Camel: Carnatic percussion music generation using n-gram models

Konstantinos Trochidis, Carlos Guedes, Akshay Anantapadmanabhan, Andrija Klaric

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In this paper we explore a method for automatically generating Carnatic style rhythmic. The method uses a set of annotated Carnatic percussion performances to generate new rhythmic patterns. The excerpts are short percussion solo performances in ādi tāla (8 beat-cycle), performed in three different tempi (slow/moderate/fast). All excerpts were manually annotated with beats, downbeats and strokes in three different registers - Lo-Mid-Hi. N-gram analysis and Markov chains are used to model the rhythmic structure of the music and determine the progression of the generated rhythmic patterns. The generated compositions are evaluated by a Carnatic music percussionist through a questionnaire and the overall evaluation process is discussed. Results show that the system can successfully compose Carnatic style rhythmic performances and generate new patterns based on the original compositions.

Original languageEnglish (US)
Title of host publicationSMC 2016 - 13th Sound and Music Computing Conference, Proceedings
EditorsRolf Grossmann, Georg Hajdu
PublisherZentrum fur Mikrotonale Musik und Multimediale Komposition (ZM4), Hochschule fur Musik und Theater
Pages475-479
Number of pages5
ISBN (Electronic)9783000537004
StatePublished - Jan 1 2019
Event13th Sound and Music Computing Conference, SMC 2016 - Hamburg, Germany
Duration: Aug 31 2019Sep 3 2019

Publication series

NameSMC 2016 - 13th Sound and Music Computing Conference, Proceedings

Conference

Conference13th Sound and Music Computing Conference, SMC 2016
Country/TerritoryGermany
CityHamburg
Period8/31/199/3/19

ASJC Scopus subject areas

  • Music
  • Computer Science Applications
  • Media Technology

Fingerprint

Dive into the research topics of 'Camel: Carnatic percussion music generation using n-gram models'. Together they form a unique fingerprint.

Cite this