Slides - McGill University

farrightΛογισμικό & κατασκευή λογ/κού

15 Αυγ 2012 (πριν από 5 χρόνια και 2 μήνες)

313 εμφανίσεις

Metadata and
jMusicMetaManager

Bruno Angeles

McGill University


Schulich School of Music

MUMT
-
621

Fall 2009

Outline


Metadata: what is it?


Problems with metadata


The business value of metadata


jMusicMetaManager


Method


Uses


Nice aspects


Issues


Possible improvements

Metadata: What is it?


Ancient Greek μετά: beyond


It is “data about data”



Music: information about composer, artist,
year, etc.

Metadata: Semantical clusters


Corthaut et. al:


Musical information


Classifier


Performance


Versioning


Descriptor


Rights & ownership


Playback rendition


Lyrics


Grouping & referencing


Identifiers


Record information




Instrumentation &
arrangement


Sound & carrier


Event


Time modeling


Musical notation


Attention
-
metadata & usage


Publishing


Composition


Production


Meta
-
metadata


Problems with Metadata


Poor, inconsistent metadata


Different suppliers with different methods, different views


Copyright laws


Researchers can’t share full music databases


But they can share some metadata!


(Careful: CSoundXML’s purpose is to allow sound synthesis)


Subjectivity


Conscious hip hop vs. rap?


Not all fields are useful


McKay & Fujinaga:
Musical genre classification: Is it worth pursuing and
how can it be improved?

Business Value of Metadata


Intermediaries: a means to an end


Faster, better management of music dbs


Finding music for certain activities


High cost of switching


Libraries: Lai et al.


The Traditional Method of
Managing Metadata


Time
-
consuming


Lots of mouse clicks!



Requires extensive musical knowledge


A British musicologist isn’t necessarily an
expert on music from China.

What is
jMusicMetaManager
?

[…] a tool for improving
metadata

associated with recordings by
automatically

detecting
inconsistencies

and
redundancies
.

Method


Calculates Levenshtein (edit) distance
between pairs of values (strings)


iTunes XML file or MP3 folder


Processing options vs Reporting options


Recursively run the software


Output can be fed to ACE or WEKA

Uses


Personal musical databases


Integration within OMEN (On
-
demand
Metadata Extraction Network) to clean up
metadata before submission


Libraries in general

An example


Subset of owner’s audio collection

Good!


Very easy deployment in Netbeans


Intuitive GUI


Detailed reports

Issues


False positives?


Problem with all metadata managers


Only one contributor


Doesn’t recognize all MP3 formats


Should (many) rap albums be
compilations?


Common to invite other artists (A feat. B & C)


Doesn’t support Cyrillic and East Asian

Possible Improvements


Support different languages


the = les (FR) = los/las (ES) = die (DE)


Change HTML report?


Apply changes to file names and tags


Fetch and edit metadata


Live versions not in clusters?

Thank you!

References

Corthaut, N., S. Govaerts, K. Verbert, and E. Duval. 2008. Connecting the dots: music
metadata generation, schemas and applications. In
Proceedings of the International
Conference on Music Information Retrieval. Philadelphia, USA. 249

54.

Datta, D. 2002. Managing metadata. In
Proceedings of the International Conference on Music
Information Retrieval. Paris, France. 249

51.


Hu, X., J. S. Downie, and A. F. Ehmann. 2006. Exploiting recommended usage metadata:
exploratory analyses. In
Proceedings of the International Conference on Music Information
Retrieval. Victoria, Canada. 67

72.

Kröger, P. 2004. CsoundXML: a meta
-
language in XML for sound synthesis. In
Proceedings of
the International Conference on Music Information Retrieval. Barcelona, Spain.

Lai, C., I. Fujinaga, D. Descheneau, M. Frishkopf, J. Riley, J. Hafner, and B. McMillan. 2007.
Metadata infrastructure for sound recordings. In
Proceedings of the International
Conference on Music Information Retrieval. Vienna, Austria. 157

8.

Mandel, M. I., and D. P. W. Ellis. 2007. A web
-
based game for collecting music metadata. In
Proceedings of the International Conference on Music Information Retrieval. Vienna,
Austria. 365

6.

References

McEnnis, D., C. McKay, and I. Fujinaga. 2006. Overview of OMEN. In
Proceedings of the
International Conference on Music Information Retrieval. Victoria, Canada. 7

12.

McKay, C., D. McEnnis, and I. Fujinaga. 2006. A large publicly accessible prototype audio
database for music research. In
Proceedings of the International Conference on Music
Information Retrieval. Victoria, Canada. 160

3.

Scheirer, E. D. 2002. About this business of metadata. In
Proceedings of the International
Conference on Music Information Retrieval. Paris, France. 252

4.

Selfridge
-
Field, E. 2006. Social cognition and melodic persistence: where metadata and
content diverge. In
Proceedings of the International Conference on Music Information
Retrieval. Victoria, Canada.