He says he has problems with recording constant turn velocity. There is no information about the surface the mouse reads off - if that is uneven, that may cause problems.
It's quite a odd phenomenon. When we have quickly changing velocity, such as while scratching rapidly back and forth it behaves fine yet when the movement is even (which may still be a high speed movement), such as when giving the platter a firm shove and letting it run out there is considerable jitter. This might be in the uneven surface or it might be some artefact of the method of scanning used by the mouse itself.
So far Rob implemented a variable interpolation that depends on the change in movement speed which does seem to work quite well. More research will be needed. We need more research anyway as it's not clear how we'd emulate the equivalent of a slipmat in digital scratching. The weight of the platter in this case seems to imply one might be useful.
For recording turns, one should use a Hall sensor, like this one:
http://www.globalspec.com/FeaturedProducts/Detail/BEIDuncanElectronics/9360_Rotary_Dual_Output_Hall_Effect_Sensor/73944/0
This is inexpensive though not USB, but netsearching for words like "USB rotary joystick Hall effect controller" may produce some results.
For my own use I already ordered a HID development board which I plan to use with a optical step-less rotary encoder. I may change my perspective here but right now relative rotation seems more interesting to me than absolute angular position. I could imagine reaching different perspectives on that and going with a Hall one though.
Yes, I see. Personally I don't care that much about ornaments ...
They are important in melodic development to provide character and style.
Again we have a small misunderstanding; I thought you meant things like volume control buttons when you said "ornamentation" and didn't realise you were refering to the musical sense of the word. My bad.
This is pretty much and individual preference - some like heavy weighted keys, and even think it helps playing faster, others the reverse. All typing keyboards seem to have an unpredictable depressing point which makes them less suitable for music interpretation.
Well, that is what I meant; typical for the Model M is that it gives both tactile and aural feedback at the exact moment the stroke is registered. This makes it a favourite of some and hated in crowded offices (where the noise may be objectionable). IBM went as far as outfitting some models with a actual speaker to give additional noise at the moment a stroke registers, that is probably taking the idea too far for most (I'm not making this up, I have keyboards with a speaker grill right here though without the speaker, I know it sounds far fetched).
These are the GPUs I mentioned. Mac OS X is developing special language support for these. So the exact driver thing will be transparent. Learning a new way to program fro every device requires too much of an effort. So this is probably the trend for the future.
As a side note; this sounds revolutionary now but the Amiga platform was well known for this. They used a aditional chip for sound as well.
Yours,
Kas.