A hybrid system for automatic generation of style-specific accompaniment

Ching-Hua Chuan, Elaine Chew

Research output: Contribution to conferencePaper

40 Citations (Scopus)

Abstract

Creating distinctive harmonizations in an identifiable style may be one of the most difficult tasks for amateur song writers, a novel and acceptable melody being relatively easier to produce; and this difficulty may result in the abandonment of otherwise worthwhile projects. To model and assist in this creative process, we propose a hybrid system for generating style-specific accompaniment, which is capable of creating new harmonizations for melodies, with proper harmonic resolutions, in a style that is learned from only a few examples. In the proposed system, a chord tone determination module first learns, then determines, which notes in a given melody are likely chord tones. According to these chord tones, triads are assigned first to the bars with unambiguous solutions, and these triads serve as checkpoints. The system then constructs possible chord progressions using neo-Riemannian transforms between checkpoints, and represents the alternate paths in a tree structure. A Markov chain with learned probabilities for these neo-Riemanian transforms then generates the final chord progression. We select four songs by the British rock band, Radiohead, to evaluate the system. Three songs are used for training, and an accompaniment is generated for the held out melody. We present the results of two case studies. We find that the system generates chords closely related to the original, and the resulting chord transitions reinforce the phrase structure of the melody.

Original languageEnglish (US)
Pages57-64
Number of pages8
StatePublished - Dec 1 2007
Externally publishedYes
Event4th International Joint Workshop on Computational Creativity, IJWCC 2007 - London, United Kingdom
Duration: Jun 17 2007Jun 19 2007

Other

Other4th International Joint Workshop on Computational Creativity, IJWCC 2007
CountryUnited Kingdom
CityLondon
Period6/17/076/19/07

Fingerprint

Hybrid systems
Markov processes
Rocks

Keywords

  • Automatic style-specific accompaniment
  • Chord tone determination
  • Markov chains
  • Neo-riemannian transforms

ASJC Scopus subject areas

  • Computational Theory and Mathematics

Cite this

Chuan, C-H., & Chew, E. (2007). A hybrid system for automatic generation of style-specific accompaniment. 57-64. Paper presented at 4th International Joint Workshop on Computational Creativity, IJWCC 2007, London, United Kingdom.

A hybrid system for automatic generation of style-specific accompaniment. / Chuan, Ching-Hua; Chew, Elaine.

2007. 57-64 Paper presented at 4th International Joint Workshop on Computational Creativity, IJWCC 2007, London, United Kingdom.

Research output: Contribution to conferencePaper

Chuan, C-H & Chew, E 2007, 'A hybrid system for automatic generation of style-specific accompaniment' Paper presented at 4th International Joint Workshop on Computational Creativity, IJWCC 2007, London, United Kingdom, 6/17/07 - 6/19/07, pp. 57-64.
Chuan C-H, Chew E. A hybrid system for automatic generation of style-specific accompaniment. 2007. Paper presented at 4th International Joint Workshop on Computational Creativity, IJWCC 2007, London, United Kingdom.
Chuan, Ching-Hua ; Chew, Elaine. / A hybrid system for automatic generation of style-specific accompaniment. Paper presented at 4th International Joint Workshop on Computational Creativity, IJWCC 2007, London, United Kingdom.8 p.
@conference{7af019feffe54f6793fbce2e00516736,
title = "A hybrid system for automatic generation of style-specific accompaniment",
abstract = "Creating distinctive harmonizations in an identifiable style may be one of the most difficult tasks for amateur song writers, a novel and acceptable melody being relatively easier to produce; and this difficulty may result in the abandonment of otherwise worthwhile projects. To model and assist in this creative process, we propose a hybrid system for generating style-specific accompaniment, which is capable of creating new harmonizations for melodies, with proper harmonic resolutions, in a style that is learned from only a few examples. In the proposed system, a chord tone determination module first learns, then determines, which notes in a given melody are likely chord tones. According to these chord tones, triads are assigned first to the bars with unambiguous solutions, and these triads serve as checkpoints. The system then constructs possible chord progressions using neo-Riemannian transforms between checkpoints, and represents the alternate paths in a tree structure. A Markov chain with learned probabilities for these neo-Riemanian transforms then generates the final chord progression. We select four songs by the British rock band, Radiohead, to evaluate the system. Three songs are used for training, and an accompaniment is generated for the held out melody. We present the results of two case studies. We find that the system generates chords closely related to the original, and the resulting chord transitions reinforce the phrase structure of the melody.",
keywords = "Automatic style-specific accompaniment, Chord tone determination, Markov chains, Neo-riemannian transforms",
author = "Ching-Hua Chuan and Elaine Chew",
year = "2007",
month = "12",
day = "1",
language = "English (US)",
pages = "57--64",
note = "4th International Joint Workshop on Computational Creativity, IJWCC 2007 ; Conference date: 17-06-2007 Through 19-06-2007",

}

TY - CONF

T1 - A hybrid system for automatic generation of style-specific accompaniment

AU - Chuan, Ching-Hua

AU - Chew, Elaine

PY - 2007/12/1

Y1 - 2007/12/1

N2 - Creating distinctive harmonizations in an identifiable style may be one of the most difficult tasks for amateur song writers, a novel and acceptable melody being relatively easier to produce; and this difficulty may result in the abandonment of otherwise worthwhile projects. To model and assist in this creative process, we propose a hybrid system for generating style-specific accompaniment, which is capable of creating new harmonizations for melodies, with proper harmonic resolutions, in a style that is learned from only a few examples. In the proposed system, a chord tone determination module first learns, then determines, which notes in a given melody are likely chord tones. According to these chord tones, triads are assigned first to the bars with unambiguous solutions, and these triads serve as checkpoints. The system then constructs possible chord progressions using neo-Riemannian transforms between checkpoints, and represents the alternate paths in a tree structure. A Markov chain with learned probabilities for these neo-Riemanian transforms then generates the final chord progression. We select four songs by the British rock band, Radiohead, to evaluate the system. Three songs are used for training, and an accompaniment is generated for the held out melody. We present the results of two case studies. We find that the system generates chords closely related to the original, and the resulting chord transitions reinforce the phrase structure of the melody.

AB - Creating distinctive harmonizations in an identifiable style may be one of the most difficult tasks for amateur song writers, a novel and acceptable melody being relatively easier to produce; and this difficulty may result in the abandonment of otherwise worthwhile projects. To model and assist in this creative process, we propose a hybrid system for generating style-specific accompaniment, which is capable of creating new harmonizations for melodies, with proper harmonic resolutions, in a style that is learned from only a few examples. In the proposed system, a chord tone determination module first learns, then determines, which notes in a given melody are likely chord tones. According to these chord tones, triads are assigned first to the bars with unambiguous solutions, and these triads serve as checkpoints. The system then constructs possible chord progressions using neo-Riemannian transforms between checkpoints, and represents the alternate paths in a tree structure. A Markov chain with learned probabilities for these neo-Riemanian transforms then generates the final chord progression. We select four songs by the British rock band, Radiohead, to evaluate the system. Three songs are used for training, and an accompaniment is generated for the held out melody. We present the results of two case studies. We find that the system generates chords closely related to the original, and the resulting chord transitions reinforce the phrase structure of the melody.

KW - Automatic style-specific accompaniment

KW - Chord tone determination

KW - Markov chains

KW - Neo-riemannian transforms

UR - http://www.scopus.com/inward/record.url?scp=84855608563&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84855608563&partnerID=8YFLogxK

M3 - Paper

SP - 57

EP - 64

ER -