Best Of
Re: How do you make the 3D pipeline less painful?
It looks like OP is working at a machine-learning startup for 3d content creation, so that's where this is coming from.
___'s machine learning and in-house art team combine to deliver production-quality assets in minutes.
___ specialises in developing cutting-edge Machine Learning algorithms for transforming images into digital 3D models. Our primary focus is serving game developers, offering a 10x speedup in their 3D asset production pipelines.
Blanked out the name to discourage SEO, but I'm sure you could find it if you looked hard enough.
Anyhow, it's all a bunch of astroturfing. Looking at you AsdiusWork, kostadina, lucaK25.
I get it, this is how you market these days, embedding things. But at the very least treat us with some level of respect and announce where you're from, don't hide behind some pseudo anonymity.
Eric Chadwick
13 ·
Re: Sketchbook: Tim Smith
I agree his head-to-body proportions are a bit off if the intention is to have a taller character.
There are a few other things that could be revised, but the most important is a structural mistake in the head which immediately stands out, specially in profile: Ear placement.
Unfortunately the skull you're using is missing the ear hole, a vital landmark for the head. When it's not being actively considered when placing features and your ref lacks it it's very likely the ear will drift. When it drifts other features tend to also drift, with jaw and sternocleidomastoid and hairline usually being the worst victims.
In this case because the ear is so high the sternocleidomastoid is attaching to the side of the skull instead of to roughly the underside of the protuberance in the temporal bone. He'd also have odd sideburns and beard as it is, and a vast expanse of empty back rim of the jaw below the ear.
There are a few other things that could be revised, but the most important is a structural mistake in the head which immediately stands out, specially in profile: Ear placement.
Unfortunately the skull you're using is missing the ear hole, a vital landmark for the head. When it's not being actively considered when placing features and your ref lacks it it's very likely the ear will drift. When it drifts other features tend to also drift, with jaw and sternocleidomastoid and hairline usually being the worst victims.
In this case because the ear is so high the sternocleidomastoid is attaching to the side of the skull instead of to roughly the underside of the protuberance in the temporal bone. He'd also have odd sideburns and beard as it is, and a vast expanse of empty back rim of the jaw below the ear.
Celosia
1 ·
Idea for Feedback on Tracking 3D Content Creation Pipelines
I’ve been thinking about a simpler, more modern way to manage the flow of 3D content creation, as an alternative to heavier tools like Shotgrid. The idea is a web app that focuses on clarity and ease of use, with features tailored for artists and teams:
- Kanban-style board that shows what stage of production each asset is at.
- 3D viewers in the browser so assets can be reviewed instantly without extra tools. After reviewing, the user can decide whether to request revisions or approve the asset to the next steps.
- Requesting revisions directly on assets, and assigning to teammates or external collaborators. This will include a simple UI to annotate on top of 3D models and add notes.
- Onboarding team members and external contributors, and assigning them to specific steps of the pipeline so responsibilities are always clear.
- To-do lists for each team member, showing just the assets assigned to them.
- Custom properties and filtering, inspired by tools like Notion and Linear, to let teams organize things the way they want.
- Grouping assets into projects and milestones, with deadlines to track progress.
I’d love to hear your thoughts: would a tool like this help simplify your workflow, or are there other pain points in your process that should be addressed first?
kostadina
1 ·
Re: How do you make the 3D pipeline less painful?
I'm thinking it's hard to tell whether if you're working for microsoft or nvidia, and what percentage of the trillions invested into generative AI is allocated to marketing
zetheros
3 ·
Re: Theoden King of Rohan
Finaly finished! Hope you like end result!
I have more shots on my artstation post: https://www.artstation.com/artwork/RKdVkm
drvoseca995
5 ·
Re: Sketchbook: Zetheros
My wife was telling me how kpop demon hunters was a good movie and how I should use the characters there as reference, so I watched it and now all the males are going to look like Jinu. I think the male lips regressed but I'll fix that next week.
