort (more neoprene?).

6. A layer of long, narrow conductive parallel strips with the same dimensions as layer 4, but running from left to right.

This forms a huge XY capacitive grid.  Sequentially drive the layer 6 strips, and watch the response on the layer 4 strips using RC time constants.

With this setup you could track loads of fingers on the thing.  You could probably DIY it for almost nothing (<= big talk).

============================================

REFERENCE STUFF:

Continuum mechanics:

 

Details of the operation can be found here.

============================================

Ok - Ready to go ;-)

Fred

"4. A flexible layer of short narrow conductive parallel strips, maybe 5mm wide spaced 5mm apart, running front to back." - Dewster

I deleted my last post because I was talking nonsense!  ;-)

Your method looks like it should work.. I use ribbon cable in my mono capacitive ribbon - I think it may be ideal as a flexible medium for your scheme.

 

I like the continuum idea because, by pivoting a rigid bar, one can obtain the Y position and force on only two sensors, and X position is easily determined by the values from adjacent sensors.

This is my idea of a simplification - no need for springs or magnets or hall sensors, and there shouldnt need to be any mechanical alignment tuning - it should be possible to DIY this.

 

Just thinking about the bars / rods - I see no reason for these to be equally spaced over the sensor pads - If one had a load of fine rigid bars even touching each other, the pads would define the spacing.. The Continuum needs to reduce the number of bars because each is sensed with hall sensors independently - However, with a capacitive sensing scheme, one is agregating all the flexing on the surface - 

The lower cost and simplicity of cap sensing would allow for more sensing pads to be fitted if required, perhaps 4 per semitone..

 Im thinking bars something like the ones in this:

If youve ever had the misfortune of needing to dissassemble and reassemble one of these, the above picture will probably frighten you ;-)

But im sure there must be a way to obtain pins like the above... No idea from where though..

Just found 3mm diameter stainless steel rod available -

 

But with all the above, I think this may be neanderthal technology..

My little Android tablet has an app (Plasma) which can sense multiple fingers and produce sounds whose pitch is determined by X and volume determined by Y - It doesnt detect force , but I suspect that the technology used could if there was a flexible surface over the sensor..

I havent got into capacitive screen sensors at all - but its probably a lot cleverer than anything I am suggesting above, and could probably be used for the above.

I like the above because its comprehensible to me, there is no direct capacitive sensing of the player, there is a solid ground between the player and the sensor / electronics, so no problems with emissions or ESD or stuff like that.. (the same is true for Dewsters scheme - but I dont think it would be true for any scheme that uses "screen" type sensing)

I suspect that "neanderthal technology" may be more robust and reliable - it often is! ;-)

Fred.

One of the things I think they do with touch grids over screens is interpolate between electrodes.  This lets them use fewer electrodes, and thus probably mostly saves on interface connections.  So they likely size the electrodes based roughly on finger tip dimensions.  They also use auto calibration, and thresholding in order to know when to do that calibration.

With a true XY grid you can get lots of multi-point, with rigid bars and endpoint sensing you are more limited, particularly when two fingers are on the same bar.

Never having touched one, I can't say I'm a huge conceptual fan of these spongy keyboards.  The Seaboard is another one: http://www.youtube.com/watch?v=8n-bEy9ISpM.  Maybe it's me, and maybe I'd totally change my mind if I played one, but the Seaboard in particular looks really cumbersome, even when played expertly.  I'd rather my fingers hit a hard, immovable surface than sponge down into foam or whatever.  The seaboard has some 3D going for it, though I wonder how much you can actually use it to locate hand position without looking - would it accidentally trigger if you were just feeling your way to the right key (or should I say sponge)?  The mass and travel of piano keys keeps this from happening on a real/digital pianos, which means you can do handy things like read the music you're playing.

But we really need alternate controllers, and I don't want to be seen as outright dumping on these things.

"but the Seaboard in particular looks really cumbersome, even when played expertly.  I'd rather my fingers hit a hard, immovable surface than sponge down into foam or whatever. " - Dewster

Yeah, I agree that the Seaboard looks way too squidgy..

