Average Velocities and Musical Conducting?

[m00nbeam360]

Hi there,

I'm trying to create a BPM (beats per minute) calculation with virtual conducting, having taken coordinates from the Kinect's system. The idea behind it is that the Kinect will record coordinates from the movements of the hand gestures and calculate average velocities of the first 5 and last 5 x and y values. If the average velocities for the y values go from negative to positive, the hand will be going up, etc. I then came up with the idea that if the average velocity from the x-values go from negative to positive, the hand is going to the left. From the average velocities, I would calculate the time between each average velocity calculation and then the average beats per minute rate.

This would be all fine and dandy except that this won't work with legato, or smooth conducting in which conductors perform like this: . Any suggestions on how I could calculate the average velocity of those gestures, or is that impossible? Any advice would be greatly appreciated. Thanks!

 

[haruspex]

Seems to me that x v. y movement is not in itself that meaningful. The first thing is to identify the strokes that are more or less linear in space, punctuated by relatively sharp turns.