A Spanish artist has created a trippy music video by training a deep learning algorithm on thousands of album covers.
Bruno López produced the video using a combination of Spotify data, Python scripting and Generative Adversarial Networks.
He first created a script that uploaded album art for each track on Spotify’s official editorial playlists.
The result was a dataset of about 50,000 covers with individual resolutions of 640 x 640.
This was used as training data for Nvidia’s StyleGAN2 architecture, the same system that previously brought us pizza, fursonas, and My Little Pony characters that don’t exist.
After several days of training, the model became able to generate her own album covers.
[Read: How do you build a pet-friendly gadget? We asked experts and animal owners]
Next, López created a Python script to organize the results according to the colors and faces in the covers. He then animated the images to the beat of the song “Interference” by Sound Stabs.
Finally, he edited the clips in Adobe Premiere to produce the final video.
While the music is too thrilling for my aging ears, the animation fits the beat perfectly and is more aesthetically interesting than many of the most popular YouTube music videos.
Published March 8, 2021 – 20:26 UTC