Using lossless audio compression algorithm to measure information content of musical messages. This approach base on Minimum Length Description and Kolmogorov complexity offers a quantitative measurement for musical complexity that can analyse micro, meso and macro-scale redundancy in music. Visualisation of complexity over time of a given musical message or of a musical corpus can be use as a tool for musicology studies or music information retrieval.
Python : numpy, scypi, matplotlib
Flac codec at https://xiph.org/flac/index.htm
For the sliding window and cumulative window :
Set the pathway of the .wav file and the size of the increment of window in seconds. The fenetre_cumul and fentre_glissante functions return a one dimentional array of the complexity at every second of the track.
filename = 'TheWellTemperedClavier.wav'
list_complexity_cumul = fenetre_cumul(filename,1)
list_complexity_slide = fenetre_glissante(filename,5)