Having played some crude immovable PCB "keyboards" I really dont like those either - Oh hell, the real semi-weighted tactile keyboard is damn hard to beat! ;-) Yeah, I love my Yamaha PDP-400.. Have had top-of-the-range Yamaha weighted MKB's, Clavinovas and Roland MKB's and this low-end Yamaha feels better than anything else I have played (except perhaps a wonderful Sequential Synth I played with for about 10 minutes in the '80's) - Wouldnt trade it for a Continuum! ;-)

I imagine that a flat flexible surface with not much "squidge" could be great for synth and instruments like that - Strings and wind and non-percussive is probably what id use this kind of controller for - The idea of using it for piano or vibes, well, IMO, its a bit silly .. The keyboard is ideal for these..

The Continuum on the other hand does look great - I think the rigid bars under a thin membrane probably give a less squidgy feel, whilst still allowing "squidge" to be applied..

But for me its all just speculation.. I cant imagine sinking my fingers into a squidge like the Seaboard and enjoying it ;-) .. But I can imagine enjoying playing a Continuum...

" with rigid bars and endpoint sensing you are more limited, particularly when two fingers are on the same bar."

I dont see this as a problem.. Two fingers placed on the same X point at different Y positions will still resolve into a force vector - it will be seen as a single finger, so yes.. one cannot play two seperate "voices" at the same pitch - But is this really a limitation? ... Perhaps it is - but it wouldnt bother me, LOL ;-) I have enough problems playing a single voice per hand..

But I do see what your saying, now that I think about it... Without this limitation you could be holding a chord with your left hand, and sweep "through" this with your right hand if you were playing a different voice on this, assuming voices were determined by the Y axis..

Fred.

"Without this limitation you could be holding a chord with your left hand, and sweep "through" this with your right hand if you were playing a different voice on this, assuming voices were determined by the Y axis.."  - FredM

You could probably track this kind of multiple trajectory in the bars / end sensors version using fancy software.  With an XY grid it would likely be much simpler.

For percussive control I think some kind of hard stop in the playing mechanism is almost essential in order to facilitate precision timing of played events, whether it be a simple hard surface the finger hits, a pivoting key bottoming out, a string being plucked, drum stick hitting the head, etc.  These squishy keyboards seem to lack that. 

And, IMO, precision location of the hand via feeling, without resulting in playing, is fairly essential.  The player's eyes should only be involved if/when absolutely necessary.  Anyone who learned to touch type on a PC or typewriter knows this - looking at your fingers is a crutch that keeps you from reaching top speed, and hinders you enormously when doing things like transcribing notes.  Not sure why, but my favorite way to play guitar is with my eyes closed.

Continuum great. Seaboard eh.

Rob, I really like your rendition of "Because" on the Continuum:

The Continuum seems like a labor of love, and I'm all over labors of love.

So you've played the Seaboard?  If so, could you give any details on its good / bad points?

"You could probably track this kind of multiple trajectory in the bars / end sensors version using fancy software.  With an XY grid it would likely be much simpler." - Dewster

"fancy software" ?

Im thinking about doing this in analogue..

What were you thinking ???  

;-)

Fred.

(ps - I keed)

 Just seen this:

 

Open hardware and source:

 

http://openmusickontrollers.github.io/chimaera/about/

Looks like it could be adapted to some other input mechanism..

"And, IMO, precision location of the hand via feeling, without resulting in playing, is fairly essential.  The player's eyes should only be involved if/when absolutely necessary. " - Dewster

Like you, I often play with my eyes closed..

But playing a conventional keyboard 'blind' is probably a lot easier to do than with something like a  Continuum.. one can hit almost anywhere on the key..

I dont really see a way 'round this.. other than an audio preview (which I am adopting for my ribbon) - As soon as one has tactile 'markers' one interrupts the movement 'flow' - when whizzing over a smooth slippery surface one can get the "flying" - any indentations or whatever will, I think, give a "trigger" as they are moved over - If they dont give a trigger signal, then IMO the surface probably isnt sensitive enough..

But I suppose its down to what sort of instrument one has in mind - I am thinking "theremin" or poly "theremin" or slide-synth, or something like that.. In my mind I have primary focus on mono, with poly perhaps coming much later - IMO, there has never been any poly instrument that replicates the mono synth, I love both - but IMO they are entirely different beasts.

Im sure you could play a mono fingerboard or ribbon with your eyes closed - I doubt you could for a poly, even if there was tactile marking.. But I suppose they would make life a little easier. For me though the loss of "flying" would make this unatractive.. may as well just use a keyboard.

I suppose if one could have an overlay to place on a smooth fingerboard, which was engraved or whatever to give tactile feedback, you could get the best of both worlds...

Fred.