chambersu
patreon
Character design tools/workflows/models update
🕑 Added 2025-01-15 10:12:50 +0000 UTCBatch models creation to test/improve my workflow,
Including 1. anime fanart model Misaka Mikoto, (from 'とある魔術の禁書目録')
2. medieval female knight Revna, (for a medieval game)
3. anime female advanturer Sophie, (for a fantasy game)
4&5&6. female ranger/artificer Elara, both virtual model and anime version and her mech. (based on my old models 'Ray', which I want to post and update on my Sketchfab page)
Download:
RevnaSophieElaraBilibili_v02_t06_Packed.blend
I model characters in 3 steps now,
1. design/sketches to preview first, just like the project files above, (ComfyUI/Blender/PhotoShop/Trellis/model library)
2. then high-poly detailed polished version later, (Blender/MavelousDesigner)
Kimiko_v02_t06_Packed.blend
Fanart Kimiko Ross from Dresden Codak is closer to step 2 (a non-exclusive model I am working on~), but her hair is still rough, the hair mesh/cards seems the most difficult part for me now...~ have to deal urregular shapes/textures by hand~
3. finally retopo and baked as low-poly version for taeget platform,(Blender/game or render engines)
like the humanoid creature, and the alchemist suits below, (non-exclusive models for a dark virtorian style game)
creature
https://1drv.ms/u/s!Am0lZ75OjQ1mrPMv8yiNX6itLdT6AQ?e=7v33yi
male suit
https://1drv.ms/u/s!Am0lZ75OjQ1mrNYqegHuoop6k3Cjcw?e=lHsqWD
female suit
https://1drv.ms/u/s!Am0lZ75OjQ1mrLJLbAKmxNJVewIfsw?e=DW43ey
Update LoRAs, (~4GB)
20250115LoRAUpdate.zip
deliciousInDungeonChar_flux-lora-000004.safetensors
matayosiStyle_flux-lora-000006.safetensors
I uploaded my trainning preset/data sets above too, feel free to test and will be happy that let me know how to improve, I am a little busy to train one by one carefully,
(I am not sure if I upload them all, let me know if any missing~
including all design tools' Pony and IL versions,
switch to dim 4 for less detail but foucus on concept,
new character style Matayosi(Not good) and Delicious In Dungeon(work fine~)
Update workflow,
workflow 1.zip
Drag the images to ComfyUI to load the workflows
- I updated the workflow to merge IL and XL now, (F1 is too slow, IL is better at prompts understanding/controllers than XL, while XL looks better~)
- 1024^2 Unlimited Tiles 3x2 Sample, which should be able to extend unlimitedly in a VRAM-friendly way by inpainting 2048^2 edge
Turn Design and Texture Projection
Also here is the screen record of the Character Turn Design Generation and Camera-Based Texture Projection,
I took a quick try and posted on Discord before, https://1drv.ms/u/s!Am0lZ75OjQ1mrPN6einG7VeCPdTUvg?e=slTqtX
also the samples' files, It's a simple version but easier to understand maybe...I hope~
or ask me here directly if any feedback, I will reply asap~
Part A
00:00-01:30 Open Blender and ComfyUI to Load the workflow
01:30-05:20 can be skipped(image generation)
05:20-06:00 Copy the camera for projection and the base mesh
06:00-06:30 Blender setting(camera size/GPU rendering/standard color space)
06:30-09:10 Projection UVs rough preparation
09:10-17:36 can be skipped(Tweak/Align Projection UVs by hand)
Part B shader, bake, and polish as you want~
I will slowly update my LoRAs on CivitAI too,
Chamber Creator Profile | Civitai
just like I used Sketchfab as my 3d models library,
ChamberSu1996 - Sketchfab
(you can use their pages to ask me for something additional like the data sets or blender projects file too~ the visual showcases are great~)
but Patreon will have the latest assets and support
(seems hard to organize assets here for me, sorry for that,
but trust me and ask me directly if needed~),
I am willing to help or chat as usual~
Feel free to contact me by leaving a comment or mailing to chambersu1996@gmail.com, if any questions or anything here interests you~
Also Discord Channel Here:
https://discord.gg/XAfS57Rrj6
will reply ASAP ~
