art

ComfyUI and AI as an Art form

ComfyUI

So, I’ve been resisting the urge to rant about how AI art is really blossoming into its own medium, and all of the alarmists out there are the equivalent of people ranting against the invention of the printing press back in the day — mostly because I don’t think I could maintain a polite tone.

Deal. With. It.

No, AI is never going to replace your experience as a human. I can’t even fathom what makes people think that’s possible …

So if you want to create art with pencils, brushes, or photoshop, you will still be able to create art with pencils, brushes, or photoshop. That’s not going anywhere, and plenty of people will still happily watch youtube art videos. Face it, art has never been a lucrative career path for the huge majority of artists throughout all of history — that’s why “starving artist” is a stereotype.

Me? I like creating art with words and AI.

The above screenshot is the user interface I use for Stable Diffusion, which, as you can see, is far more complicated than typing in, “magical Autumn forest”.

All of those settings in all of those boxes mean something — they are the difference between a beautiful picture and something totally broken — and ComfyUI lets you go even deeper than that.

I’ve spent the last six months learning how to write prompts, because the order of the words and even the punctuation in between them matters. I’ve also been learning what all of the settings do. Most of this only has a little bit of documentation because of how brand new the technology is, so the majority of my knowledge has come from trial and error.

Fails include:

And pretty soon here, SDXL 1.0 will be released and I’ll probably end up re-learning it all over again, ha ha.

It’s great though. I thoroughly enjoy AI art and the process of creating it. Not so thrilled with how many people are reacting negatively to a new technology, but go figure — people freaked out over every single technological advancement in all of history, and it didn’t stop the world from changing anyway.

art

AI Art – Alien Worlds

Stable Diffusion XL is the hottest thing hitting the AI art world at the moment, and they still haven’t released the official 1.0 version to the public yet.

So, aside from a fancy new Stable Diffusion model, what’s different about these pictures?

Multi-phase generation.

Typically you type in a prompt, and you get an image. But with this updated format, you type in one prompt to generate an idea of an image, and a different prompt to fine tune it. The results are AMAZING.

art

AI art – Designs

Playing around with Stable Diffusion’s capabilities.

art

AI art – Setting the scene

I’ve mentioned before that I think very heavily in words, while my visual imagination tends to be fairly vague. With writing, this means that I focus more on the pacing and sound of words of the story, while glossing over visual descriptions.

But, with the availability of Stable Diffusion now, I’ve realized that I can describe a setting, and the AI will fill in the little details for me.

At some point in The Scions, Carol and Hartmann go hiking, and I wanted to properly describe the landscape to really set the scene. After telling Stable Diffusion what I needed, it gave me this picture:

Perfect!

Instead of trying to visually imagine it in my head, I had a much easier time describing the picture that I was actually looking at. The nice thing was that I didn’t have to slog through stock photos trying to find something that kinda sorta fits but not really because why can’t the most obvious pictures just exist when I’m not asking for anything special … you know. 😉

I think this will really benefit me as a writer.

Then from there, of course, I added more keywords to see what sorts of pretty effects I could achieve. Lol.

art

The Stable Diffusion Rabbit Hole

I’m thoroughly terrible at documentation, so I didn’t write a single thing down. Whoot whoot, go me!

At this point, my husband has decided that we need to get Stable Diffusion set up at home, so we can make lots of pretty pictures without spending “credits” through host sites. Also because we’re the sort of nerds that enjoy making minute changes to see what effect it has, and that’s very difficult when one is working with limited numbers.

So, with this round of learning, I took the same prompt, “Yellow rose in a glass vase. Centered composition,” and changed the art style. I do remember that one of them I put down “Lisa Frank,” but I bet that you can’t guess which one. Another one was, “Magic realism.”

Also, since I used the Stable Diffusion Demo for these, I had no control over the seed or any other technical settings.

I need to get into the habit of documenting.