Pattern identification in melody via wavelets

Gissel Velarde

Research output: Contribution to conference without publisher/journalPosterResearch

117 Downloads (Pure)

Abstract

This study presents a new approach for music
representation and pattern identification in melody.
Melody is consider as one of the most important elements
of analysis in musical units. Western musical examples
taken from Stein’s book Structure and Style: The Study
and Analysis of Musical Forms were digitalized into
vectors of events and analyzed via wavelets. The
examples correspond to the following uses of figures:
repetition, sequence, contrary motion and interlocking or
overlapping. The wavelet technique was used because its
capability in self-similarity detection caused by scales
and dilatations. The coefficients found by the discrete
and continuous wavelet transforms were plotted in color
scales, revealing similarity degrees. The visualization of
the coefficients allows identification of patterns and
segmentation of musical units. Future work should study
systematically a greater corpus of musical examples and
uses of figures and extend the analysis through the
structure of the form.
Original languageEnglish
Publication date13 Aug 2010
Publication statusPublished - 13 Aug 2010
Externally publishedYes
EventInternational Society for Music Information Retrieval Conference - Utrechts Conservatorium , Utrecht, Netherlands
Duration: 9 Aug 201013 Sept 2010
Conference number: 11

Conference

ConferenceInternational Society for Music Information Retrieval Conference
Number11
LocationUtrechts Conservatorium
Country/TerritoryNetherlands
CityUtrecht
Period09/08/201013/09/2010

Fingerprint

Dive into the research topics of 'Pattern identification in melody via wavelets'. Together they form a unique fingerprint.

Cite this