ISMIR 2004 Audio Description Contest

Pedro Cano, Emilia Gómez, Fabien Gouyon, Perfecto Herrera,
Markus Koppenberger, Beesuan Ong, Xavier Serra,
Sebastian Streich, Nicolas Wack
Music Technology Group, Universitat Pompeu Fabra

MTG-TR-2006-02
April 6, 2006


Abstract: In this paper we report on the ISMIR 2004 Audio Description
Contest. We first detail the contest organization, evaluation metrics, data
and infrastructure. We then provide the details and results of each contest
in turn. Published papers and algorithm source codes are given when
originally available. We finally discuss some aspects of these contests and
propose ways to organize future, improved, audio description contests.













This work is licenced under the Creative Commons
Attribution-NonCommercial-NoDerivs 2.5. To view a copy of
this licence, visit
http://creativecommons.org/licenses/by-nc-nd/2.5/ or send a
letter to Creative Commons, 559 Nathan Abbott Way,
Stanford, California 94305, USA.

MTG-TR-2006-02 1
ISMIR 2004 Audio Description Contest
Pedro Cano, Emilia Gómez, Fabien Gouyon, Perfecto Herrera, Markus Koppenberger,
Beesuan Ong, Xavier Serra, Sebastian Streich, Nicolas Wack
Universitat Pompeu Fabra, IUA, Music Technology Group
Ocata 1. 08003 Barcelona, Spain.

Abstract: In this paper we report on the ISMIR 2004 Audio Description Contest. We first detail the
contest organization, evaluation metrics, data and infrastructure. We then provide the details and
results of each contest in turn. Published papers and algorithm source codes are given when originally
available. We finally discuss some aspects of these contests and propose ways to organize future,
improved, audio description contests.

Music Information Retrieval (MIR) established itself in the last few years as a very active
multidisciplinary research field. This is clearly shown in the constantly growing number and subjects
of articles published in the Proceedings of the annual International Conference on Music Information
Retrieval (ISMIR, the first established international scientific forum for researchers involved in MIR)
and also in related conferences and scientific journals such as ACM Multimedia, IEEE ICME or
Wedelmusic, to name a few. The standardization of world-wide low-latency networks, the extensive
use of efficient search engines in everyday life, the continuously growing amount of multimedia
information (on the web, in broadcast data streams or in personal and professional databases) and the
rapid development of online music stores (as e.g. Apples iTunes, Walmart or MusicMatch ) set great
challenges to MIR researchers. Indeed, applications are manifold, from automated music analysis to
personalized music recommendation, online music access, query-based retrieval (e.g. “by-humming”,
“by-example”) and automatic playlist generation.
Among the vast number of disciplines and approaches to MIR (an overview of which can be
found in (Downie 2003)), automatic description of audio signals in terms of musically-meaningful
concepts plays an important role. As in other scientific endeavor, long-term improvements are