Sunteți pe pagina 1din 2

Algorithmic Composition Using a Computational Model of

Music
jur.byu.edu /
admin
Morgan Quigley and Dr. Steve Ricks, School of Music
Algorithmic composition involves the application of systematic mathematical methods to the compositional process.
Sequences of pitches, rhythms, or other musical elements are manipulated as if they were abstract mathematical
symbols, and the resultant music contains patterns that can either be simple enough to recognize audibly or so complex
that they are revealed only through detailed score analysis. To various extents, algorithmic compositional techniques
have been employed since at least the late medieval times, where composers wrote melodies on top of fixed, repetitive
rhythmic phrases called isorhythms. Modern composers have developed the technique of chance music, where
pitches or rhythms are randomly chosen from a collection of source material to form compositions which are only
partly determined by the composer. Traditionally, compositions that use algorithmic techniques such as those just
mentioned were created using time-consuming paper-and-pencil methods. Since the era of the personal computer, many
software packages have emerged that automate the creation of algorithmic compositions. However, these packages
have tended to be very complicated, requiring the composer to learn some sort of computer programming language in
order to use the software. My goal in this project was to write a program that allowed composers who are not computer
programmers to easily create algorithmic musical textures. Another design goal was to not completely disrupt the
traditional compositional process by allowing composers to mix computer-assisted compositional techniques with
traditional hand-editing of musical scores. The first stage in this project consisted of creating a computational model of
the compositional process that would enable rapid generation of algorithmic musical textures. The following paragraphs
describe the hierarchical structure of my model from the bottom up.

An element is the basic and fundamental unit of the compositional model. There are various types of elements in my
model which roughly correspond to several of the traditional elements of music taught in music theory courses, namely
pitch, rhythm, and dynamics. Pitch elements can take the form of literal pitch locations (such as middle-C), pitch classes
(such as D# in any octave), chords (such as a C-major triad, C-E-G), or rests. Rhythm elements can specify a
traditional duration, such as an eighth note or a dotted quarter note, or they can be a composite duration whose length
is a multiple of a traditional note value, such as five quarter notes or nine sixteenth notes. Dynamic elements specify a
traditional Italian dynamic marking, such as forte or mezzo piano.

A phrase is a collection of elements, a method to expand them into a final sequence, and the length of the desired
sequence. Phrases can be expanded sequentially, in which case the elements are merely repeated in the order they
were defined until the desired phrase length is achieved. Phrases can also be expanded randomly, achieving the
desired phrase length simply by choosing elements at random. In either case, if a phrase contains a reference to
another phrase, the referenced phrase is expanded recursively to form a sub-expansion, which eventually will form a
sequence of elements which are copied into the expansion of the original phrase.

A track is formed by combining a phrase of pitches, a phrase of rhythms, and a phrase of dynamics. Each track has a
desired length, which is used at expansion time as the desired length of its pitch, rhythm, and dynamic phrases. Once
these three elemental phrases are expanded to the desired length, the final track expansion is formed by combining
the results of the pitch, rhythm, and dynamic expansions to form a sequence of complete musical notes.

A session is a collection of tracks and sets, which are simply phrases that can be referenced by name in anywhere in
the composition. Organizing the compositional session in this way allows for multiple-voice compositions as well as
repetitive thematic ideas one or more voices. For example, a loud-soft-soft dynamic pattern may be defined as LSS
and referenced in the dynamic phrases of any of the tracks of the composition.

The example at right shows several of the features of this program. The pitches of the upper voice are a random
ordering of Db, Eb, E, F#, Ab, and Bb. The rhythm of the upper
voice is straight eighth notes, and its dynamics follow a repetitive
pattern of loud-soft-soft. The pitches of the lower voice are a
random ordering of C, D, A, and Bb, and a sequential phrase of
F-E-Eb. This three-note descending pattern occurs three times in
the example. The lower voice shows a repetitive rhythmic pattern
(an isorhythm) of two-eighths, quarter, eighth, two-sixteenths.

As the previous definitions and example have shown, my


compositional model has attempted to capture the two main phases
of algorithmic composition: the creation of source material, and
the expansion of the source material into a usable musical
segment through application of a systematic mathematical process.
To implement these ideas and experiment with their effectiveness,
Dr. Ricks and I designed a simple interface that allows for the
creation of complex algorithmic music without requiring the user to
write any computer code. Composers enter source pitches,
rhythms, dynamics, and sets into the left side of the window, push
the large button in the middle of the window to expand the source
material into a final sequence, and listen or export the output using
the controls in the right side of the window. We called this program
the Universal Music Machine, and I wrote over ten thousand lines of C++ computer code to create it for both Windows
and Macintosh OS X v.10.2. Music playback is achieved using the native MIDI sequencer provided by the operating
system, and composers can export their algorithmic sequences in the standard MIDI format for hand-editing in score
editors such as Finale or Sibelius. Sessions can be saved and loaded in a cross-platform XML format. Dr. Ricks and I
made this program freely available on the Internet, and thousands of musicians across the globe have downloaded it.
Composers from Australia to the Netherlands have sent emails offering suggestions for the improvement and refinement
of the program, and I have periodically been implementing their ideas. At time of writing, the program is available at the
following URL: http://www.universalmusicmachine.com

S-ar putea să vă placă și