an anthology of impossible artists | 2022
10-Track Digital Album
A Recurrent Neural Network model (sampleRNN) was trained for each short piece using inconsistent training-data (prism-samplernn implementation by Dr. Christopher Melen).
Each training corpus contains two "incompatible" positions (aesthetically, ideologically, historically or stylistically as you like). Artist names indicate data sets origin or style.
Purely computer generated, no pre- or post-processing, just curated and assembled.
Made for the festival Bowed Electrons @ University of Cape Town
3rd-7th Sept. 2022
“The idea becomes a machine that makes the art”. (Sol LeWitt)