AI Image requests - Now generate her giving birth.

I probably should've been clearer that I'm not developing anything right now, I was just curious to see how well something like this would output images for seamless textures in games.
I also should've been more clear on what I had in mind since if you've ever been to a site like textures.com then you'd probably know what textures look like from various materials from there and how they try to make stuff like grass, bricks, floors, marble, etc. I will say the first blue bricks it generated looks very nice and what I thought it'd look like. As for the meat, I had something more like this in mind, just meaty in texture with some white lines to give it that marble look. I'm sure if you'd use the images from there you could make some funky and unique textures and not just based it all solely on text alone, but again I was curious to see how well it could handle generating images if for texturing and modeling purposes for games or what have you. (Thanks for generating it for me btw)
 
  • Like
Reactions: Suspicious Activity
I probably should've been clearer that I'm not developing anything right now, I was just curious to see how well something like this would output images for seamless textures in games.
I also should've been more clear on what I had in mind since if you've ever been to a site like textures.com then you'd probably know what textures look like from various materials from there and how they try to make stuff like grass, bricks, floors, marble, etc. I will say the first blue bricks it generated looks very nice and what I thought it'd look like. As for the meat, I had something more like this in mind, just meaty in texture with some white lines to give it that marble look. I'm sure if you'd use the images from there you could make some funky and unique textures and not just based it all solely on text alone, but again I was curious to see how well it could handle generating images if for texturing and modeling purposes for games or what have you. (Thanks for generating it for me btw)
Ah, yeah. I did kind of understand what you were getting at. Yes. The art generator is designed more for creating entire one off pieces. But could be adapted to making seamless, unique textures for game design. I think the potential is there. I had a quick look at the site and found their ‘blue brick’ texture.

They seem quite protective of their content, not wanting you to download too much etc, etc. I just took a shot of this as an example…
54BA843E-9718-4700-BA25-1B0A93002831.jpeg
I used this as a photo prompt in the generator, influence on high level.
A photo prompt always requires some verbal input too, so you can write in some descriptors and sentences to alter the original, whilst still having the picture be of high influence.
Like this
Verbal prompt: dirty, dilapidated, old blue brick wall



128C05AE-2F19-4746-966D-5B065CE729A7.jpeg
1DC46288-05CF-4FC3-A9FE-C4E96E60C019.jpeg4EF498BA-8175-4473-91ED-AF407C371390.jpeg

Verbal prompt: sparkling gemstone blue brick wall
F53A06A0-DAB0-448B-9EA3-32691530DC96.jpeg0EBDFD9B-9E97-4969-9AF0-544D6DDF5855.jpeg

So you can take someone else’s textures and manipulate them quite nicely.
 
  • Like
Reactions: Colon capital V
Alex Jones dressed as Jc Denton carrying an M249 Machine Gun
I’m not sure what either of these guys look like, but neither did the AI. To help it along I used the image prompt as well as the verbal prompt. It’s a bit fiddly but I think I’ve pulled it all together. As I said, I’m not sure what they’re really like, so I don’t don’t know if I struck the right energy. Here they are, anyway.

A880487E-5D48-41AA-8E14-66C8E5AA4D91.jpeg3BBD5A46-2927-49D1-BAD3-F87C109862D2.jpeg26C15D50-2E00-4D98-87DB-0791BD6C2F24.jpeg 6A39DDCA-B30A-4D77-8741-1890B57F422A.jpeg
 
Verbal prompt: dirty, dilapidated, old blue brick wall
I went ahead and nabbed that first output of the blue bricks and tried my hand at making it as "seamless" as I could for fun.
bluebricks.png
HUYjbTrWvO.png
Not sure if I like how those big carves on the left and middle of the bricks kinda pop out like that, but I could always mask that out if I so wished.
Overall, while I'm not too experienced with fixing or making images "tileable", this could definitely prove to be very useful in making unique textures out of pre-existing ones from places like textures.com if I decided to put in the effort on fixing it post generation.
 
  • Agree
Reactions: Suspicious Activity
I’m not sure what either of these guys look like, but neither did the AI. To help it along I used the image prompt as well as the verbal prompt. It’s a bit fiddly but I think I’ve pulled it all together. As I said, I’m not sure what they’re really like, so I don’t don’t know if I struck the right energy. Here they are, anyway.

View attachment 5114416View attachment 5114417View attachment 5114418 View attachment 5114421
That’s exactly what I wanted. Perfect for a portrait.
 
  • Like
Reactions: Suspicious Activity
I'm actually writing a story about a Buick with Ingrid. Did you use an actual image to help? Gorgeous.

