AI images

Runnen

Vyemm Raider
1,747
4,027
Avatar of fear / room in temple of Cazic Thule

There's been several revamps of the Temple of Cazic-Thule with different rooms for the Avatar of Fear, let me know if you wanted a different one.

_5673e25f-0385-41f6-9b31-7a2261a8d5e7.jpg
 
  • 6Like
Reactions: 5 users

Siliconemelons

Avatar of War Slayer
10,917
15,392
The OG I spent soooooooo much time there it was my jam, I soloed a ton, takes forever as a SK but I trained the nubs groups they cycled in on camps and how to pull etc was a big part of my EQ life

there is a real revamp? Not just when they messed it up with new mobs?
 

elidib

Vyemm Raider
2,063
3,495
Is there some super-secret phrase you can add to your prompt to get it to confine results to actual high-quality photorealistic images? Bing loves to go on a tangent with cartoony bullshit like this:

fail 2.JPG


fail.JPG
 
  • 2Like
Reactions: 1 users

Runnen

Vyemm Raider
1,747
4,027
Is there some super-secret phrase you can add to your prompt to get it to confine results to actual high-quality photorealistic images? Bing loves to go on a tangent with cartoony bullshit like this:

Some keywords tend to get your more cartoony / animated results. Going with "small" or "short" people tends to make it more cutesy. If you do need a small person who looks realistic, you can try to add some other descriptions such as "threatening" or "serious-looking" to steer it back out of cartoony stuff, but sometimes the AI just tries different things as you submit the prompt again and again because it's trying to explore different avenues. Usually after enough prompts it goes back out of Toon Town and you get some realistic results. I'd try and modify my prompts to remove any cute-adjacent elements if I were you.
 

Runnen

Vyemm Raider
1,747
4,027
The best part of the Deep though:

Your own fault for not using the totally-safe-super-not-fake bridge thoughtfully provided by the local.. *checks notes* ..thought horrors.

_9f4e12c7-3c54-479d-a6a2-33c121465641.jpg
 
Last edited:
  • 6Like
Reactions: 5 users

Chris

Potato del Grande
18,339
-262
I am totally blown away by these images, what the fuck.

Are you guys doing anything special or just using base dall-e or midjourney?

My mind keeps going to how to make money off this and I need a new career lol.
 
  • 1Like
Reactions: 1 user

elidib

Vyemm Raider
2,063
3,495
I'm just using Bing's chat, which uses Dall-E 3. the key is to make your prompts really descriptive. the better you can describe what you want, the better results you get. so it becomes more about being literary than being artistic. After doing some research, you can kinda mold your results by telling it what style you want the image in. so for example, i made this Akanon picture using the below prompt:

OIG.s8MnsrxZAQm5eCTb47ZH


OIG.ZW3wtDznrn.I7XpI6u2a




A slightly lighter wooded area with mountains in the background. Roaming around are some clockwork spiders and rats. There are a couple of clockwork people with headlights for eyes. In the background is a big gear and there are mountains in the distance. A clockwork building is added in the far distance. The atmosphere is dark and eerie. The scene is captured from a wide angle with a 20MM lens.


Here are some tips for which angles and lenses create which type of exposures
and here's a really good reference document that helped me pick out some style choices: https://dallery.gallery/wp-content/uploads/2022/07/The-DALL·E-2-prompt-book-v1.02.pdf

another good keyword i started adding to prompts is something like "make it in the style of a screenshot from X game" where X is something very current and next-gen. Like Elden Ring or something.
 
  • 11Like
Reactions: 10 users

Pasteton

Blackwing Lair Raider
2,609
1,724
There’s definitely a talent /finesse to prompting for images. I think there’s enough of a skill curve that it could become a marketable skill, but one that will subsequently vanish once an LLM becomes a far better prompter than any human.
 

Siliconemelons

Avatar of War Slayer
10,917
15,392
I love Castlevania and that's my Halloween I suppose so I figured I would let Midjourney play with a lady castlevania thing

1698783496023.png
 
  • 5Like
Reactions: 4 users

Runnen

Vyemm Raider
1,747
4,027
Runnen Runnen - time to start on world of Warcraft?

There's a literal gazillion WoW-like AI artworks out there, and while I do still "play" WoW for some reason, I prefer to work on EQ stuff because it scratches the nostalgia itch and I like trying to make the most out of outdated graphics, I like to imagine things from EQ as the developpers might have imagined them, but then the technical limitations of the era and budget came in and they had to do what they could.

WoW has improved graphically over the years but its cartoony style tends to age better (like to this day Zelda A Link to the Past still looks good to me because it's cartoony pixel art as opposed to N64 games with polygons). EQ looked impressive when it came out but today it's comically awful-looking for the most part but, again, it's so nostalgic (the newer models they used for zone revamps / new content in EQ look for the most part horribly out of place to me even though ironically they have a bunch more polygons and higher resolution but it's like seeing an Unreal engine Goomba in my NES Mario Bros).

I've had less spare time to do AI stuff last few days and the AI keeps crashing a lot, so it's a bit slower, I'm trying to see if I can make it do stuff in the style of Keith Parkinson, or maybe try some fun crossovers, or just striking EQ places, designs or memories I have.. I dunno, sometimes I just get an idea and I write it down to try some AI later.

Requests are fun too, don't hesitate to suggest stuff you'd like to see, or if you want a particular character / screenshot you cherish redone, I can try that. It can be from other games/media too, I just have a soft spot for EQ but the sky's the limit.
 
  • 3Solidarity
Reactions: 2 users