Abstract
Content-based search of music collections presents differing challenges at different scales and according to the task at hand. In this paper, we consider a number of different use cases for content-based similarity search, at scales ranging between a detailed investigation of a single track to searching for fragments of a track against a collection of millions of media items. We pay particular attention to the varying tradeoff between precision and recall in these contexts, both from the point of view of system evaluation and from the point of view of a user of a system searching an unknown collection. We present the audioDB software for content-based search, and describe how it has been used to address use cases across these different collection sizes; in addition we show that the interpretation of similarity as a distance which can be modelled statistically, initially motivated by our desire to achieve sublinear retrieval time on large databases, can be used to improve the precision of searches over small and medium-sized collections.
Acknowledgements
We are grateful to Daniel Leech-Wilkinson, Andrew Hallifax and Martin Haskell for providing files from the Kings Sound Archive collection, and to Denzyl Feigelson for providing media from the Artists Without a Label collection. Some of the larger-scale investigations we describe were performed with Malcolm Slaney on the media collections at Yahoo! Inc. We have benefited greatly from conversations with Michael Jewell, and we thank the anonymous reviewers for their perceptive comments and constructive suggestions. This work was supported by EPSRC (grant number EP/E02274X/1).
Notes
4such as the ‘googlewhack’: searching for particular terms which have low absolute numbers of matches in the database.
5The Culverhouse Collection http://www.filmandsound. ac.uk/collections/culverhouse.shtml
7Retrieved 21 January 2010 from http://www.iprospect.com/premiumPDFs/iProspectSurveyComplete.pdf
9An alternative, concatenating multiple-side recordings into complete movements—as on a modern CD re-issue—was not done for reasons of time.
10Audio file available at http://images.cch.kcl.ac.uk/charm/liv/audio/flac/DB_1010_Cc_9888-2.flac