Have a great week!
Have a great week!
zetheros
1 ·
Re: Sketchbook: Zetheros
More work done with the male human basemesh, I think it's in a pretty good spot for now. I'm now adding the same changes to the female human basemesh.
Have a great week!
Have a great week!
zetheros
1 ·
Re: BRAWL2 [Vehicle] Rust Rhino, Hang Li
I’ve completed the 2D vehicle design and am moving on to 3D modeling, while continuing to refine the 2D work in parallel. 
HangLi
3 ·
Re: How do you make the 3D pipeline less painful?
Hehe indeed ; and if it comes from a genuine place, then it sounds like "whatifs" from someone who hasn't worked on many models.
To the OP :
"A UV co-pilot overlays stretch/TD heatmaps, proposes seams you can accept/tweak,"
"wouldn't something like this be cool?"
Well not really, since seams are placed (or at least planned ahead) *while* modeling the low, as they very much influence the modeling. Hoping for an assistant to design them for you after the fact is somewhat absurd and wouldn't be much better than the many automatic or semi-automatic UV tools already available today.
What you are describing is close to one-day throwaway models that are not meant to go in the final game. And while such quick models can be interesting experiments (and as said they can certainly be done in less than a day with the currently available tools, ie tools that have been available for about 10+ years already), in actual practice they're not so hot as ingame placeholders since they take about the same time to do as a regular clean blockout ... and a clean blockout has the big advantage of being a starting point for the final model later as opposed to being fully throwaway. As a matter of fact it seems to be what Valve is doing with their Deadlock models. And if production needs to speed up unexpectedly, a clean blockout ready for quick cleanup and texturing will always be more useful than a sculpted model bruteforced into a pseudo-realtime model that needs to be redone from scratch.
Now of course new tools are always fun to think about (and certainly interesting to develop). But I genuinely believe that the time of an artist is better spent on developing efficient working habits with the tools available today, than on daydreaming about the next automagical thing.
To the OP :
"A UV co-pilot overlays stretch/TD heatmaps, proposes seams you can accept/tweak,"
"wouldn't something like this be cool?"
Well not really, since seams are placed (or at least planned ahead) *while* modeling the low, as they very much influence the modeling. Hoping for an assistant to design them for you after the fact is somewhat absurd and wouldn't be much better than the many automatic or semi-automatic UV tools already available today.
What you are describing is close to one-day throwaway models that are not meant to go in the final game. And while such quick models can be interesting experiments (and as said they can certainly be done in less than a day with the currently available tools, ie tools that have been available for about 10+ years already), in actual practice they're not so hot as ingame placeholders since they take about the same time to do as a regular clean blockout ... and a clean blockout has the big advantage of being a starting point for the final model later as opposed to being fully throwaway. As a matter of fact it seems to be what Valve is doing with their Deadlock models. And if production needs to speed up unexpectedly, a clean blockout ready for quick cleanup and texturing will always be more useful than a sculpted model bruteforced into a pseudo-realtime model that needs to be redone from scratch.
Now of course new tools are always fun to think about (and certainly interesting to develop). But I genuinely believe that the time of an artist is better spent on developing efficient working habits with the tools available today, than on daydreaming about the next automagical thing.
pior
5 ·
Re: [WIP][UE5.3] Forgotten Waters
No, no need for new lights-- maybe shadow casters, i.e. invisible meshes that cast shadows. You can place them close to the houses, mostly, and your existing lights should do the work there.
Just an example from google, but this kind of higher frequency dappling:
As for how to lessen the existing dappling, that may be harder to do, if it's actually cast by trees that are in the shot. If they're off-camera it may be easier.
Just an example from google, but this kind of higher frequency dappling:
As for how to lessen the existing dappling, that may be harder to do, if it's actually cast by trees that are in the shot. If they're off-camera it may be easier.
Joopson
1 ·