1. INTRODUCTION
We believe that research into musical information processing will
be advanced if music databases (DBs) become available that can be
used by various researchers. The main purposes and advantages of
such commonly available DBs can be summarized as follows:
• Researchers will be able to use the DBs as a common benchmark
for comparing and evaluating various methods related to
musical information processing. The lack of common music DBs
available worldwide for research purposes at almost no cost has
made it difficult to establish benchmarks (evaluation frameworks)
for much of the research done regarding musical information
processing.
• The DBs will accelerate the progress of various forms of research
using statistical methods. Recent progress in the use of statistical
methods in other research fields such as speech recognition has
been largely due to the availability of large DBs.
• Researchers will be able to use the DBs for research publication
and presentation without conventional copyright restrictions.
It has been difficult to demonstrate research using copyrighted
musical pieces that will be included in, for example, conference
videos and CD-ROMs.
Although there is an enormous amount of music available on commercially
distributed compact discs, it is difficult to use this music
for the above purposes because of copyright issues. Commonly
available DBs with copyright-cleared pieces are therefore essential
to encourage the healthy development of this field of research.
Various commonly available DBs have been built in other research
fields since the importance and significance of such DBs have
been widely recognized. In the research field of speech information