A beat histogram is formed using a perceptual onset detection frontend, then leaky integrators for each of 100 periodicities. On the ensuing histogram, statistics are taken.
Operate feature extraction at control rate
fft |
Output of FFT UGen |
[0] entropy of beat histogram
[1] ratio of the largest to the second largest entries in the beat histogram
[2] diversity (Simpson's D measure) of beat histogram
[3] metricity (consistency of high energy histogram entries to integer multiples or divisors of strongest entry)