Post your AI images
Discussion
Stealthracer said:
"Malevolent, sentient cheeses in a parody of the Warhol Marilyn painting, photorealism, hyperrealism, highly detailed"
The second one looks more like Maggie Thatcher! (Well she was born the same year.)
Strangely not photo-realistic. The cheese is way off.The second one looks more like Maggie Thatcher! (Well she was born the same year.)
But then hyper real, whatever it is, may be distorting that.
There really need to start being defined terms or image based swatch toggles so you can get an idea what different tags/descriptions will get you.
In my experience 4k, hyper real, photo real, ultra detail, etc, are just as good as the original tags (done by authors of original works and ultimately meaningless), and so essentially are just random seeds, rather than strongly indicative of outcomes.
Mr Whippy said:
Which generative AI are these from?
It seems to be following prompts quite well.
Just the Bing oneIt seems to be following prompts quite well.
https://www.bing.com/images/create
I've also tried imagebot.ai which is better for some things, although like Bing it has some quirks. It's good at mixing up colours for example - in this one the lady was supposed to have a pink top and red skirt, and the Mini Cooper was supposed to be light grey with a white roof. (And I didn't ask for the dog.)
In these next two, specifying the era of the car also meant the outfit was era-correct. (To be fair, Bing does the same.)
And when I asked for Space Echo, it sort of got the idea.
(Bing just gave me a load of planets.)
In these next two, specifying the era of the car also meant the outfit was era-correct. (To be fair, Bing does the same.)
And when I asked for Space Echo, it sort of got the idea.
(Bing just gave me a load of planets.)
Just tried imagebot.ai with 'cheese eating surrender monkeys'
imagebot.ai has an 'improve' button that changes the prompt into keywords.
cartoon, monkeys, eating, cheese, Eiffel Tower, berets, sunset, humorous, exaggerated expressions, vibrant colors, caricature, playful, Paris, picnic, baguette, wine, grass, joyful, whimsical, detailed
So it knows I'm talking about the French
imagebot.ai has an 'improve' button that changes the prompt into keywords.
cartoon, monkeys, eating, cheese, Eiffel Tower, berets, sunset, humorous, exaggerated expressions, vibrant colors, caricature, playful, Paris, picnic, baguette, wine, grass, joyful, whimsical, detailed
So it knows I'm talking about the French
Details though.
Both are real looking bar the actual woman.
The first the mini looks too big. And odd with back on pavement but wheels sitting right, it’d have a lean on.
The second looks ok, but wheels have no bolt holes and the perspective looks off on the car vs the background.
Both are completely fine images, but neither might be useful without pretty much starting again/hours arsing in photoshop etc.
Both are real looking bar the actual woman.
The first the mini looks too big. And odd with back on pavement but wheels sitting right, it’d have a lean on.
The second looks ok, but wheels have no bolt holes and the perspective looks off on the car vs the background.
Both are completely fine images, but neither might be useful without pretty much starting again/hours arsing in photoshop etc.
Yes, faces in Bing tend to look like paintings - putting the image through KREA to enhance it can sometimes make them look more natural, whereas in magebot they are quite 'photographic' in quality to start with, like the lady racers in my previous post.
When Bing gets a car right, it can look pretty good.
When Bing gets a car right, it can look pretty good.
Mr Whippy said:
Details though.
Both are real looking bar the actual woman.
The first the mini looks too big. And odd with back on pavement but wheels sitting right, it’d have a lean on.
The second looks ok, but wheels have no bolt holes and the perspective looks off on the car vs the background.
Both are completely fine images, but neither might be useful without pretty much starting again/hours arsing in photoshop etc.
The 2nd lady is Eccentrica Gallumbits AICMFPBoth are real looking bar the actual woman.
The first the mini looks too big. And odd with back on pavement but wheels sitting right, it’d have a lean on.
The second looks ok, but wheels have no bolt holes and the perspective looks off on the car vs the background.
Both are completely fine images, but neither might be useful without pretty much starting again/hours arsing in photoshop etc.
Kowalski655 said:
Mr Whippy said:
Details though.
Both are real looking bar the actual woman.
The first the mini looks too big. And odd with back on pavement but wheels sitting right, it’d have a lean on.
The second looks ok, but wheels have no bolt holes and the perspective looks off on the car vs the background.
Both are completely fine images, but neither might be useful without pretty much starting again/hours arsing in photoshop etc.
The 2nd lady is Eccentrica Gallumbits AICMFPBoth are real looking bar the actual woman.
The first the mini looks too big. And odd with back on pavement but wheels sitting right, it’d have a lean on.
The second looks ok, but wheels have no bolt holes and the perspective looks off on the car vs the background.
Both are completely fine images, but neither might be useful without pretty much starting again/hours arsing in photoshop etc.
My attempt at a Lamborghini Espada gave me a Countach instead - also she was supposed to be leaning on it, not sitting in front of it.
But I did like the way the clasp on her handbag was a little Lambo badge.
Currently trying to do the Eiffel Tower n the style of M. C. Escher, but no joy so far.
But I did like the way the clasp on her handbag was a little Lambo badge.
Currently trying to do the Eiffel Tower n the style of M. C. Escher, but no joy so far.
Gassing Station | Computers, Gadgets & Stuff | Top of Page | What's New | My Stuff