The guitar chord-generating algorithm based on complex network

• Published in: Physica A Vol. 443; pp. 1 - 13
• Main Authors: Ren, Tao, Wang, Yi-fan, Du, Dan, Liu, Miao-miao, Siddiqi, Awais
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.02.2016
• Subjects: Chord progression; Chord-generating algorithm; Complex network; Guitar chord; Musical motif; Two-tiered network; Musical motif; Chord-generating algorithm; Chord progression; Two-tiered network; Guitar chord; Complex network
• ISSN: 0378-4371; 1873-2119
• DOI: 10.1016/j.physa.2015.09.041

This paper aims to generate chords for popular songs automatically based on complex network. Firstly, according to the characteristics of guitar tablature, six chord networks of popular songs by six pop singers are constructed and the properties of all networks are concluded. By analyzing the diverse chord networks, the accompaniment regulations and features are shown, with which the chords can be generated automatically. Secondly, in terms of the characteristics of popular songs, a two-tiered network containing a verse network and a chorus network is constructed. With this network, the verse and chorus can be composed respectively with the random walk algorithm. Thirdly, the musical motif is considered for generating chords, with which the bad chord progressions can be revised. This method can make the accompaniments sound more melodious. Finally, a popular song is chosen for generating chords and the new generated accompaniment sounds better than those done by the composers. •The analysis of the chord progressions encountered in pop music by complex network theory is interesting and useful.•The concept of musical motif is considered for generating chords automatically.•The bad chord progressions can be revised that can make the accompaniments sound more melodious.