The conflicted expression was intentional, yes. In addition to the long term worries you mentioned, she also has the immediate concern of being the center of attention while in such a state.
For sure, the AI faces all have the eyes looking at something because that's what's in the training set. But for a 'surprise from behind' expression, you want the eyes staring straight forward not focused on anything. Also the teeth clenched "eee" mouth shape is more startled, where the "o" mouth is more surprised, but it just can't do the more emotive facial expressions.
The biggest production potential I can see right now is using it for backgrounds. It's able to match my colors and lighting pretty well with a lot more detail. And even in top tier anime like studio Ghibli, they're using painted over photographs for backgrounds a lot, so it's not that unusual.
The character renders are just an novelty. I like seeing my characters in different styles that I'd never be able to manage.
Facial expressions are a weak point of the AI models I've tried. They can do angry, happy, crying, and neutral pretty reliably but a sideways glance or looking down in embarrassment or a conflicted expression are difficult to wring out of it.
This one used the leaked NovelAI model. For the more anime looking ones, I used AnythingV3. For the more realistic looking ones, I used a 50/50 blend of AnythingV3 and F222.
If we want to get literary, it's a physical representation of her sexual desire, which in this scene has been embarrassingly exposed by her soon-to-be lover. She makes a surface level attempt to conceal it but exposes herself in the process, priming her for contact with her man. Indeed, she is distracted away from her simulacrum by the real thing moments later. He is not so easily covered up. He has seen her desire. She knows that he knows. They are inevitable.
This is based off the earliest sketch of Alison, before the full lore was established. In the inspiration page https://satinminions.com/Comic-Page-Light-Chains.html it's implied she complies immediately rather than the slow burn from the full story.
Also, AI has a lot of difficulty with her chunky collar because it's an uncommon accessory compared to a choker. In the Disney-esque one https://satinminions.com/SD-16881.html it turned into a weird ribbon thing and I just gave up.
-itten by a smart person. This painting has lots of detail? Must be a good artist. But now AI is here that can generate surface level bullshit on demand. An offhand comment can become an essay, a doodle can become a painting. For some people, their entire way of judging things is being disrupted. Now the content and specificity of a work needs to be understood, and that is much more difficult.
I am not above this frustration myself. AI is used to crap flood plausible looking content onto all platforms now. I find myself quickly checking images for the tell-tale signs of AI, listening for verbal ticks in youtube shorts, just to quickly determine if I should ignore them. But it ain't going away. In the future there is going to be more AI in the world, not less. It's better to understand it than be taken by surprise.
This turned into a minor blog post, don't mind me:
Stable diffusion works best on single character scenes in standard poses. Pinups basically. It's also pretty good at backgrounds.
The issue is that it doesn't reliably understand relative conditional statements or adjectives. So if you prompt it "a blue ball on top of a red box", it will give you that... sometimes - but you'll also get random combinations of red, blue, ball and box. It gets worse the more adjectives and conditionals you add in.
In the case of multiple characters it's almost impossible to apply the conditionals correctly. If you say "a woman with tan skin and red hair putting bunny ears on a kneeling woman with short blonde hair wearing a red leather bustier", you're gonna get red hair on both of them a because that's a common thing and you have the "red" token in there twice. It's also going to screw up normal features twice as much because there are more things in the scene. Even getting a single character that looks good often requires rolling the dice dozens of times because random bits will be screwed up.
I have done some multiple character scenes by inpainting each part individually, but the problem is you end up with a slightly different style and lighting on each part. Plus it's very time consuming. The patchwork look is a problem even on single character composite images, it's easy to accidentally stray into something that looks like a collage or that un-tooned homer simpson meme. The recent gold dress pinup kind of strays into that territory.
Dall-E is better at comprehension but it's proprietry and too large of a model to run on consumer hardware at this time. I'm generally not interested in AI models that I can't run locally and I'm especially disinterested in corporate mandated artificial brain damage. Even if you're not trying to make edgy stuff, unfiltered models are just better because they have a more complete understanding of the world. Even stable diffusion has fallen into the "trust and safety" trap and it looks like future developments will have to be underground.
The control net extension, which I used to make the latest Alison pic, is a major improvement, but it still doesn't solve multiple characters or animations. I think the possibility is there. I could see something being added like a segmentation map where each segment could be given different prompts. Temporal stability has been shown to be possible in things like nvidia's styleGAN, and some newer text-to-video models. At some point you will be able to go from a sketch animation to a perfect render. The capability in the AI model is there, it just needs to be activated appropriately. Similar to how chat-gpt is an activation layer on top of gpt-3.
