Design of an Automatic Music Transcription System for the Traditional Repertoire of the Marovany Zither from Madagascar: Application to Human-Machine Music Improvisation with ImproteK

Design of an Automatic Music Transcription System for the Traditional Repertoire of the Marovany Zither from Madagascar: Application to Human-Machine Music Improvisation with ImproteK

Dorian Cazau (Sorbonne Universités, France), Marc Chemillier (Centre d'Analyse et de Mathématique Sociales, France) and Olivier Adam (Sorbonne Universités, France)
DOI: 10.4018/978-1-5225-0270-8.ch010
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

This chapter presents an original approach for the development of an automatic music transcription system of a Malagasy traditional plucked string instrument, called marovany zither. Our approach is based on a technology of multichannel capturing sensory system, which allows breaking down a complex polyphonic audio signal into a sum of monophonic sensor signals. A very high precision in transcription is obtained, i.e. & gt; 95% on the average note-based F-measure metric. A second part of this chapter consists in using these transcripts in the human-machine improvisation system ImproteK. Details of an exploratory working session with a local Malagasy musician are reported and discussed.
Chapter Preview
Top

Background

Music Information Retrieval (MIR) (Schedl, Gomez & Urbano, 2014) is a new emerging area of multimedia research dealing with the analysis of music in digital form to retrieve audio information efficiently and effectively, in response to the tremendous recent growth of music-related data digitally available. Some of the problems that the MIR community attempts to solve include the classification and organization of music, recommendation systems and the complex analysis of large musical databases by musical experts. Many of these problems are of a very tangible commercial interest, but most are related to a simple desire to understand basic music functions by using large databases and the power of computer processing. Among the MIR tasks, the goal of Automatic Music Transcription (AMT) is to extract the fundamental frequencies of all (possibly concurrent) notes within a polyphonic musical piece (Klapuri, 2004). The goal of AMT is then to transform a music performance, given in the form of an audio recording, into a symbolic representation by the use of signal processing methods. More specifically, AMT consists of an automatic estimate of notes in music recordings, through four attributes: onset time, duration, pitch and velocity. On the long range, AMT systems, with the purpose of retrieving meaningful information from complex audio, could be used in a variety of user scenarios such as searching and organizing music collections with barely any human labor.

Complete Chapter List

Search this Book:
Reset