Title :
Form recognition for Johann Strauss´s waltz centos based on music features
Author :
Liu, Dan ; Zhang, Nai-yao ; Zhu, Han-Cheng
Author_Institution :
Dept. of Autom., Tsinghua Univ., Beijing, China
Abstract :
Presents a form recognition method for Johann Strauss\´s waltz centos. In this method, a multilayer process strategy is adopted to make the recognition much more efficient. Three assumptions based on music theory are put forward to simplify the form recognition procedure. In addition, a new structure called "melody tree" is created to capture the essential aspects of melody, which is the most important feature in the form recognition. We have done many experiments on John Strauss\´s, most famous waltz centos to check the validity of this method, and the average correct ratio is 98.21%. This experimental result shows that the presented form recognition method is quite effective.
Keywords :
feature extraction; music; Johann Strauss; MIDI; features extraction; form recognition; melody tree; multilayer process strategy; music features; music theory; waltz centos; Art; Automation; Data mining; Instruments; Multiple signal classification; Music; Nonhomogeneous media; Rhythm; Shape; Timbre;
Conference_Titel :
Machine Learning and Cybernetics, 2002. Proceedings. 2002 International Conference on
Print_ISBN :
0-7803-7508-4
DOI :
10.1109/ICMLC.2002.1174492