We have been conducting research to develop a music information retrieval system which retrieves music based on user preferences. In order to conduct evaluation experiments, it is necessary to accumulate an experiment set of music data, and collect user ratings of each music data included in this data set. Two music data sets were prepared for user evaluation: the RWC Popular Music Database, which is a collection of 100 copyright-free pop songs, and a collection of songs recorded in on-the-market CDs. Comparison of user ratings towards the songs in each data set shows that user ratings of the RWC songs were far lower than that of the songs in the other music data set. Based on this experience, the artists believe that the construction of a standard list of on-the-market CDs is necessary for the development of an entertaining MIR system.
[1]
Keiichiro Hoashi,et al.
Personalization of user profiles for content-based music retrieval based on relevance feedback
,
2003,
ACM Multimedia.
[2]
Keiichiro Hoashi,et al.
Implementation of relevance feedback for content-based music retrieval based on user prefences
,
2002,
SIGIR '02.
[3]
Jonathan Foote,et al.
Content-based retrieval of music and audio
,
1997,
Other Conferences.
[4]
Masataka Goto,et al.
RWC Music Database: Popular, Classical and Jazz Music Databases
,
2002,
ISMIR.
[5]
Gerard Salton,et al.
The SMART Retrieval System—Experiments in Automatic Document Processing
,
1971
.