In honor of Bicycle Day and the work of  Albert Hofmann Trippin Thru Town is a video art series inspired by his unexpected ride.

The NFT space is flooded with lazy style transfer pieces, but I’m still fascinated by what these deep learning applications can generate, particularly with a little coaching and encouragement.

For this series I wanted to try Ebsynth again, now that I’ve done a few doodles with it and learned what it needs. I learned even more with Trippin Thru Town that will produce clearer images in the next project. The best frames generated in all 4 pieces are in the middle and the further Ebsynth gets from the frames I hand painted, the less precise the shading and details become. It worked out marvelously for a loop inspired by a psychedelic bicycle ride.

I’m only able to scale and crop the very low resolution source up to 4K-ish resolution through the power of DaVinci Resolve’s node magic.

This is the original footage:

The key to quality rendered frames I’ve discovered is to create input keyframes that have the texture that you want. I only hand textured one frame from the center for Trippin Thru Town. For other projects I’ll probably feed Ebsynth more keyframes.

This is how the Watercolor Wine edition turned out:

Viewable in 4K on YouTube:

---

I’m only releasing 3 editions of each this year and I plan to only release another of each on Bicycle Day each year until they are all published.

https://www.hicetnunc.xyz/objkt/39929 Watercolor Wine

https://www.hicetnunc.xyz/objkt/39920 Electric Fire

https://www.hicetnunc.xyz/objkt/39945 Eggplant Juice

https://www.hicetnunc.xyz/objkt/39422 Neon Melon

---