Hi I'm a student at Blekinge institute of technology located in the southern part of Sweden.
I have some questions I want to ask you guys.
I have just recently realized that I want to become and environment artist in the future. I'm a novice when it comes to environment creation and in my upcoming school project I'm going to try to figure out what different areas I should study, to improve myself and hopefully be able to work as a Environment Artist in the industry sometime in the future.
I'm doing this so I can adjust my future studies to help myself and hopefully other students to become better environment artists.
So I would like to ask you to share your experiences working as a environment artist.
I would also like to ask for advice on things you think is important for a student to learn? This can be anything you think is relevant, this can be your own tips and tricks and all from tutorials, books and articles and anything else you can come up with. The more you're willing to share and suggestions the better and I would very much appreciate it.
I'm also planing to look up different companies employment requirements but I'm not sure how specific those are going to be and thats why I'm also asking for your opinions here on polycount because I think it's to gather as many opinions as possible when it comes to things like this.
Thanks in advanced /onewinged
Replies
Cheers /onewinged
The Basics -- prop creation: create small and medium sized props with 1:1 textures. Get a feel for proper modeling and texturing techniques. Develop a solid understanding of normal maps, and how to create them. Might also be good to try your hand at natural objects (rocks + plants) as well as man made things. Nature tends to be tricky to emulate.
More Advanced -- Tiling textures and large objects: Work on techniques for creating large objects such as landscapes, buildings, and interior architecture. develop a feel for breaking down large things into modular parts, and using tiled or stripped textures in situations where a 1:1 can't provide enough coverage.
Scene creation -- Putting it all together: Create fully realized scenes in engine. Things like lighting and composition come into play here. Engine specific features may also play a role at this point. Often there are special systems for terrain, foliage, modular instancing, etc.
In reality there's a fair amount of overlap between these areas, but this might be a handy way of looking at it.
1) Start with just a basic prop like a crate, barrel, gun, rocks, chair, etc. Learning the high poly to low poly workflow first is pretty important so having a fundamental understanding of that is key. Learning sub-division modelling, sculpting and such are the first steps to making the high poly. I'll make an emphasis that you're making the high poly FOR the low poly, so big regards in terms of angles, smoothing, details you can condense etc. After you learn how to do that, building the low poly and learning efficient unwrapping layouts and understanding normal maps and baking is critical. After doing that, gaining a fundamental understanding of texturing techniques is key too, going from how to how hand paint details and using photo sourced textures and manipulating them to define your materials. Learning presentation here like lighting techniques, composition and such are the next step. Essentially, make a bitchin looking prop first and then you can move on.
2) Learning about tiling textures and large objects is next I think. Having a good understanding of UV's and unwrapping is critical here, also understanding trim textures and how tileable textures work is very important here. A common beginner mistake is like, "Hey, how about I just uniquely unwrap this skyscraper and then put an AO bake on it, it'll look great! Wait, what why is it so blurry, even with my 4096x4096 texture?!" Essentially, learning Material IDs, modular sheets, decals, blending, trims, etc is what you wanna do.
3) Building an environment! Essentially this is where all those techniques come into action in order to build a scene. Generally the workflow is a white box blockout, followed by rough colour schemes or concepts to figure out the composition. then to building the tileable textures and modular sheets to build it up. Then usually people build individual props in order to break up repetition and such to build it to the final scene. Environments take a while but it's pretty impressive when people can do it well!
Anyways hope this advice helps, good luck!
http://worldofleveldesign.com/categories/level_design_tutorials/how-to-plan-level-designs-game-environments-workflow.php
Ehhh, while I agree that composition and lighting are important, I'm still of the mind that being able to make a good looking asset is essential. To me, if you can't make a good looking prop to save your life, there's no way you'll be able to make an entire environment look good. My opinion is that starting smaller scale makes more sense than throwing a beginner in the deep end first and being like, "GO!" If you can create an asset thats good, then presenting it well with good lighting and composition then you move forward and can worry about the entire scene compositions and environment lighting theory.
Tnx man, you are right, i completely forgot about that site, should pay more attention to it....
