Diferentes abordagens evolutivas aplicadas no processo de transcrição automática de partituras musicais em tablaturas

• Ano de defesa: 2015
• Autor(a) principal: Ramos, João Víctor
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Tecnológica Federal do Paraná; Cornelio Procopio; Brasil; Programa de Pós-Graduação em Informática; UTFPR
• Palavras-chave em Português: Algorítmos computacionais; Música por computador; Tablatura (Notação musical); Computer algorithms; Computer music; Tablature (Musical notation); CNPQ::CIENCIAS EXATAS E DA TERRA; Ciência da Computação
• Link de acesso: http://repositorio.utfpr.edu.br/jspui/handle/1/2547

A tablature is an alternative music notation format for string instruments. It contains the exact position (string and fret) witch needs to be played to produce the desired sound. The conversion of the musical staff notation (sheet music) to a guitar tablature, also known as transcription, is not a straightforward process. The reason is that each note on the music notation format can be played in different positions of the guitar fretboard. So this problem can be classified as a combinatorial optimization problem. In this work, we have employed a comparative study of different algorithms: A*, genetic algorithms (AG), genetic algorithms with use of subpopulations (AG-SP), ant colony optimization (ACO) and differential evolution (ED). I was also included heuristics based on local search 2-opt and 3-opt in the aproaches AG, AG-SP and ED. The experimental results with a dataset of 87 monophonic musics indicated that the approaches ACO, AG-SP 2-opt and AG 2-opt obtained the best results. Finally, each approach were statistically compared using the ANOVA with the Tukey post hoc, where the analysis indicated that there are significant differences between the evaluated algorithms.