On Operationalizing the Musicological Concept of Tune Family for Computational Modeling
Year: 2011
Type: book-chapter
Abstract: We present a computational investigation of melodic similarity among a large collection of Dutch folk song melodies, hosted at the Meertens Institute (Amsterdam). The primary aim is to develop a similarity measure for melodies that corresponds to the way musicological experts would evaluate the similarity between folk song melodies. Therefore, we aim to relate our computational solutions to existing knowledge from ethnomusicological studies on Western folk song melodies. The most important concept from the musical domain is the concept of tune family. It appears that the existing algorithm for sequence alignment (Needleman-Wunsch) with appropriate musicological knowledge incorporated is adequate to retrieve members of the same tune family from the full collection of melodies. Based on the automatic similarity assessments, the musicological collection specialists of the institute reconsidered the classification of several melodies. Furthermore, they could classify several melodies they were not able to recognize before.
Language: English
Topic: Music and Audio Processing
Subfield: Signal Processing
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W1918576746