Análise do vibrato e do bending na guitarra elétrica a partir dos descritores de expressividade da ferramenta Expan

• Ano de defesa: 2015
• Autor(a) principal: Tairone Nunes Magalhaes
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: bending; sonologia; expressividade musical; computação musical; guitarra elétrica; vibrato; Musicologia sistemática; Musica; Música para guitarra; Música por computador; Vibrato (Musica)
• Link de acesso: http://hdl.handle.net/1843/AAGS-ADAK4R

On the electric guitar, bending and vibrato are techniques that, according to common sense, play a key role in a guitarists playing signature. This work aims at investigating this hipothesis by means of the features of bendings and vibratos that might dierentiate one guitarist from another. Moreover, we propose a model for the categorization of performances composed by four stages: (1) pitch extraction; (2) note segmentation; (3) extraction of descriptors for bending and vibrato; (4) analysis and categorization of performances based on the proposed parameterization, using KNN (K-Nearest Neighbors) and SOMs (Self-Organizing Maps). In order to validate the model, we analyzed a set of several performances of two short musical excerpts, performed by 8 guitarists. We found that there is a consistency in the data obtained by applying the proposed model to performances executed by the same guitarist..