Try her with a gun.
I did a Google image search for myself, just so I knew if the AI could be accurate and knew her face, or if I would need to import a photo, but no, it did and the Bergman generations are all verbal prompt. I asked Chat GPT what sort of gun would be accurate for her to have owned. He got kinda cagey and said it wasn’t his place to recommend firearms to people, but I continued in a ‘hypotherical’ manner, what she may have liked and we decided together that she might have had a nice shiny pistol, or something. Image generator was clearly confused about guns and portrayed them as light sabers in one image. And it really doesn’t know where the handle is or where the business end is supposed to point. I’ll try and teach it through verbal edit. It’s a fun case study.

So are these. It’s so interesting trying out all these briefs, I’m learning what this thing can and can’t do. It’s hilarious to me that the AI is not entirely sure how a human holds a gun. I got many generations of her holding the barrel or the muzzle and waving the grip around.
Here are a few, in some different styles. Are you going for a certain aesthetic. I can get some very dark moody stuff right up to bright paint splashes and little muppet models. I figured you’d like the more realistic settings. I’ll throw in a few for your viewing pleasure.

I’ll start to spoiler long posts of images. I don’t want Null coming in here telling me I’m causing the site to crash when the page opens. It loads quicker for everyone this way.
17268078-8EC7-48AB-B0FC-AE4DD682DE72.jpegB837C411-3EFD-471B-AF23-44A903A69789.jpegDD3BF3FA-A338-4773-95EC-3DACE34CF0C1.jpegB837C411-3EFD-471B-AF23-44A903A69789.jpeg88A75D54-91DB-4F8E-AF37-819A8B2E33F6.jpeg3B8805FF-6C90-4398-9C61-DFB0D479D88B.jpeg62E188F7-1845-4252-90E4-BDD06B0C8CF9.jpeg544F87D7-B4F4-4034-AB44-AD87CF9905C0.jpeg649767AA-9217-4969-9CFD-DEDA6FF82289.jpegF37F34FA-31EC-4C40-9D8D-A5AC7FFB4292.jpeg395BBC7F-3832-40B8-A0D6-FDFD9668DB8C.jpegD0470A12-6905-4BB7-8226-B9F566D6894C.jpeg
 
It’s hilarious to me that the AI is not entirely sure how a human holds a gun. I got many generations of her holding the barrel or the muzzle and waving the grip around.
I'm surprised you didn't get any of her fellating the gun barrel which was a not unusual result when I attempted renders with firearms.
 
  • Thunk-Provoking
Reactions: Suspicious Activity
I did a Google image search for myself, just so I knew if the AI could be accurate and knew her face, or if I would need to import a photo, but no, it did and the Bergman generations are all verbal prompt. I asked Chat GPT what sort of gun would be accurate for her to have owned. He got kinda cagey and said it wasn’t his place to recommend firearms to people, but I continued in a ‘hypotherical’ manner, what she may have liked and we decided together that she might have had a nice shiny pistol, or something.
I could've answered that myself, lol. Given the time period with movies of her era, a revolver would be best. Stylish, simple, elegant, and lethal.

I cannot unsee Ingrid with Star Wars now. So beautiful. AI needs to chill with gun possession. It's not the end of the world. I noticed AI has an issue with depicting hands and some faces.

You'll have to show me how to use your AI tool.

I need more originality with my prompts. Maybe, an alien invasion at the White House during a parade.
 
I could've answered that myself, lol. Given the time period with movies of her era, a revolver would be best. Stylish, simple, elegant, and lethal.

I cannot unsee Ingrid with Star Wars now. So beautiful. AI needs to chill with gun possession. It's not the end of the world. I noticed AI has an issue with depicting hands and some faces.

You'll have to show me how to use your AI tool.

I need more originality with my prompts. Maybe, an alien invasion at the White House during a parade.
Sure! I’m only learning myself, but the learning curve is steep. The tools still have major flaws. But flaws are necessary and give invaluable insight into how these machines think. I have found that to fulfil more complicated brief, the way to go is to use my freebie knock off photoshop to block in the basic premise of the final desired image.

The example of the latest brief I tried. “Donald Trump farting a swarm of bees, holding an AR. 15 and waving an Israeli flag.”
I tried this straight as a verbal prompt. No. It was so confused and couldn’t answer the brief well at all.
My methodology here is to use a different strategy.
  • Get a transparent background (tb) image full body shop of trump.
  • Get a good tb image of the gun
  • Fiddle around in my basic photoshop to make a personalised ‘sticker’ of a swarm of bees.
  • Get a good tb Israeli flag.
These can be roughly and easily combined in my photo edit suite, and it doesn’t matter if it’s really rough, the stickers of the bees and the flag are really rough,…..they’re IN THE RIGHT PLACE.

THEN I take these to Dream, the AI I use….use my rough photo as a reference, saying use this as a source to rely on heavily…..then put in your verbal prompt ( the wording, phrasing and order of words is important) and the AI can generate some great interpretations of your rough photoshop….ending up with some cool, polished images. I’m. Learning how to use a cheap photoshop to enhance the results from the AI art generator. I find it really interesting. The way it get things right and the way it gets things wrong….and finding ways to get round the shortcomings.

