I did look at this (angular acceleration rather than angular velocity) a couple of years ago, but I don’t think I got very far. I remember playing with it for a while with Maths Channels, and I think I quickly came to the conclusion that the number of teeth on the flywheel (around one for every 10 degrees of rotation) was far too few for accurate results (plus there was the interpolation at the missing teeth). My interest lay in determining the health of the engine by looking for discrepancies between cylinders. Instead of looking at the acceleration of the crankshaft, I looked at the reaction: the acceleration of the engine block. I used the 3-axis accelerometer and interface from the NVH kit and placed the accelerometer on a cylinder head bolt and looked at a single channel in PicoScope 6, taking advantage of the equal and opposite reaction.
To give credit where it’s due, this was a collaborative effort: it was Steve Smith who came up with the idea and gave me advice and guidance with it.
Yes, my interest is in determining cylinder contribution balance (just intellectual curiosity). At lower engine speeds it could be possible to identify engine misfires. It is the same the way the engine control ECU determines this (I live in Toyota World).
Apparently ATS Escope has an application that performs this task.