How do you create giant, non-repeatable environment structures for games?
polycounter lvl 4
Here, an example. We're assuming that the player WILL be able to get close to every part of the structure. So you can't just disregard the texture quality above 5 meters, because "the player will never see that anyway". The player will see that, because they can climb it.
I guess the main block of the structure could be texture bombed or assembled out of many objects. However, what about the circle? How to achieve that without making an enormous object/decal?
Replies
It doesn't, because the object in the image is between, let's say, 500-1000 meters high. It's bigger than entire locations in many games. I can't imagine sculpting, baking or painting it as a single object.
Make it out of lots of small bits
I think you misunderstood my answer above.
https://www.artstation.com/artwork/3dmL5Y
https://www.artstation.com/artwork/2R36a
The AC Valhalla one even uses a nice combination between modular and Masks.
If some of the explanations given here make no sense to you, that's ok - you're delving into tech art with something like this.
I'm interested in the why. Is this personal or professional work? "Its huge and has to look great from all angles and up close" is the kind of vague client request that hurts my soul, so I hope for your sake its personal work...
Identify repeating parts, turn them Into modules, fill the gaps with unique geometry apply some vertex/mask based blending, do some fiddling with shaders to apply some more variation and you're done .
Plenty to get your teeth into, none of it's rocket science
It's neither really. I'm trying to switch from advertisement/product visualization industry to games and I'm learning environment art. And this is one thing I can't figure out and it bothers me to no end when I don't know how to do something. I want to avoid a situation where I get an assignment like this and all I could say would be: "yeah, here is a thing - I have no idea how to do that", and they say: "well, you should, aren't you an environment artist? I think you should rethink your future in this company".
Change your mindset. There almost always be situations that you aren't comfortable in, but those are called challenges, and it's what you learn the most from. You're likely working with a team, both with people at your role (enviro artists), and overseeing roles with some deeper knowledge (tech artists, developers, art directors/leads). One person or the other will come up with a solution, if not, a step in the right direction.
And if you're working as an independent freelancer with no team to bounce ideas off of, then it's up to the client to explain the tech limitations. Unreal? Unity? WebGL?
Anyhow, lots of solutions in this thread for you to dive into. And check out our wiki http://wiki.polycount.com/wiki/MultiTexture
I think what gets missed by amateurs/students (I was the same!) is that a huge part of game art is misdirection. When you're making something for your portfolio, you want it to look great from all angles, up close and far away, which is fair enough when you have a free camera and you're trying to get a job. Its different in production - "You'll never see that in-game" is a well worn phrase...
So, to take your example of being asked to make this as part of your job, you'd have a lot of supporting information:
- The kind of camera your project uses, and how close it gets to the environment.
- The mechanics the player traverses the world with and how fast they move.
- The way the rest of the world is built and streamed.
So, you already have a ton of info that will guide and constrain how this gets built. There will be conversations with tech art and world building. There might (probably) be revisions as you all collectively work out how to execute this thing.Embrace the doubt; this sort of problem solving will be a big part of the job
My advice: just try anything. Don't wait until you got every step figured out in your brain. Just form a half concocted idea and start working. Once you have just an ounce of "paint on the canvas", it becomes much easier to bring it all together.
Chances are, this first time that you finish won't be the very best way. But once you've got something that works, it is then much easier to refine and do a better job.
Most of the time, that second or third iteration is where I've discovered the "typical" way people do these things.
It is frustrating and tiresome but it is more productive to work like this than to sit around thinking and planning and phoning the audience, in my experience.
Anyway, there is some motivational, but more specific to the problem I agree that the object looks more like it's in the tens of meters tall, not hundreds. I'd just make it in a few distinct parts. I wouldn't worry how many material sets it uses. Just make it look right, and then figuring out how to repack and optimize later is very easy.
I don't think the size of this object would necessitate using some macro/micro textures that blend based on screen size. I mean you could do that but I would test and see if the most simple approach would work before doing a more involved process. (if you want to learn how to do material stuff like that, check out Ben Cloward on youtube. He shows how to blend macro/micro textures based on either distance of screen size in one of his terrain material tutorials).
And if the goal is to make portfolio pieces to get a job, and you aren't a super technically inclined person... just do something that doesn't twist your brain in a knot. There is plenty of awesome art that you can make which is technically more simple to create. Teams will hire person(s) specifically for more technical problems.