|
General information
Important dates
Submission information
Conference program
Detailed program
Keynote speakers
Special
Metadata Session
Accepted
papers
Accepted
posters/short papers
Tutorials
Panels
Professional visit
Exhibits
Social events
Registration
Conference venue
Conference
committee
Photos from the conference
Conference proceedings
Mailing-list
Hosted
by
Sponsors
and partners
|
ISMIR
2002
3rd International Conference on
Music Information Retrieval
IRCAM – Centre
Pompidou
Paris, France
October 13-17, 2002
This page now includes a link
(through the ►icon) to the full text of the final version of the
conference papers. These texts are © IRCAM – Centre Pompidou in this form.
Authors have retained the rights to their original texts.
The printed
version of the proceedings, which also contains introductions, the abstracts,
a table of contents and the author index, can be ordered here.
SPECIAL KEYNOTE SESSION
ON METADATA
This special session, consisting of invited distinguished
speakers, will discuss the importance and variety of metadata for music
information retrieval. Whether derived automatically from contents (audio,
score...) by extracting a variety of features (samples, notes, melody, form,
lyrics...), or produced manually by specialists (e.g., as in libraries),
supplemented with such mechanisms as authority control, and put in relation
with musical documents and documentation, very powerful music searching tools
may be developed. This session will also address the issues of creating,
organizing, managing and using potentially very large collections of
metadata, including the technical and economical/marketing aspects.
This session will be held on the afternoon of Wednesday, October
16, 2002, and will be open to Résonances visitors. The participants and the
topics of their talks are:
“Beating Babel -
Identification, Metadata and Rights”
The development of rights trading for
music over networks depends, crucially, on persistent identification and the
availability of machine readable descriptive and rights metadata that is
interoperable. However, with the proliferation of metadata schemas and
identification systems, the future is potentially very confusing.
The presentation describes an approach
to this subject, beginning with the <indecs> analysis, developed as
part of the Info2000 programme. Now widely accepted in the content
industries, the <indecs> analysis has informed a number of initiatives,
including the MPEG-21 Rights Data Dictionary and the GRID project (the Global
Release Identifier) for the release of music on-line.
The presentation also describes the work being
carried out by the Contecs:DD Consortium, based on the original
<indecs> work, which has been adopted as the baseline technology for
the MPEG-21 Rights Data Dictionary, currently at Committee Draft stage.
“Technology
and art – putting things in context”
A set of standard
technologies already developed or under development by the
Moving Picture Experts Group (MPEG) promises to bring back to authors the
control of their works. The technologies are those of Content Representation,
Digital Item Declaration, Interoperable Intellectual Property Management and
Protection, and Metadata. ►
“Managing Metadata”
The All Media Guide (AMG) is a technology
company that maintains the world’s largest database of metadata relating to
the entertainment industries. This document describes some of the goals
of AMG, the issues uncovered during the evolution of our databases, and
discusses some of the implementations we have chosen. ►
“Why Not MARC?”
Traditional library cataloging records in the United States, based
on AACR2R cataloging rules and MARC standards, constitute a solid foundation
for many of the descriptive metadata elements needed for searching and
retrieving works of music. However,
there are significant weaknesses associated with these records and the online
environment in which they live as users seek access to digitized
representations of music. While music
metadata in the library catalog records offer less than a perfect solution,
they can and should have an important role in the total solution. Variations2, the Indiana University
Digital Music Library, builds on the advantages of AACR2R and MARC and offers
a domain-specific data model and search environment that address many of
the identified problems. ►
“About this Business
of Metadata”
A brief discussion presents some
of the opportunities and challenges involved with creating metadata-centric
businesses that bring Music Information Retrieval technologies to the
marketplace. In particular, two
related difficulties -- that of the difficulty of proving incremental value
for new metadata systems, and that of the relative influidity of the
marketplace for MIR -- are highlighted.
Potential directions for resolving these issues are also discussed. ►
|