Talk:Musically relevant data: Challenges

On the importance of the quality of the meta-data associated to the audio (annotated corpus), this paper is very relevant: "Semiotic Structure Labeling of Music Pieces: Concepts, Methods and Annotation Conventions Frédéric Bimbot, Emmanuel Deruty, Gabriel Sargent, Emmanuel Vincent" http://ismir2012.ismir.net/event/papers/235-ismir-2012.pdf

On discussion about the reliability of largely-used corpus, these two papers are relevant: "An Analysis of the GTZAN Music Genre Dataset", Bob Sturm, ACM-Mumtiledia-MIRUM, 2012 "Ambiguity in Automatic Chord Transcription: Recognizing Major and Minor Chords" Antti Laaksonen, AMR, Denmark, 2012

Posted by Geoffroy Peeters on 16:38, 5 November 2012 (UTC)

Also this one on the importance of the quality of the meta-data "Creating Ground Truth for Audio Key Finding: When the Title Key May Not Be the Key", Ching-Hua Chuan1 and Elaine Chew http://ismir2012.ismir.net/event/papers/247-ismir-2012.pdf

Posted by Geoffroy Peeters on 16:44, 5 November 2012 (UTC)

Workshop on Crowd-sourcing at ACM-Multimedia 2012 http://crowdmm.org/

Posted by Geoffroy Peeters on 17:15, 5 November 2012 (UTC)

Geoffroy---maybe I misunderstand, but is that the paper about metadata quality you intended? I think the paper by you and Karën Fort, "Towards a (Better) Definition of the Description of Annotated MIR Corpora," is also highly relevant. http://ismir2012.ismir.net/event/papers/025-ismir-2012.pdf

Posted by Jordan Smith on 12:29, 20 November 2012 (UTC)