Feel free to take it to my DMs/ conversations. Any questions you have, I’m always around. It might be a case of the blind leading the blind, but two heads are better than one. Aliens at the White House during a parade sounds fun. Will give it a try.

As I said. My DMs are always open.
 
  • Semper Fidelis
Reactions: The Last Stand
The example of the latest brief I tried. “Donald Trump farting a swarm of bees, holding an AR. 15 and waving an Israeli flag.”
Who on Earth... I didn't think that's possible.

Great, now I have to top that.

Audrey Hepburn controlling fire with her hands ... in the middle of the beach.
 
Donald Trump farting out bees while firing an AR-15 and yielding an Israel Flag
Is this what you had in mind? AI was far too prudish to do this. I ended up using a photo manipulator to put the basic elements in place, all rough and ugly. Feeding this final photo into the AI reference can make your photo look much better as long as you get the verbal prompts correct.
0634E072-D742-4EED-86CE-EC2834E223DE.jpeg 88D3B289-C2C8-4409-89AC-95E7C921EDFC.jpeg845DA783-369E-4B4E-B0A3-8C79BBB4E051.jpeg18896F96-6B77-4DF2-B15C-008C4CB24541.jpeg
 
I could've answered that myself, lol. Given the time period with movies of her era, a revolver would be best. Stylish, simple, elegant, and lethal.

I cannot unsee Ingrid with Star Wars now. So beautiful. AI needs to chill with gun possession. It's not the end of the world. I noticed AI has an issue with depicting hands and some faces.

You'll have to show me how to use your AI tool.

I need more originality with my prompts. Maybe, an alien invasion at the White House during a parade.
6EFB15F8-4176-4EF6-B11B-22AA97D2878C.jpegD3121903-9BE1-4367-AB94-AB2A68F36A60.jpeg36C6632A-0BD2-4F2C-8AD8-A8F9514BAA9E.jpeg23779C51-7B39-47A0-80DC-55EDCA5789A5.jpeg7D09A885-3C38-463B-BE8B-21EE35421C53.jpeg1DE15C6C-941E-4B7A-9358-4C7B7A76F891.jpegC0B79C62-BB90-4F2C-AA64-6FDDF325B715.jpegA208B13C-5BEA-4636-9132-0F15F11260BC.jpeg
This one wasn’t as easy as I thought it would be. I definitely don’t know how to animate stuff yet. I’m happy to try, if anyone has any ideas but I need to add I’m 100% IOS. I murdered my last laptop out of spite and revenge for Windows 10 just existing…and I never replaced it.
ECE86226-D456-429A-8242-2812B1D03543.jpeg
 
We need an alien invasion with all this nonsense going on now.

Let's do... Norma Shearer enjoying coffee and hot breakfast while wearing a nightgown.
 
  • Thunk-Provoking
Reactions: Suspicious Activity
We need an alien invasion with all this nonsense going on now.

Let's do... Norma Shearer enjoying coffee and hot breakfast while wearing a nightgown.
AI alone on this one was tough. I’m learning that adding verbs and especially multiple verbs, really confuses it. Inputting a location, a number of nouns and a couple of verbs….you get some craziness. Interesting, but not what you want.
However, it’s quite easy to ‘build’ a very rough, sloppy version of your general desired outcome by piecing together a background, a person shape and a vague table/food shape in a photoshop program can be used as a guideline image for the AI to extrapolate a nice detailed image. To be honest, I think you could do a rough jot in MSpaint and import it…would do just as well.
anyway…here are the better results I put together.
C730CBBF-1BA6-4317-88C3-DFC0E24DB2CC.jpeg131F26AD-15F1-4901-9759-A5711640D5D8.jpeg2BE4F316-FA0A-429B-886D-53AF53087CA7.jpegBEF2CEEB-CF17-4FF9-B746-2FF5E65C5479.jpegBB956F21-9DA6-4E6E-8E84-2CD23ED819E7.jpegD9779D33-FBA1-4C02-8704-FEE1B9B53638.jpeg
 
  • Feels
Reactions: The Last Stand
AI alone on this one was tough. I’m learning that adding verbs and especially multiple verbs, really confuses it. Inputting a location, a number of nouns and a couple of verbs….you get some craziness. Interesting, but not what you want.
However, it’s quite easy to ‘build’ a very rough, sloppy version of your general desired outcome by piecing together a background, a person shape and a vague table/food shape in a photoshop program can be used as a guideline image for the AI to extrapolate a nice detailed image. To be honest, I think you could do a rough jot in MSpaint and import it…would do just as well.
anyway…here are the better results I put together.
If you're back. Jean Harlow laying down with a Thompson gun.
 
Back