About this story

Authors: Andrea Poltronieri, Jacopo de Berardinis

Part of: choco

Source: SPARQL endpoint


Chord Corpus Statistics

Exploring Chord data from different genres and historic periods

Besides being one of the most studied subjects in music, harmony has attracted the interest of musicologists, computer scientists, and psychologists due to its multi-faceted and inter-disciplinary nature. The basic constituents of harmony are chords, groups of simultaneous notes, and their progressions, which convey notions of consonance (pleasant, euphonious) or dissonance (unpleasant, discordant) in music. The Chord Corpus (ChoCo) is a large-scale dataset that semantically integrates harmonic data from 18 different sources in various representations and formats (Harte, Leadsheet, Roman numerals, ABC). We leverage JAMS (JSON Annotated Music Specification), a popular data structure for annotations in Music Information Retrieval, to effectively represent a variety of chord-related information (chord, key, mode, etc.) in a uniform way. On top of it, we propose a novel ontology modelling music annotations and involved entities (artists, scores, etc.), build a 30M triple knowledge graph on chords, and find 3K links to other datasets (MIDI-LD, LED).