Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


jshafer_dissertation_final.pdf (23.07 MB)

Supplemental Files

File List


jshafer_supplement.pdf (86.35 MB)

ETD Abstract Container

Abstract Header

Classification and Pattern Extraction: Application of Wavelets in Music Analysis

Shafer, Jennifer Christine
http://rave.ohiolink.edu/etdc/view?acc_num=osu1460977726

Abstract Details

2016, Doctor of Philosophy, Ohio State University, Music.
This project investigates the use of wavelets for pattern recognition in music, with applications to classification tasks and analysis. Wavelet analysis is a digital signal processing technique whose applications include data compression and pattern recognition and extraction. Wavelets are designed for multiresolution analysis, and thus are a useful tool for examining patterns at both small and large levels; in application to music, this could be compared to patterns at the level of motive and patterns at the phrase level. Wavelet analysis as a step in the analysis process is an idea that is still in the experimental stages, and in this document wavelet analysis is applied to symbolic representations of music and two applications focusing on pattern recognition are explored. A conceptual explanation of wavelet analysis is presented first, with care to keep the discussion focused on applications to music analysis. This first explanation is accompanied by a series of entry-level examples demonstrating the capabilities of wavelets for recognizing musically relevant patterns; these capabilities are the primary motivation for the two projects which conclude the study. In the first project, a corpus study of melodies of French and German art songs explores the use of wavelet-analyzed representations for encoding melodic similarities and using this information to classify phrases and songs. The wavelet-analyzed data is compared to use of “raw” unprocessed data for the same tasks, and other parameters including the type of similarity or distance measure and the musical content (pitch, rhythm, or combined pitch-and-rhythm) are also evaluated. The behavior of individual composers is examined, with the goal of studying what information about stylistic tendencies can be observed from the results of the corpus study. In the second project, more traditional analysis work is executed on Karlheinz Stockhausen’s Tierkreis cycle and on Iannis Xenakis’s Analogique A; an excerpt from Xenakis’s Evryali is also examined. In these applications, the wavelet data is used as a preliminary step to help to motivate structural and motivic analysis of the works. Motivic relationships are examined in detail in the Stockhausen analysis, along with larger-scale patterns that link the cycle as a whole. Larger-scale analysis is the main focus of the work on the Xenakis compositions, and the results reveal interesting elements of the “structure” in some of Xenakis’s stochastic music.
David Clampitt (Advisor)
Johanna Devaney (Committee Member)
Anna Gawboy (Committee Member)
844 p.
Music
wavelet; pattern extraction; classification; Xenakis; Stockhausen; corpus

Recommended Citations

Citations

  • Shafer, J. C. (2016). Classification and Pattern Extraction: Application of Wavelets in Music Analysis [Doctoral dissertation, Ohio State University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=osu1460977726

    APA Style (7th edition)

  • Shafer, Jennifer. Classification and Pattern Extraction: Application of Wavelets in Music Analysis. 2016. Ohio State University, Doctoral dissertation. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=osu1460977726.

    MLA Style (8th edition)

  • Shafer, Jennifer. "Classification and Pattern Extraction: Application of Wavelets in Music Analysis." Doctoral dissertation, Ohio State University, 2016. http://rave.ohiolink.edu/etdc/view?acc_num=osu1460977726

    Chicago Manual of Style (17th edition)

osu1460977726
700
© 2016, all rights reserved.
This open access ETD is published by The Ohio State University and OhioLINK.