Lukepham101 - some great advice!! I also love your Tactical Operations Centre piece in your portfolio. I love the journey I can see your work has taken.
Would you rather have a bunch of great assets composed and lit poorly, or some mediocre assets with good composition and lighting? Slap those high-res textures and DX12 shaders on assets as much as you want -> if you can't make a good looking level with them you've ended up with something arguably worthless.
The things I have gather so far is that I would start with Preproduction following the guide from the book Preproduction blueprint: How to Plan Your Game Environments and Level Designs Tutorial that is available at http://www.worldofleveldesign.com/store/preproductionblueprint.php really good book. And then I would do composition.
Followed by a simple block out stage. Then I would do the actual final assets (includes textures), level assemble (in my case UDK), and then lighting.
I think we're essentially on the same page here in that we both think composition and lighting are critical to making environments look good and I wholeheartedly agree with you on that. Even if you've got the best looking AAA assets ever created to man, if you have no understanding of things like focal points, lighting theory, colour, etc then you're still gonna have a shitty looking scene.
What I'm just trying to advocate here is being capable of the lower level tasks like prop creation as well as a thorough understanding of composition and lighting. To me, if you have a limited knowledge of texture creation, UVW mapping, material definition, modelling and such, then you're not going to be a good environment artist. Vice versa, if you don't have an understanding of modularity, colour, space, etc then you're not going to be able to make very good environments either no matter how great your crate looks.
An analogy I can think of is a bad art lead who have little to no understanding of art creation and gives management deadlines and create schedules based on nothing. Nothing is more frustrating than that and doesn't help in gaining respect either. An environment artist in my opinion should be capable of doing all the tasks below him, like modelling, texturing, lighting and all that. If he can't do those things, it looks incredibly bad if he's like, "I have no idea how to create normal maps except throwing a texture through CrazyBump." Furthermore, while a lot of large studios do require you only to do a very specific role, smaller studios need you to be able to take on many tasks and responsibilities so you can't expect every studio to be like, "Only lighting for you!" This is especially true if the country you live in has virtually little to no big studios anymore. Anyways my point is not to neglect the other aspects that are really important to environment creation!
And also, build up a library for images for the real world places as well- from Urban to Ancient Ruins- anything. It'll probably will take you a good awhile to build up an inspirational library folder but I guarantee you that it will be worth it! Hope that helps
I agree, both are important. My point was that I think composition and lighting are *more* important, as if you nail those you can get away with having sub-par assets. If you're lighting and composition suck, good assets won't save your scene.
I agree here, even though it's awesome to be able to make super detailed meshes everywhere. If you're to make you're an environment (I mean, not a one-angle thing) you have to choose your battles. Spend time to detail the main pieces and leave the rest to less detailed and focus on composition, shapes, colors and overall lighting.
Lighting and composition. As others have stated.
Tiling textures. This is a large one, but learn how to make tiling textures with values and surface details that will work well with a defined palette. Use of photos with good color correction, sculpted/high res baked elements, and hand painted creation.
Shaders for blending tiling textures. Learn how to create or use shaders that allow vertex painted and masked blends between multiple materials and textures. This will inevitably take you back to the last topic which is textures that will work well together.
Baking of unique objects. Understand smoothing, baking process, creating high res objects in 3d package, sculpting, retopoing. Normal bakes, height map bakes, masks, AO bakes, etc.
Photoshop. Seriously. Masking tools, blending modes, color correction tools, painting tools, actions, color channels. If you suck at Photoshop, you will never be a productive environment artist.
These are just a few of the critical components. I'd also recommend working on some natural subject matter (rocks, trees, grass, etc.), and manmade stuff (buildings, manmade props). In my day-to-day work, I bounce back and forth between these things a lot. I'll be making buildings one week, then trees and grass the next, then cars the week after that. So try to cover the most common bases there.
I would contest this point, 95% of our environment artists barely touch Photoshop. It seems a bit weird compared to old methodologies, but I can see art production plausibly moving further in this direction in coming years (at least for top-end graphics work).