Proceeding of the past Conferences /

Below you can read the papers of the past Conferences

New Music Concepts 2020

Automatic Identification of Melody Tracks of Piano Sonatas using a Random Forest Classifier

Part of the New Music Concepts book series (NMC volume 7)
ISBN: 978-88-944350-3-0

Author(s): Po-Chun Wang, Alvin W. Y. Su

Abstract: In this paper, an identification method of melody tracks of classical piano sonatas is presented. The tracks which are regarded as 'melody lead' are important cues in music interpretation when symbolic sheet music is concerned, especially when computer synthesis of emotional and expressive music is desired. In this work, four new features are proposed. Combined with five conventional features, there are nine features to be extracted from a standard MIDI file. Then, random forest classifier is applied to determine whether a measure is 'melody-like' or 'accompaniment-like'. There are 8 manually annotated classical piano sonatas used to validate the proposed method. Over 90% accuracy is achieved and is 6% higher than the previous work in this art.

Keywords: Melody finding, Music analysis, Music perception, Symbolic representation