quite an older topic now but I couldn't find any big discussion about it so If you have a link feel free to post it.
Obviously it's still in development and some features are still not available but let's go with the case that UE5 is ready, Would it change the Industry?
There are still quite a lot of questions about it, the main question (for me) is do we have to adapt to that new Technology or will Normal Maps continue being a thing, I am very sceptical about that whole topic. Importing extreme High Poly models without having the game weight 1000 GB?
apparently the UE5 Demo is at 24 GB which is relatively nothing, considering the size and that an average AAA game today weights about 60 - 100 GB, 24 GB seems really good.
And then again, some Software can't handle such High Poly meshes, imagine importing a 50 Million Poly Mesh in Maya, 3ds Max or Blender, and UV Mapping that Mesh? Hell No.
I am also aware that every company got it's own Game Engine where Technology like Nanite (might or might not?) exist, so If we are talking about Portfolio Stuff I think Optimization will remain for a while, however I see more and more UE5 projects with insane high poly meshes on Artstation so I don't know.
What are your thoughts? Will this Technology affect the Industry at all? Should we start stepping over from UE4 to UE5?
Thanks in advance!
Replies
Short answer.. you still have to author the models which is 95% of the effort so it'll make fuck all difference
Longer answer..
Nanite and alternative prefiltered mesh stuff is very cool but contrary to most people's interpretation the primary benefit is not in allowing us to make shit meshes without the performance impact, it's in rendering efficiency.
As an asset artist you basically save a day on lods. You're still going to want to bake most stuff because it's just quicker and easier than dealing with fully built high res models.
The real benefits come from scene composition where a lot of restrictions are going to disappear.
Basically tech artists and world builders can celebrate, for the rest of you it's business as usual
no
I know this is very exciting and the tech is very cool...BUT I worked really really hard to become a 3D artist, basically from scratch and to have all my work potentially be replaced by an unthinking, unfeeling machine that can't even appreciate what it does is extremely depressing for me.
Yes yes, I know that photogrammetry can't replace creativity and you could never create a game like Doom Eternal with scans but why make a medieval house when you can just scan it and just slap it into the game with no effort? Why learn all the intricacies of a gun if you can just scan all the parts and have better textures than you could ever make? I LOVE modeling even humble props like a gas pump, chair or whatever, knowing that my work will be seen in many areas of the game. Part of what makes 3D art exciting is knowing someone put a lot of effort into understanding the asset, what kind of paint it uses, how it peels, how the rust forms...I feel like all of that is going away. Above all, I just love modeling, especially guns. Blocking out shapes and at the end finding ways to optimize it. I love doing that, it's like a video game to me. It gets even more depressing when AI is advancing at a very rapid pace. You now have AI that can replicate specific art styles. Who knows where the tech will be in 5 years.
I was thinking about learning character modeling to become "future-proof" but with MetaHuman who knows if that's even worth pursuing..All that work I put might be irrelevant in 5 years, hell maybe less.
I don't like the trends the industry is heading to. What's gonna happen in 5 years? Every dev is gonna drop scanned assets in their game? Or are they going to tell their AI that they need an object and then it scans google images for pictures and recreates it from those?...
dont be silly,
this is the same as arguing against the development of farming on the grounds that food is more satisfying if you have to go out to the woods and wrestle it from a bear
"I was thinking about learning character modeling to become "future-proof" but with MetaHuman who knows if that's even worth pursuing..All that work I put might be irrelevant in 5 years, hell maybe less."
That's not how things work though... NPCs in a game requiring believable realistic characters are not individually sculpted, retopoed and textured from scratch anyways - they are always derived from master heads established early in the project. MH is not changing any of that, the tech (toolset and models) just makes the process orders of magnitudes easier by providing some excellent bases perfectly setup for animation and facial capture. That's a good thing that all artists involved in that kind of stuff will benefit from. And if you imagined that the job of a character artist at a studio consists of spending all day sculpting details in Zbrush, than that's just you having a false image of what the job consists of :) I would say that the only time when this was the case was probably ... 10 -ish years ago already ? Around the time of the first Assassin's Creed games, early Witcher, and so on.
Now if your goal was indeed to get hired to work on highly detailed hero characters (like in Uncharted, Horizon Zero Dawn, GoW ... ) then nothing changes for you either - as a matter of fact all that means is that you have now direct access to the highest fidelity tech today, for free and directly from your home, allowing you to drive pre-made models as well as your own custom ones with an incredible amount of fidelity. How can someone not see the exciting side of that ?
The people worth hiring are going to be the ones who are able to identify how to use the tech at their benefit ... as it has always been. The only jobs that will disappear are the ones when someone contacts you out of the blue asking you to make a realistic character model for 50USD :D
- - - - -
Now to be fair I do see one downside : by making realistic character imagery much easier to achieve, artists who just follow where the tech goes will not get the opportunity to brush up against some of the fascinating challenges and explorations that low tech from decades ago naturally favored (handpainting, chunky lowpoly models, bright colors, and so on). But that's all irrelevant to anyone only interested in realistic rendering in the first place. But ironically, the best ones at that stuff ... are the ones with strong art and photography background anyways.