All the images were generated using a custom StyleGAN2-ada implementation trained on a handcrafted dataset of 1200 nebulae and abstract pictures. The only post-processing applied was some standard color correction and a little bit of image sharpening. You can download the PNG images at full resolution here. Extra: I've generated 20 additional artworks and turned them into NFTs just for the lols. You can check them out on Opensea.

Music video

This audio-reactive video is the result of sequentially interpolating the latent vectors of 370 unique generated samples. The procedure is straightforward: different audio tracks (s.a. the kickdrum) are mapped to different vectors that get triggered whenever they sound, which makes the generated images smoothly react to the music. You can find more details about the implementation on my Twitter.

Streaming platforms

The track is also available on Spotify and all other streaming platforms.