general texturing-uv question (game assets)

goekbenjamin

Is it ok to have "one mesh" and two materials (and 2 uvs oc)?  One mesh like merged together on the edges.

Like "Head + bust" is one mesh, and this mesh has 2 materials and uvs?

Is there a nogo for game assets, in any way?

icegodofhungary

From what I understand you want as few materials per object as possible. But two materials per object isn't unheard of. With environment art people seem to use a tiling + trim material for a single object at times. Objects usually have more than 1 UV channel in a game setting so I think that's okay.

If it's a matter of not having enough room for your head and bust to be on the same channel, there are methods around that. I think UDIM textures is the term? It's when you put your UVs outside the 0-1 range so you can fit more unwrapping in one channel. I'm wholly unfamiliar with that but it may point you in the right direction for what you're trying to do.

http://wiki.polycount.com/wiki/UDIM

Alex_J

You can answer the question for yourself by studying the work of professionals. Link is to site that has a lot of models ripped from many of the latest AAA games. Star Wars battlefront 2 models are really nice to study because they include marmoset toolbag render scene and neatly organized folder directories.

https://p3dm.ru/

Obviously this resource is for study only. I don't believe there is any issue linking it here but if its no-go you still have plenty of options via gumroad and artstation.

short answer - yes you can assign multiple materials to a single mesh island. its standard practice

goekbenjamin

@icegodofhungary

thanks for the response.
i think udims ar not for games?

@Alex Javor thanks i know that site is super good for study.

i am currently working on a realistic character which does have pants with the same amount of tricount as the "leon kennedy"  from RE2, and some people said this tri count is way to much for pants?! so i am not sure, maybe those are mid-low models for video cut scenes? i am not sure how reliable this source is....


What i did is i split the head from the bust, with different uvs and materials of course (like head_mat, bust_mat)

i think thats ok too when i want to have the head to hve s much uv space as possible right?

Alex_J

yep. if there is a large disparity in texel density that might show where the uv seam is so you may need to be strategic where you place it.

you can also find the winning character art entries from artstation competitions, look at the marmoset viewers, and you can find where they split the materials and how big of a texel density difference they had.

Alex_J

i didn't see the first part about pants tri count.

I'd say not to put too much stock in outside opinions. not because they dont know what they are talking about, but because they dont know the full project. so its speculation based on limmited information.

What you have to do is test. You are working digitally so setting up a pipeline for iterations is very painless and easy. You don't have to nor should you expect to get the best solution on the first go. So design your workflow to accomodate lots of experimentation.

So if topology density is a concern, you want to get to teh animation as soon as possible. Test it and see how it goes. quickets and easiest way to test animations for non-tech artist is upload your model to mixamo. the auto rigger works pretty good and you can quickly download every kind of animation imagineable. Then based on that, see what can you improve. Reduce the topology, change it, just experiment and see what you can do better. understand that there is solutions for imperfect deformations that the tech artist would solve though, so don't get too carried away reworking topology a million times. just find tech artist and ask, "is this a problem for me to solve or is it good enough?"

In real life project you have to be careful about wasting time. Dont want to spend time working on something with limited return. Sometimes this means you dont painstakingly follow all the best practices if they aren't going to pay back. For making tiny indie games, this means I cut tons of corners. Just strategic decisions have to be made. But if its AAA they are always pushing to the limit so knowing how to squeeze the most bang for the buck into your model is going to be important. You'll only gain that knowledge by experimentation. (but i think for AAA most important thing is that you are pushing the quality, so if you need to allow excess resources to hit that its probably better to do so versus adhering to arbitrary limitations)