Our great sponsors
-
InfluxDB
Power Real-Time Data Analytics at Scale. Get real-time insights from all types of time series data with InfluxDB. Ingest, query, and analyze billions of data points in real-time with unbounded cardinality.
Using Minecraft GAN: https://boredhumans.com/minecraft.php (made with https://github.com/jeffheaton/pretrained-gan-minecraft) . Unlike #2 and #3 below, the images were AI-generated but not based on my lyrics.
Using Story2Hallucination: https://boredhumans.com/music_videos/Story2Hallucination_withwords.mp4 (made with https://github.com/lots-of-things/Story2Hallucination). The problem with it was that I had no easy way to match the timing of the lyrics on the screen with the real singing. So I then made a new version at https://boredhumans.com/music_videos/Story2Hallucination_nowords.mp4 where I edited the code so it did not show the words at all. But it is somewhat boring to watch.
Using Deep Music Visualizer - https://boredhumans.com/music_videos/Big_GAN.mp4 (I made it small as a fast demo but it could easily be normal sized) . Made with https://github.com/msieg/deep-music-visualizer. Although the output is good, it uses 12 random classes (or you can pick them manually) of items from ImageNet, and the 1000 classes ImageNet uses really suck for making music videos. They don't have general categories such as beach, ocean, stars. Instead it is very specific things like moving van, drilling platform, and envelope.