Post-apocalyptic neural audio synthesis
I have been experimenting with RAVE a lot lately, and more specificly with its real-time audio generation abilities (coming soon !). This time I used a dataset of songs that can be heard in the Fallout franchise, and created a small max/msp patch to unconditionally generate audio in real-time!
For the artwork I have used a clip guided diffusion model with the prompt a forgotten radio, which produced those nice images !