The images
The origin is a body of drawings made by hand. From those drawings, and from an archive of the same graphic language, a number of LoRA models were trained.
The work is made with Flux and Krea 1. LoRAs are combined with each other and with styles; within those styles the original drawings are combined again with visual references gathered over time — animals, Egyptian faces, material found intuitively and kept in a Pinterest archive used for research, not as a source.
The prompts are visual. A single combination can take days of testing, comparing and discarding. What is kept is a small part of what is made.
The films
The films are not generated or animated automatically. Selection, animation, cutting, rhythm, repetition, transitions and montage are done by hand.
The editing is self-taught, and the tools changed as the work went on: iMovie, then CapCut, now Final Cut Pro.
Authorship
It is in the first drawing, in what the models are trained on, in which references are chosen, in what is discarded, in what is processed again, in what is kept, in the order it is put in, and in how it finally becomes a film.
Sound
The music was created and developed with Suno, in a self-taught process of searching, generating versions, selecting, editing and setting the result against the images.
Format
Modular and adaptable: projection, installation, mapping, or a live collaboration. The stable core is an audiovisual piece of approximately 45 minutes.