Information Technology Reference
In-Depth Information
These patterns have a length less than or equal
to 4 notes and possess at most 21 occurrences.
Owens established a hierarchically classified
lexicon of 193 patterns from his corpus (Owens,
1974). A large number of Owens' patterns were
satisfactorily extracted by the system, in par-
ticular patterns designated by 11A, 4Ea and 5B
in Owens' nomenclature. Moreover, a number of
new patterns, not signalled by Owens, have been
effectively “discovered” by the system, and are
considered by Rolland as “musically meaningful”
(p. 348). Significantly different sets of musical
descriptions lead to the extraction of significantly
different patterns. In particular, the use of too-
restrictive sets leads to the extraction of patterns
without any musical pertinence.
Meredith et al. approach (2002) has been il-
lustrated with one single practical example: the
analysis of the beginning of Bach's two-part
Invention in C major
(BWV 772), where only
one pattern is presented in the paper, composed
of the seven eighth notes forming the entry of
the main theme.
results obtained using our
proposed approach
Our own approach can be applied for the moment
to the analysis of simple melodies. The model has
been tested using different musical sequences
taken from several musical genres (classical music,
pop, jazz, etc.), and featuring various level of com-
plexity. The experiment has been run using version
0.8 of our algorithm, called
kanthus
. This section
presents the analysis of two pieces: a medieval
Geisslerlied
and a Bach
Invention
. A large set of
Figure 16. Analysis of the Geisslerlied “Maria muoter reinû maît” (Ruwet, 1987). Each motivic repeti-
tion is represented by a line below the corresponding stave, each labelled with a letter.
Search WWH ::
Custom Search