D&D General DALL·E 3 does amazing D&D art

100% did not.intend that: I assume that it is an algorithmic quirk when requesting a group of people with the same description, cut and paste. It was notable that I got diverse clothing when I stipulated "rainbow."

It's sometimes called the "twin effect" where all your characters end up being clone. The more there are, the more often it occurs. I think it's pretty strong with Dall-E because they obviously tried hard (and succeeded...) to avoid concept bleeding (that's when you ask for a fighter with a green sword and black armor and you get random green armor pieces). If the prompt we type for a generic member of a group is reapplied for each, similarities are bound to happen more often.
 

log in or register to remove this ad



Some experiments with color and action

_f619adeb-0ba0-472a-9ad6-a2e431cf00fe.jpeg
_0acab781-c861-41fb-a4bc-67b3d68f6e7d.jpeg
_67e4bf0a-812e-4fea-ac6a-1a26704b3fb5.jpeg
_298b80d2-72d6-4ed3-94f4-696eea7ef88b.jpeg
_9962d034-631e-43cd-976d-309b577439b0.jpeg
_76aee0fe-8889-4c8f-b2f9-190011cf285b.jpeg
_6f75d68b-890e-4df5-b6b4-e53c274cdd3e.jpeg
 



It's fascinating the things the AI "gets" and the things it doesn't. Obviously it "understands" elf babes (or more generally, plenty of nouns and adjectives). But getting it to draw a specific number of things, or get things positioned (ie, prepositions) is a mess. It needs to spend some serious time watching Sesame Street, imo.

Anyway, here's few experiments of D&D themes in different styles. The AI gen has been painfully slow to finish my requests, so I haven't been able to learn how to fine tune any better, but they're oddball fun I think.


Security camera footage in a certain temple:
_4ea230f3-5faa-4e97-9a6f-9801534cb2a4.jpg


After some redecorating of his demiplane, here is a portrait of Conde Estrahda at his hacienda, Desván de los Cuervos
_7623fd15-23ee-4aef-8f4a-c8eceb3efd40.jpg


A polaroid of the real-life event that inspired the cover of the Red Box Basic set.
_fbee036c-c898-487c-a4d5-93a83a829832.jpg


Proud Dragon hunters in the old west:
_7089413e-c7ab-48ca-911d-57eb17f2a1f1.jpg
 


It's fascinating the things the AI "gets" and the things it doesn't. Obviously it "understands" elf babes (or more generally, plenty of nouns and adjectives). But getting it to draw a specific number of things, or get things positioned (ie, prepositions) is a mess. It needs to spend some serious time watching Sesame Street, imo.

Anyway, here's few experiments of D&D themes in different styles. The AI gen has been painfully slow to finish my requests, so I haven't been able to learn how to fine tune any better, but they're oddball fun I think.


Security camera footage in a certain temple:
View attachment 298922

After some redecorating of his demiplane, here is a portrait of Conde Estrahda at his hacienda, Desván de los Cuervos
View attachment 298923

A polaroid of the real-life event that inspired the cover of the Red Box Basic set.
View attachment 298925

Proud Dragon hunters in the old west:
View attachment 298924
It doesn't "understand" anything, really.
 

Oh, !I see you were talking about thw green leather jacket and floral dress. They are not the same, each is diatinctly different in the details.
Yeah, they're not exactly the same style - but it's still a green leather jacket with a white dress with pink flowers. Did you specify those in your prompt? They're uncannily similar, that's all.
 

Remove ads

Top