Content-Based Indexing of Symbolic Music Documents - Intelligent Music Information Systems: Tools and Methodologies - page 3

not be directly employed for other media. On the
other hand, these specific techniques should be
integrated whenever different media are present in
an individual item. The core information retrieval
(IR) techniques based on statistics and probability
theory may be more generally employed outside
the textual case and within specific nontextual
application domains, like music. This is because
the underlying models, such as the vector-space
and the probabilistic models, are likely to describe
fundamental characteristics being shared by dif-
ferent media, languages and application domains
(Sparck Jones & Willett, 1997).
The requirement for a music content-based IR
has been stressed, since many years, within the
research area of music information systems as
well. The developments in the representation of
music “suggest a need for an information retrieval
philosophy directed toward non-text searching and
eventual expansion to a system that encompasses
the full range of information found in multime-
dia documents”, as stressed by McLane (1996).
As IR has dealt with the representation and the
disclosure of content from its early days (van
Rijsbergen, 1979), it is natural to consider that
IR techniques should be investigated to evaluate
their application to music retrieval. By conclud-
ing his survey, McLane stressed that “what has
been left out of this discussion, and will no doubt
be a topic for future study, is the potential for
applying some of the standard principles of text
information retrieval to music representations”.
Since 1996, many approaches have been applied
to music access, browsing, retrieval, personaliza-
tion, both proposing original techniques tailored
to the music domain and adapting IR techniques
(Downie, 2003). Many approaches to music re-
trieval are related to the field of digital libraries
(Bainbridge, Nevill-Manning, Witten, Smith &
Mc-Nab, 1999; Agosti, Bombi, Melucci & Mian,
2000). Because of their multimedia and multi-
disciplinary nature, digital libraries may profit
from results in music indexing and retrieval.
In particular, projects on the preservation and

dissemination of cultural heritage can be the
result of the combination of digital library and
information indexing techniques (Ferrari & Haus,
1999). Yet most of the projects involving digital
libraries, such as Harmonica (2006), are still based
on bibliographic values rather than on indexing
document contents, meaning that research on
content-based approaches is required.

Indexing

One of the main components of an IR system is
indexing (Baeza-Yates & Ribeiro-Neto, 1999). In-
dexing can be defined as “the process of analyzing
the informational content of records of knowledge
and expressing the informational content in the
language of the indexing system. It involves: (1)
selecting indexable concepts in a document; and
(2) expressing these concepts in the language of
the indexing system (as indexing items)” (Borko
& Bernier, 1978). An indexing system is composed
of a number of automatic procedures that allows
for the organization of document contents, and for
their access, retrieval and dissemination.
Indexes are used as guidance towards items in
a collection of documents. In particular, the fact
that indexes can be ordered, stored in complex
data structures, and accessed with fast techniques
such as hashing functions or tree searches, allows
for efficient retrieval of documents in a collection.
The effectiveness of indexes is part of everyday
life, when looking up a dictionary or looking for
the content of a topic through its list of relevant
names and concepts (which is precisely called
“index”). Because it allows fast access to a syn-
thetic description of documents content, index-
ing allows for the scalability of an IR system.
Efficient data structures, such as inverted files
(Baeza-Yates & Ribeiro-Neto, 1999) have been
proposed to connect the indexes—which are used
for retrieval—to the documents—which are of
interested for the user.
Many approaches to music retrieval are based
on online searches, where the user's query is