Document details

Automatic cymbal classification

Author(s): Almeida, Hugo Ricardo da Costa

Date: 2010

Persistent ID: http://hdl.handle.net/10362/4923

Origin: Repositório Institucional da UNL

Subject(s): Automatic classification; Cymbal classification; Music classification; Music information retrieval (MIR); Drum kit; Cymbals


Description

Most of the research on automatic music transcription is focused on the transcription of pitched instruments, like the guitar and the piano. Little attention has been given to unpitched instruments, such as the drum kit, which is a collection of unpitched instruments. Yet, over the last few years this type of instrument started to garner more attention, perhaps due to increasing popularity of the drum kit in the western music. There has been work on automatic music transcription of the drum kit, especially the snare drum, bass drum, and hi-hat. Still, much work has to be done in order to achieve automatic music transcription of all unpitched instruments. An example of a type of unpitched instrument that has very particular acoustic characteristics and that has deserved almost no attention by the research community is the drum kit cymbals. A drum kit contains several cymbals and usually these are treated as a single instrument or are totally disregarded by automatic music classificators of unpitched instruments. We propose to fill this gap and as such, the goal of this dissertation is automatic music classification of drum kit cymbal events, and the identification of which class of cymbals they belong to. As stated, the majority of work developed on this area is mostly done with very different percussive instruments, like the snare drum, bass drum, and hi-hat. On the other hand, cymbals are very similar between them. Their geometry, type of alloys, spectral and sound traits shows us just that. Thus, the great achievement of this work is not only being able to correctly classify the different cymbals, but to be able to identify such similar instruments, which makes this task even harder.

Dissertação apresentada na Faculdade de Ciências e Tecnologia da Universidade Nova de Lisboa para a obtenção do grau de Mestre em Engenharia Informática

Document Type Master thesis
Language English
Advisor(s) Cavaco, Sofia
Contributor(s) Almeida, Hugo Ricardo da Costa
facebook logo  linkedin logo  twitter logo 
mendeley logo

Related documents

No related documents