Baldur's Gate 3
Vox Machina style Baldur's Gate 3

Image information

Added on

Uploaded by

AnteMaxx

About this image

What would Baldur's Gate 3 look like if it were animated in the style of The Legend of Vox Machina? Perhaps something like this, but better.

Image created with Stable Diffusion with the following models: AnyLoRAMistoon_AmethystMistoon_Emerald and Vox Machina Style LoRA.

Mods used: Improved UIInvisible Head & BodyBasket Full of Equipment and Custom Backdrops for Character Creation and Level Up.

6 comments

  1. Kkaras
    Kkaras
    • member
    • 6 kudos
    Love the work (and the detailed explanation on how you worked your magic). Honestly, I've always preferred this kind of graphical design as opposed to the modern trend toward extreme realism. Of course the downside to such types of graphics in game is visual customization becomes nigh impossible (or far more limited in scope). But, that doesn't mean I wouldn't have loved a BG3 with this kind of art style!
  2. v1ctin
    v1ctin
    • supporter
    • 0 kudos
    hi AnteMaxx, how did you make it generate a full body character like that?, i tried using AnyLoRA and Vox Machina LoRA but it still has weird eyes when it's a full body image.. also how do you use multiple stable diffusion model thingys?
    1. AnteMaxx
      AnteMaxx
      • premium
      • 384 kudos
      This isn't a typical one and done AI image. I used img2img to make the background from a screenshot of just the CC background, which I then added as a backdrop to the actual CC screen to get the options on top of the image. Then I used img2img to make the character from a screenshot of just the character with a white background. Finally Photoshop was used to combine the images.

      img2img from BG3 screenshots to Vox Machina style wasn't the easiest thing to do. So I can't really explain in detail how this was accomplished.

      As for the weird eyes. If the resolution is too low it will generate weird eyes. You can fix this by generating higher resolution images, use inpainting or img2img after creating the first image through txt2img. Other settings might also be causing this issue

      You can use multiple models in one image by merging them. You should be able to find guides online on "how to merge stable diffusion models". Though for this image different models were used during different parts of the process in addition to using merged models.
    2. v1ctin
      v1ctin
      • supporter
      • 0 kudos
      oh alright! thanks for explaining everything <3
  3. MendyCorsair
    MendyCorsair
    • premium
    • 0 kudos
    I love it
  4. lostsoulman
    lostsoulman
    • premium
    • 143 kudos
    Hi AnteMaxx

    What magic is this? It looks fantastic!

    Great work as always