A Music Annotation System for Multiple Interpretations of Tunes

PDF
Katsuhiko KAJI
Graduate School of Information Science, Nagoya University
Katashi NAGAO
Center for Information Media Studies, Nagoya University

Abstract

1 Introduction

2 Annotation Management

3 Annotation Acquirement

3.1 Tune Annotator

Example of Tune Annotator

Fugure1: Example of Tune Annotator

3.2 Tune Annotator

Definition of ElementPointer for Timeline Segmentation

Fugure2: Definition of ElementPointer for Timeline Segmentation

A Part of Annotation Definition

Fugure3: A Part of Annotation Definition

Example of Timeline Annotator

Fugure4: Example of Timeline Annotator

Example of Segment Annotation

Fugure5: Example of Segment Annotation

A Part of Annotation for Score Generation

Fugure6: A Part of Annotation for Score Generation

A Part of Generated MusicXML

Fugure7: A Part of Generated MusicXML

3.3 Tune Annotator

Example of Score Annotator

Fugure8: Example of Score Annotator

Annotation to the Selected Objects

Fugure9: Annotation to the Selected Objects

4 Annotation Availment

4.1 Music Retrieval System

4.1.1 Playlist Recommendation

4.1.2 Query by Keyword

Music Retrieval Form

Fugure10: Music Retrieval Form

4.1.3 Query by Chord Progression

4.1.4 Narrowing down by Musical Structure

4.2 Playlist Recommendation System

4.2.1 Annotation Definition

4.2.2 Feature Values of Tunes

4.2.3 Playlist Recommendation

Example of Plyalist

Fugure11: Example of Plyalist

5 Experimentation

Percentage of Reason

Fugure12: Percentage of Reason

Relationship between Highlight and Chorus

Fugure13: Relationship between Highlight and Chorus

Deviation of Reasons

Fugure14: Deviation of Reasons

Destribution of Highlight and Chorus

Fugure15: Destribution of Highlight and Chorus

6 Related Researches

6.1 CDDB, MusicBrainz

6.2 TS-Editor

7 Summary