Originally Posted by ArtPE
300 hz is plenty...that's a sample every 3.3 ms...generally speaking you need to sample at 2x the frequency of the signal you are trying to capture data for...
so if the shift is 30 ms, 3.3 is 9x's...
3.3 is theoretically enough to capture 8 ms...
If a function f(t) contains no frequencies higher than W cps, it is completely determined by giving its ordinates at a series of points spaced 1/(2W) seconds apart.
Not if you want to be more descriptive of what exactly is happening during the shift. It would also help identify the transitions better. It is possible that the first shift is longer than 35 ms for instance. But the noise needs to be reduced first.