Variation II of Variation II

Kinetic Score/Audio-Visual Installation


work in progress

In this piece, I use Cage’s Variations II, which a compositional and notational model that John Cage developed over the period from 1958 to 1961, as an audio-visual generative method. In Variations II, Cage's original notation consisted of five points and six lines on eleven individual transparent plastic sheets. A composer/performer superimposed the transparent sheets and measured the distance between the dots and lines that represented parameter of musical events and sounds. In other words, the contingent relationship between the lines and dots determined the parameter of musical events and timbres. In my interpretation of Variations II , I apply the score to a kinetic installation. The lines and the dots are replaced by computer-controlled moving panels and visitors’ bodies, respectively. The algorithms analyze the relationship between these components and generate musical parameters in real time.

Computer vision dissects the movements of the audience and transforms them into hallucinated images which are analyzed, enhanced, and synthesized by Neural Network methods. Moving panels are juxtaposed with the walls of a gallery space and display what the computer vision and AI understand, analyze, and synthesize through interaction with the audience. This may include data such as the visitors gender, height, or age, as well as the process of analysis and synthesis from the perspective of computer vision and AI.

The contingent relationship between the moving panels and movements of the audience body determines musical events, timbres, and movements of the images. Through this process, the audience members explore new physical movements of their bodies. Furthermore, I expect the audience to raise questions, such as how much of the creativity is derived from the computer program and how much from the programmer, what it would be like from the computer’s perspective to be an artist, what creativity is in our digital age, or how AI bias actually happens.


work in-progress

 © 2019 by Jinku