comparison signal_processing.py @ 3:4be70944d21b skeleton_v2

Add some basic signal processing methods we need

+
+import numpy as np
+
+def rms(samples):
+    """Return the root mean square of the given samples"""
+    return np.sqrt(np.mean(np.array(samples)**2))
+
+def autocorrelation(samples):
+    """Return the autocorrelation of the given samples, as an array with zero lag at the start"""
+    acf = np.correlate(samples, samples, mode='full')
+    return acf[len(acf)/2:]
+
+def bpm_to_lag(bpm, hops_per_sec):
+    """Return the autocorrelation lag corresponding to the duration of a beat at the given tempo in bpm"""
+    return int((60.0 / bpm) * hops_per_sec)
+
+def lag_to_bpm(lag, hops_per_sec):
+    """Return the tempo in bpm whose beat duration corresponds to the given autocorrelation lag"""
+    return (60.0 * hops_per_sec) / lag
+