New sound reactive experiment is live on my site! Some background, a little over a decade ago I created an exhibition where I used sound reactive graphics created via Xcode that could be manipulated via several digital and physical MIDI controllers. That project sparked now a lifelong fascination with sound and form that I love exploring, which led to the latest web-based experiment. Check it out at: https://thehonestape.com/play/instrument
1
11
ComfyUI is one of my favorite things to experiment with. Together with some other open source things from Hugging Face, you can do a lot of cool things with your local machine. This is a collection of things created experimenting with animated diff workflows.
1
1
56
Sometimes I look back at older outputs or "errors" from Stable Diffusion and find that they are more interesting than "better" results using the latest models and hardware. I think it may be because the interpretation of the the images still leaves a lot to be figured out by the human mind.
This batch is all from October 2023, using M1 Mac and Stable Diffusion Web UI by Automatic1111.
1
46
Today's meditation on writing resume's and cover letters for design teaching positions and thinking about raising the dead.
0
31
Some recent meditations on form and color. These are illustrations that I make freeform in Illustrator as a way to decompress. Just put on some music and go at it with boxes and colors.
I also recently read "Libre Novis" by Carl Jung and realized these are probably mandalas, but I have not delved deep at all to interpret any meaning. Hope everyone is well.
1
3
52
More experiments with Krea's real-time model. These are all direct prompts using modern scientific terminology to describe the features of the medieval imagery that I used to train this style. A lot of the source images are symmetrical (medieval "sacred geometry"), deconstructing them, then piecing them back together via AI with language, produces some interesting and beautiful images.
12
118
Experimenting with a generative poster design. Typography was set beforehand in Illustrator, real-time model trained on Krea.AI (https://Krea.AI). None of them has the composition I was envisioning yet; I probably need to do more actual designing with hands and fingers for a bit to get the composition I wanted.
Some interesting things to look at and study for the next round. Probably need to improve the prompt and provide a better initial composition, but the medieval calligraphic style is working well.
Anyone else here using Krea?