Tessera
Generative visuals that listen to music.
Tessera creates abstract visuals by listening to a track the way a musician does. It picks up loudness, attacks, the weight of bass and treble, harmony, and a sense of building tension. Each of these becomes a control for a generative scene, so the music decides how forms move, melt and change color. Every frame is a function of musical time, so any moment can be rendered as video.
Now I’m making it run in real time and respond to MIDI, so it can be played during live shows, for example with my new electronic music duo Particle Flow, which will release music soon. On the AI side, I’m exploring fine-tuning multimodal models to link music understanding and abstract visual generation directly in embedding space.
No music here yet, so you play it: come closer to the form, drag across it, scroll, click.