Maybe Arrive Musical Metacreation Concert, MUME 2018 at the Ninth International Conference on Computational Creativity in Salamanca, Spain, June 2018.
Anticipation and arrival, or lack of arrival, nested within other anticipations and outcomes, forms a self-similar landscape of rhythmic building blocks. In this piece, bass lines and melodies are parsed into these rhythmic building blocks and then regenerated from nearby points on that landscape, steered by the performer in real time. The geometry of the GUI affords navigation and shaping of music at a relatively subjective level, via simple, nested operations on patterns of rhythmic expectation.
These algorithms are implemented by custom macOS apps (Coord) that analyze and generate MIDI note patterns in Ableton Live. One app directly manipulates individual Live clips and the other app morphs between selected Live clips. Remaining elements of the Live set are also under the control of the performer but unaffected by the algorithms.
The composer and performer of the piece is Harun Gezici. He has been producing different styles of electronic music for more than 15 years as LowNoiz, and he is founder of netlabel Rauscharm Recordings at www.rauscharm.com. His music can be found at lownoiz.bandcamp.com and soundcloud.com/ultratech. He resides in Ennetbaden, Switzerland.
Harun Gezici has been producing deep minimal house and atmospheric Techno for more than 10 years as Ultratech. In this piece he weaves algorithmic, real-time musical variation and hybridization into the context of a larger musical production/performance, incorporating algorithmic music software designed by Jay Hardesty and outboard analog gear. Here, music algorithms are not intended to create material that sounds generative or experimental; instead the aim is to harness, splice and steer generative structure that already exists in a familiar genre (EDM).
Here the composer’s skills, intuition and judgment are leveraged by algorithms that act as a sort of improvisational coprocessor, interactively refactoring local, low-level rhythmic and melodic structure. There is no attempt to model or replace compositional strategies or to algorithmically evaluate the results. The software complements the composer/performer in a manner similar to the relationship between a driver and a car’s dashboard and navigation system. (That is, a self-driving car would be exactly the wrong metaphor.)
Particular melodic elements are morphed to generate a spectrum of variations, which are explored by traversing an interactive landscape where selected melodic inputs serve as musical landmarks.
Certain other parts are algorithmically recombined in a coarse- grained manner, governed by patterns that have unpredictable sequences, governed also by simple relations that derive from a tiny set of rules.
But the main shape and content of the piece remain in the hands of the composer/performer. The overall form and many elements within the piece are completely outside the reach of the software. The aim is a piece of music that sounds like music without any algorithmic component, but which has organic “moving parts”.