Abstract

MusicHastie is a hierarchical music representation language designed for use in human and automated composition and for human and machine learning based music study and analysis. It represents and manipulates musical structure in a semantic form based on concepts from Schenkerian analysis, western European art music and popular music notations, electronica and some non-western forms such as modes and ragas. The representation is designed to model one form of musical perception by human musicians so can be used to aid human understanding and memorization of popular music pieces. An open source MusicHastie to MIDI compiler is released as part of this publication, now including capabilities for electronica MIDI control commands to model structures such as filter sweeps in addition to keys, chords, rhythms, patterns, and melodies.