ZEITGEIST Using Midjourney image AI to imagine Zeitgeist

Rukaiko

Explorer
In the end I found a style that I really like for Zeitgeist, so one by one I decided to convert all the important npcs in this style, trying to keep them as close as possible to the original idea.
 

Attachments

  • 00234.jpg
    00234.jpg
    361.1 KB · Views: 107
  • 00207.jpg
    00207.jpg
    451.3 KB · Views: 109
  • 00187.jpg
    00187.jpg
    461.8 KB · Views: 108
  • 00186.jpg
    00186.jpg
    379.8 KB · Views: 104
  • 00185.jpg
    00185.jpg
    418.2 KB · Views: 105
  • 00184.jpg
    00184.jpg
    551.3 KB · Views: 108
  • 00181.jpg
    00181.jpg
    499.3 KB · Views: 103
  • 00178.jpg
    00178.jpg
    797.4 KB · Views: 98
  • 00301-4185520716-(illustration___.png
    00301-4185520716-(illustration___.png
    444.6 KB · Views: 107
  • 00227.jpg
    00227.jpg
    408.9 KB · Views: 104
  • 00302-1310925048-(illustration___.png
    00302-1310925048-(illustration___.png
    383.3 KB · Views: 103
  • 00311-3419970453-oil painting____.png
    00311-3419970453-oil painting____.png
    547.2 KB · Views: 105
  • Dale.png
    Dale.png
    221.5 KB · Views: 96

log in or register to remove this ad


Rukaiko

Explorer
Sure, but since it was with imgtoimg at low denoising and controlnet with depth/canny the prompt and model alone won't give the exact same picture.
"portrait of [description here, for Dale I used handsome man, serious face, wearing leather armor, pale skin], (illustration:1), (best quality), (masterpiece:1), <lora:Colorwater_v4:0.8>, detailed eyes, (strong shading), hard shadows, style of Lois van Baarle and charlie bowater and Sui Ishida, thick brushwork, detailed face and eyes"
I used Moistmix_v2 as model and colorwater_v4 as lora.
I think I used this guy in the zeitgeist npc portraits as reference:
1680193785762.png


In the last ones I even added another lora with a low weight that adds even more "watercolor" effect: <lora:freaksOfTheHeartland_v01:0.3>
1680192680338.png
Yesterday I made some NPCs, Tinker, nathan jierre and the eladrin slave on the train with and without veil.
 

Attachments

  • schiava2.jpg
    schiava2.jpg
    816.7 KB · Views: 83
  • schiava1.jpg
    schiava1.jpg
    1 MB · Views: 85
  • nathanluc.png
    nathanluc.png
    3.9 MB · Views: 87
  • 00409-3325984080.png
    00409-3325984080.png
    2 MB · Views: 85
  • Tinkerr.png
    Tinkerr.png
    2.1 MB · Views: 85
  • Hyacint.jpg
    Hyacint.jpg
    920.5 KB · Views: 79
  • Valerie.jpg
    Valerie.jpg
    659.5 KB · Views: 77
  • Sera.png
    Sera.png
    1.6 MB · Views: 79
  • Jarres D'lane.png
    Jarres D'lane.png
    2.9 MB · Views: 82
  • 00248.jpg
    00248.jpg
    465.5 KB · Views: 80
  • 00239.jpg
    00239.jpg
    546 KB · Views: 85
  • 00236.jpg
    00236.jpg
    626.1 KB · Views: 83
  • 00182.jpg
    00182.jpg
    865.3 KB · Views: 80
  • randal pell.jpg
    randal pell.jpg
    673.9 KB · Views: 82
  • old stag.png
    old stag.png
    2.3 MB · Views: 73
Last edited:

Echolocation

Explorer
Sure, but since it was with imgtoimg at low denoising and controlnet with depth/canny the prompt and model alone won't give the exact same picture.
"portrait of [description here, for Dale I used handsome man, serious face, wearing leather armor, pale skin], (illustration:1), (best quality), (masterpiece:1), <lora:Colorwater_v4:0.8>, detailed eyes, (strong shading), hard shadows, style of Lois van Baarle and charlie bowater and Sui Ishida, thick brushwork, detailed face and eyes"
I used Moistmix_v2 as model and colorwater_v4 as lora.
I think I used this guy in the zeitgeist npc portraits as reference: View attachment 280359

In the last ones I even added another lora with a low weight that adds even more "watercolor" effect: <lora:freaksOfTheHeartland_v01:0.3>
View attachment 280315Yesterday I made some NPCs, Tinker, nathan jierre and the eladrin slave on the train with and without veil.

Thanks so much for this. Using your advice, I made some tokens for my player's adventures in Alais Primos (Book 11). Unfortunately, the art was a bit rushed as I made them right before session.

Nevertheless, here is:
savina.png

Legate Savina Tullius

Arch Secula Natalia Degaspare.png

Arch Secula Natalia Degaspare

advorel.png

Advorel, Heavenly Emissary
(getting an accurate 'whole body' art for this angel was tough, so I settled for a helmet)

And some NPCs specific to my campaign:
1681261752831.png

Oscan Ligurio the Reformed. My players handed Oscan to the family following Adventure Seven. Subsequently, Oscan was tortured and his mind crushed, unable to truly repent for his sins. Now insane, he serves as a hivemind scout in Plaza Hyperion during the God Trials.

antonio.png

Antonio Aneenya, Father of the Volcano.
Father Tony once headed the cathedral atop Enzyo Mons. Now, he reluctantly serves the godhand Sigismund to kill gods. Somehow, Tony keeps finding himself with blood on his hands, despite leaving the Family long ago.
 
Last edited:

Rukaiko

Explorer
Wow, those are really good! May I ask what model and lora did you end up using? Expecially for Oscan Ligurio and Antonio, they look a lot less anime syle.
 

Echolocation

Explorer
I'll show my ignorance here, and I've just been typing things to Discord's MidJourney Bot 🙃 I doubt the code is working as written, especially the colourwater lora.... I should tidy up the prompts.

E.g., here is a prompt used to generate Savina Tulluius.
Code:
portrait of dark skin wood elf tall woman with pointed ears, glinting, glowing, with long, tied-up, nicely styled red hair, wearing red and white robes, singing an opera (illustration:1), (best quality), (masterpiece:1), <lora:Colorwater_v4:0.8>, detailed eyes, (strong shading), hard shadows, style of Lois van Baarle and charlie bowater and Sui Ishida, thick brushwork, detailed face and eyes
 

Rukaiko

Explorer
Aaaah, ok, sorry I forgot to mention that I'm using Stable Diffusion, not Midjourney, the procedure and prompts are totally different!
 

Echolocation

Explorer
Ha, I realised after checking what the prompts meant and figured it was Stable Diffusion. Strangely though, I'm happy with the MidJourney results, so I'll likely clean up the prompt and continue to use it.
 

Crispy120286

Explorer
I have an image request! On Thursday my players are about to go through the Near Future and Far Future visions at Ingatans Refuge. I only found one single image of the Gyre constellation and that's the one with Rock Rackus. I would like to know if someone can use the AI to make an image of just the sky with the Gyre as it would appear in the sky. I want to traumatize them more than I already have >:)

Thank you in advance!
 


Remove ads

Top