I've done a lot of AI tinkering instead of drawing lately, and some people don't like it - but I hope everyone can appreciate that this an existential crisis in art. Lots of people are in "anger" and "denial" stages of grief. I've had some truly bizarre discussions on other forums where I try to demonstrate SD's ability to generate backgrounds and they will start picking apart some 3 pixel high blob of a bush in the distance because it's not an exact technical drawing. Like, have you ever seen a painting by a person? Bob Ross? The guy just smooshed his brush on the canvas and it looks great. A bunch of people were upset that netflix made an anime short using AI for the backgrounds - but tons of anime have been using crappily filtered stock photographs and 3D models for backgrounds for decades, AI could only improve this situation. Even big budget titles frequently use painted over photographs because even among artists, very few people can generate an accurate scene entirely from their mind.
A big problem is that a lot of people are walking around without any comprehension of what they're looking at or reading or listening to. They just make value judgements based on surface level traits that, in the past, have reliably served as proxies for quality. There's a bunch of big words here? Must be wr
Writing all that makes it seem like some huge process. It's way easier than drawing something. But a breakdown of the drawing process is like:
Draw some lines. Color 'em in.
The main bottleneck for AI is the time it takes to generate images. If my computer was infinitely fast, it would only take minutes. As it is now I usually queue up a batch to run when I'm afk, which isn't often.
The first step is to dial in the prompt and settings. I test different models and denoising strengths to see what looks good. Sometimes it has trouble understanding a pose until you give it the right prompt. In this case it understood it fairly well, but specifying hands_on_hips, short_sleeves, and pencil_skirt helped.
Once it starts looking good, I generate a bunch of images with the those settings. Then I piece together the good ones in photoshop and do a little retouching. It's rare for one image to have good everything so I'll take the face from one, the body from another, etc. The hands are usually messed up and have to have parts redrawn.
Then I export the composite image and upscale it 3x. Then I'll do AI inpainting on key parts to high-res-ify them. In this case, her face, her boobs, and her waist. I tried her skirt a couple times but I ended up keeping the original.
The main thing for this one is that I knew it would work since I had already done the prompt testing part at low resolution on my old video card. In general a simple pinup pose like this is easy since there is lots of training data for it. I have yet to get good results on unusual angles like this https://satinminions.com/Suction-Dildo-Shower-04.html or first person shots.
It is possible to just run an image through and get pretty good results. But if you want a high res output where you struggle to find flaws, more work is required. Here is the raw output for this image: https://satinminions.com/SD-15675-Raw.html
The premise is that she has an organic brain inside a synthetic body. The brain would need food, water, and oxygen to survive, but a greatly reduced amount compared to a full human body. These needs would also be regulated and buffered by the synthetic systems. A chemical oxygen generator and CO2 scrubber can be hundreds more efficient natural hemoglobin, allowing her to hold her breath for hours if necessary - although that would be uncomfortable as the brain is accustomed to breathing.
The brain would also still require sleep, although the requirement for the "slow wave" period where they body repairs itself may be reduced. So she may feel suspiciously un-fatigued at times.
A synthetic body would probably need to repair itself though. Think about it, how many times can you bend a usb cable back and forth before it stops working, how long do your clothes last? Anything soft is going to need to be self-healing or it'll turn into dust in a matter of months. And then there's the matter of expendable lubrication.
So I'd imagine she'd have some kind of digestive system and eat small specially prepared supplements. I had an idea for a little story page about this.
For power, she'd probably have to be nuclear. You just can't get enough power density out of chemical batteries and the charge anxiety would be too much. I had an idea where when she's first installed she's on battery power and feels cold and lethargic, then she goes down to the reactor room to have her permanent power supply installed and immediately feels much warmer and energetic.
AI Smoothing does not work well here because the original animation is kind of jank and the smoothing it just makes the jank more obvious.
For a full remix, stable diffusion does not have temporal consistency, which means you can't really get the same character out of it twice, even with the same settings. It ends up looking like every frame was drawn by a different person.
AI really struggles with hands and feet, especially in unusual orientations like this. I've run about 100 iterations and they mostly turn out as weird blobs.
I've discovered that things can be greatly improved with inpainting and a little touchup in photoshop. I was able to fix her face and hand. If I can manage to find a new graphics card I'll be able to iterate a lot faster.
