Patrick Hartono is an electroacoustic composer, audiovisual artist, researcher, and currently a Lecturer in Computational Arts at Goldsmiths, University of London. In this Roundtable, Patrick will be sharing his recent research project on neural network-driven parametric control for generative visual systems, where AI functions not as an image generator but as a real-time controller of procedural visuals. Using TouchDesigner with a custom machine learning model trained on motion capture data from Lengger dancer Rianto, the system interprets the movement data and maps it to the generative process, creating a fluid, dynamic visual representation of the dance’s expressive motion. Lengger, a traditional Indonesian dance, is known for its fluid representation of gender, where male dancers embody feminine roles while imagining an intangible, non-binary presence within the performance. This research seeks to translate this concept into a dynamic generative visual system that visually reflects the fluidity and expressiveness of the dance.
