Certainly that would be the better solution, but writing an acoustical model of a floppy drive is a little ott don't you think
?
I think the easier and faster solution, would be to have the 84+ samples and play the relevant one depending upon the head position.
I may give this a go myself when free time permits.
Some samples of the formatting process (All 84+ tracks please) would certainly help speed things along a bit
.