Conducting, dancing, and tapping your foot in time to music all involve periodic motions. When you play note on a piano, only the down part of the cycle is expressed. The idea behind Cranker was to extract information about the gesture involved in the rest of the cycle.
In the Cranker interface, the crank is attached to an optical shaft encoder and some simple logic circuitry; it attatches to the parallel port of an IBM PC:
What I found was that it's very hard to turn a crank evenly.
There are a couple of directions I'd like to go with this. The first is to create a circular display of the score, with a line showing the current position of the crank:
The other is to provide haptic feedback. A motor attached to the shaft would provide resistance to motion, and a pressure sensor on the handle would detect how hard the performer was squeezing. When the position of the crank was between notes, it would move without resistance, but when it was with a note, there would be resistance related to the dynamic envelope of the note (as stored in the score) and proportional to the pressure on the handle (so that there'd be more resistance if you were squeezing harder). The force applied during a note would control the dynamics of the resulting sound, giving the performer control analogous to that associated with plucking or bowing a string: the more pressure you apply, the harder it is to pluck/bow, and the more sound (and, usually, the more strident sound) is produced.