Music composed by Genetic Algorithm
Dead Slow, which consists of a continuous overlapping harmonic progression, was generated using a harmonic analysis of 87 compositions by Pat Metheny, and a third-order Markov model based upon this analysis. Durations, dynamics, playing style, range, and harmonic spread were determined using patterns generated by a genetic algorithm. These continuous harmonies are interrupted by Look Left – contrapuntal sections that interpret tendency masks that define such parameters as sequence length, number of instruments, subdivisions, playing style, number of playing styles, dynamics, and the number of gestures in a section. In these sections, each instrument operates independently within its own melodic, rhythmic, and tempo plane. The title comes from traffic signs observed while in England, where this piece was conceived.
Performed by the Yaletown String Quartet: Mark Ferris (violin), Cameron Wilson (violin), Henry Lee (viola), Finn Manniche (cello), with Daniel Tones (percussion).
One of the Above (2011)
Four movements for solo percussion
These movements were created using evolutionary algorithms to develop a population of rhythmic motives, initially generated through probabilities. Each population is a single rhythmic phrase, and can last from two to over twenty beats. Once a series of generations (up to fifty) were derived, a genetic operator was run to determine the best ordering and repetition of individual populations, while another genetic operator was run to combine divergent populations. The evolution of the individual rhythmic ideas are clearly heard as they evolve, mutate, and combine to form new individuals.
Movement 1 is the “baseline” – composed by myself without the use of the computer, whereas the other three are entirely machine-composed. The purpose is to compare my “real” compositional aesthetic (when I compose by hand) with my metacreative system (which is a codification of my personal compositional aesthetic).
Performed by Daniel Tones
Armar
for four percussionists
i. Guaguanco
ii. Guiro
iii. Mambo
iv. Mozambique
(Spanish: To Assemble; put together) was composed by Kinetic Engine, a real-time software system created by the composer that uses multi-agents to create complex polyphonic ensemble rhythms. Armar also uses a genetic algorithm to evolve a population of rhythms that are derived from an analysis of source material, in this case, Afro-Cuban music. It then used these rules to generate its initial population. Each “parent” rhythm was then analysed again to generate new rules for subsequent children, all during performance. Each movement was “performed” live in the composer’s studio, and the output was recorded in a music notation program, with only minor editing by the composer.
Performed by Brian Nesselroad, Daniel Tones, Martin Fisk, Timothy van Cleave (percussion)
Other, Previously (2009)
for string quartet
The software was given 16 measures of the traditional Javanese ensemble composition Ladrang Wilugeng, which it analyzed to derive musical rules. These rules were used by a genetic operator to create a population of ever- evolving melodies and rhythms that the system reassembled in a multi- agent environment over a rotating harmonic field. The end result is a piece of music that reflects many of the tendencies of the original, without direct quotation. The composerʼs role was limited to dynamic markings, orchestration, and assembling sections. The title comes from the phrase “memories, forgotten”, which was run through Google Translate many times across various unrelated languages.
Members
Research papers
Eigenfeldt, Arne. "Corpus-based recombinant composition using a genetic algorithm." Soft Computing 16.12 (2012): 2049-2056.
Eigenfeldt, Arne, and Philippe Pasquier. "Populations of populations: composing with multiple evolutionary algorithms." Evolutionary and Biologically Inspired Music, Sound, Art and Design. Springer Berlin Heidelberg, 2012. 72-83.
Eigenfeldt, Arne. "The evolution of evolutionary software: intelligent rhythm generation in Kinetic Engine." Applications of Evolutionary Computing. Springer Berlin Heidelberg, 2009. 498-507.
Eigenfeldt, Arne, and Philippe Pasquier. "Evolving structures for electronic dance music." Proceeding of the fifteenth annual conference on Genetic and evolutionary computation conference. ACM, 2013, 319–329.