chambersu

2025 Aug Asset Update, Snails, Ju-Eun, Nexaes, Mia, Room...~

Added 2025-08-23 19:02:25 +0000 UTC

If you have any questions or would like to discuss anything further, please don't hesitate to:
Leave a comment
Email me at chambersu1996@gmail.com
Join my Discord channel: https://discord.gg/XAfS57Rrj6

I'll respond as quickly as possible.

Thanks a lot to all my supporters on Patreon. I am sure I can't insist without your help.~
If you read from somewhere else, consider supporting me on Patreon if possible.~
patreon.com/chambersu
If still suffering from the bill like me or other plights, please relax and feel free to enjoy the contents, which will encourage me to post more and better too~

SnailGroup_v02_t04_Packed.blend ~214MB

Modeled the snail-like monsters for a dark Victorian theme RPG, the baked low-poly version, and posted here with their permission.
DarkVLib_v01_t02_Packed_2K.blend ~318MB

Also packed a library for the project to preview more easily (less editable, lower texture res, and fewer progress details than the full projects posted before).
Also tested the stylish/style shifting/aligning workflows to try to make the series for the same theme more harmonious in the future.
Preview And Output Images Including Workflows.zip
(Drag the raw output image from the zip file to ComfyUI to load and read the whole workflow)

L33Ju3un_v03_t04_WithoutPhoto.blend WIP ~277MB

Used an old hairstyle asset to preview here, will replace her real hairstyles later.
The posted version here has replaced the basecolor map/photo material with some common textures, which are still able to read/learn the necessary info to build a similar thing.

This is a fan art for Lee Ju-Eun. I should not post the whole project directly to avoid abuse.
It is for education and research only.
And may show the rendered images/video to support her in the future when finishing the whole character(head/body/suit...)/setting.
The method to build such a character I have mentioned before, all about the head turn design workflow(also face LoRAs training) and camera space texture projection.

Also tested the stylish/style shifting/aligning workflows to try to create an anime/painting face from a real face.
(Drag the raw output image from the file “Preview And Output Images Including Workflows.zip” above to ComfyUI to load and read the whole workflow)
Head_F1_LoRAs.zip ~43.9MB
including Delicious in Dungeon, Final Fantasy Tactics, and Tactics Ogre Wheel of Fortune style head Flux LoRAs.

Nexæs_v06_t02_Post.blend WIP ~64.5MB
Model this sci-fi male character design for one of my bosses~ Post here for education and research only with their permission.

Mia_v05_t03_RigDEFWeight_Post2.blend WIP ~62.5MB

retopology, deformed bones position, and weight transfer for this girl character, for a anime game project,
Also, it is a sample for me to rebuild and test a new advanced rig system/workflow mentioned before.
Her final face mesh and textures (which are more anime/Genshi/ZZZ-like, cleaner, and flatter)are created by another talented artist.
I used the version modeled by me here.

Room_v01_t12_Retopo_Post.blend WIP ~1.6GB

Modeled a bedroom, based on where I am typing now(80% similar to it, lol~),
to test the real scene/level models by AI 3D, which is related to the real face test above.
The raw 3D models(meshes/topology/UVs/textures) by AI did harm to real-time performance/editing/storage, but were really cool for preview as placeholders.
The smart retopology functions I have tested are exaggerated (at least 10-20k triangles for a simple asset to keep the details I want with bad topology),
So I started retopo and baked them by hand(to use 10% triangles of the "smart" ways with simpler and clearer data, but as we know, the work hours are expensive...but necessary sometimes),
about 30 low-poly baked assets from 100 messy high-poly detailed assets(most by AI 3D, and some industry product models like laptop, monitor, and keyboard are from other libraries).

As I mentioned earlier, I still believe the 3D scan/measurement method for rebuilding photoreal assets has the best quality, but it also incurs significant costs.
So, I should try one by AI 3D to determine the actual budget/work hours/quality in this way.
Maybe it is useful for some requirements.
I felt some potential during modeling, Cool~

For working with AI texturing, I guess anime/painting styles will be better, so I tested style shifting too, and will try further with them.
(Drag the raw output image from the file “Preview And Output Images Including Workflows.zip” above to ComfyUI to load and read the whole workflow)

Seiji23.safetensors ~18.4MB, trained from scanned paintings from the artbook Houses with a Story by Seiji Yoshida (the Japanese version, whose government sounds ok to use different materials for training AI, one of the reasons why most of my LoRAs relate to Japanese anime theme), an Alpha version LoRA for anime anime-style scene.

Plan to test further about the painting-style character and scene.
Read lots of awesome artists' masterpieces during these weeks,
Some may be in the public domain, like paintings by
Suzanne Valadon,
Никола́й Петро́вич Богда́нов-Бе́льский,
Sir James Jebusa Shannon,
Alfred Sisley,
Исаак Ильич Левитан,
Frederic Leighton,
John William Godward,
Charles Sillem Lidderdale,
John Singer Sargent,
Alphonse Maria Mucha...
Try to search and have fun with them~
The others out of the range(date or countries) are awesome too, but I should try to avoid using them. (like Christine Lashley, impressive too~)

Read an interesting article about AIGC,
Mean Images, by Hito Steyerl,
I read the translated version somewhere else, then found an English version here.
File:Steyerl Hito 2023 Mean Images.pdf - Monoskop
We did see in 2025, the giants ran out of all resources online without being fully licensed, but provided unreliable/revocable, need-to-pay, and closed-source cloud services, and hinted that their tools are so powerful that would be necessary.lol~ Some topics like this are cool.~

I typed enough documents and plans for my personal game projects, and starting to work further on engines(tried Godot for personal use)/coding now, will show them too if anything interesting/meaningful/useful during development.

Some other libraries and assets we are discussing how to deal with,
I won't post this time, but I will in the future (after the projects' release, I will be 100% able to post them, for we made a deal at the very beginning for such kind of co-working~)