Is Sub D/Poly modeling dead?

5rettski

I know I made a few posts about poly modeling and subd modeling to try to learn new things, but even if I try to look for timelapses, or videos of people modeling stuff, just to watch them to pass the time. Whether it be characters or even some weaponry modeling, everyone uses sculpt and retopo. I know that's been used for some while now, but still, can't seem to find subd modeling anywhere anymore.

kanga

sub  model a character

Neox

5rettski said:

I know I made a few posts about poly modeling and subd modeling to try to learn new things, but even if I try to look for timelapses, or videos of people modeling stuff, just to watch them to pass the time. Whether it be characters or even some weaponry modeling, everyone uses sculpt and retopo. I know that's been used for some while now, but still, can't seem to find subd modeling anywhere anymore.

overwatch is mostly subd modelling. far from dead

SubD is the reason overwatch look so damn clean. zBrush is very much an afterthought in that production. and one can certainly see if a skin/character was done using mostly zbrush or modelling.

everything the light touches (hehe) has a basemesh and is subdivided twice on this one.

various artists use various approaches. and yeah some of these parts have been sculpted first, and the received a retopo to create the basemesh. but being able to subD model is THE core value of this project.

pior

[...] "everyone uses sculpt and retopo. I know that's been used for some while now, but still, can't seem to find subd modeling anywhere anymore."

Well, the real world isn't limited to Youtube or to whatever gets posted on the internet. You'll get a better feel for what is actually useful for you to learn by joining some collaborative projects, or by working on some Steam Workshop items.

You'll also learn faster and better by trying things out first-hand. It only takes a few attempts at "Zbrushing everything" on a character similar to what is posted just above to realize that sculpting it all is a dead end for efficient and clean asset production - always been, and always will be.

But regardless : worrying about creating highres models for baking is probably not the best thing for you to do ATM. You have a mediumpoly Kratos fanart bust to get done first !

FrankPolygon

The short answer is: Yes...

But it also depends on the artist, art style, industry, role, specialty, etc.

So, the short answer is: No...

In reality a lot of things are contextual: the answer isn't absolute or mutually exclusive and isn't dependent on what tutorials are trending. Framing it that way makes the premise of the question seem a bit contrived.

Content creators, who make more tutorials than art, tend to focus on the latest tools and workflows because that's usually what the largest audience is searching for. Since there's already a lot of poly modeling and subdivision modeling content there's not a lot of incentives for anyone to create more of it. Just because content creators aren't creating new tutorials on these subjects doesn't mean older workflows are completely irrelevant or aren't being used.

At this point there's at least a decade's worth of video content available on these topics and the hardest part is sifting through it to find tutorials where working professionals are teaching best practices. Older tutorial content may be a bit dry or boring but the information is still relevant. Mostly because the workflows haven't changed that much in the last half decade or so.

Here's a few examples of professional hard surface artists (working in games, VFX, concepting, etc.) that make content talking about their poly / subdivision modeling workflows:

https://matthiasdeveltere.com/
https://www.artstation.com/devmatt3d
https://www.youtube.com/@matthiasdeveltere4192

https://www.torfrick.com/
https://www.artstation.com/snefer
https://www.youtube.com/@TorFrick/videos

https://andrew-hodgson.com/
https://www.artstation.com/andrewhodgson
https://www.youtube.com/@AndrewHodgson3D/videos

https://makehastecorp.com/
https://www.youtube.com/@MakeHasteCorp/videos

https://www.behance.net/colorsponge
https://www.artstation.com/colorsponge

https://www.altcinc.com/
https://www.artstation.com/ashthorp
https://www.youtube.com/@alt-c/videos

Also, here's a link to a previous discussion about the relevance of different modeling workflows for games:

https://polycount.com/discussion/234693/

It's worth mentioning that poly modeling is still a foundational technical skill that has significant overlap with other skills like subdivision modeling, manual retopologizing, etc. and it's certainly possible to make the argument that it's still relevant to contemporary hard surface modeling workflows like boolean re-meshing and mid-low poly bevel baking. 

Both workflows still require block outs and base meshes. If the model isn't being exported from CAD then it needs to come from somewhere. I.e. poly modeling, subdivision modeling, boolean modeling, etc.

Examples of bevel baking workflows:

https://marmoset.co/posts/toolbags-bevel-shader-in-action-3-artists-weigh-in/

https://polycount.com/discussion/217834/

Example of workflows evolving from subdivision to re-meshing:

https://polycount.com/discussion/comment/2418170/#Comment_2418170

https://polycount.com/discussion/203187/

https://polycount.com/discussion/168610/3ds-max-zbrush-proboolean-dynamesh-hardsurface-workflow-tutorial/p1

Examples of non-destructive block outs and streamlined, modifier based subdivision modeling:

https://polycount.com/discussion/comment/2769252/#Comment_2769252
https://polycount.com/discussion/comment/2777123/#Comment_2777123

https://polycount.com/discussion/comment/2779298/#Comment_2779298

There's also single application modeling and re-meshing workflows in Blender that enable using poly, subdivision, and boolean modeling together (along with a lot of other non-destructive modeling operations) in a way that makes creating complex hard surface shapes incredibly easy. Native non-destructive re-meshing and polishing modifiers streamline the high poly process into a very flexible workflow.
https://www.artstation.com/learning/courses/JQG/basemesh-and-high-poly/chapters/1Qg6/procedural-basemesh

https://www.artstation.com/learning/courses/JQG/basemesh-and-high-poly/chapters/GM5A/creating-the-highpoly

These different workflows require the same basic modeling skill set and that overlapping skill set is largely transferable. A game artist's hard surface pipeline doesn't have to be limited to a single modeling workflow for every part of a model.

If it makes more sense to use subdivision modeling for curved parts and boolean re-meshing for everything else then that's fine. Both of those workflows can be used to create the high poly for the same asset. Same goes for CAD and other parametric modeling workflows.

Most of the time it all gets baked down to a low poly with a normal map anyway. Deciding which modeling workflow to use really just comes down to what the technical and logistical constraints are... Along with tool set, personal preference, and art style.

Digging into the question: is poly modeling still used to create game art assets? Yes. Is subdivision modeling still used to create high poly game art assets? Yes. Is the subdivision modeling workflow (in game art) as prevalent or relevant as it used to be? No. Does that really matter though?

Even if subdivision modeling was completely irrelevant to creating games or any other media, there's still nothing wrong with learning how to do it. Provided that's what you want to learn how to do. Sometimes over intellectualizing things is just an advanced form of procrastination. If the goal is to learn subdivision modeling then avoid finding or making excuses for not practicing and getting feedback.

Set a goal, pick a project, jump into the modeling process, and solve the problems as they occur.