Abandoned Swamp House Environment - [UE4]
polycounter lvl 8
LATEST IMAGES:
More images/info/updates on last post...
__________________________________________________________________________________
Hey guys and girls,
This is my first real post. I am a 3d generalist and want to make the transition to an environment artist. I have been doing some tutorials and reading through tons of post here and elsewhere. I have come to the point where I think I gotta start an environment. I have decided to work mainly from this concept from Andree Wallin. I will add and remove stuff as I see fit.
My main goal for this project are:
My plan is to start with a few basic props which will allow me to grow and learn with smaller items that could be replaced at the end easily.
My starting prop list includes:
In browsing these forums I can tell I have tons to learn but I think with a little help from you guys I can figure this stuff out.
Thanks for all your feedback, crits, and patience in advance. I'm sure I'll have plenty of questions.
..oxy..
More images/info/updates on last post...
__________________________________________________________________________________
Hey guys and girls,
This is my first real post. I am a 3d generalist and want to make the transition to an environment artist. I have been doing some tutorials and reading through tons of post here and elsewhere. I have come to the point where I think I gotta start an environment. I have decided to work mainly from this concept from Andree Wallin. I will add and remove stuff as I see fit.
My main goal for this project are:
- Create, at minimum, a realistic swamp house ext. diorama.
- Develop/learn prop/asset workflow.
- Learn to create tileable textures in Zbrush.
- Learn/implement modular workflow.
- Build the environment in either CryEngine or UDK.
My plan is to start with a few basic props which will allow me to grow and learn with smaller items that could be replaced at the end easily.
My starting prop list includes:
- Oil Barrel
- Wooden crate/box
- A rock
- A tree
In browsing these forums I can tell I have tons to learn but I think with a little help from you guys I can figure this stuff out.
Thanks for all your feedback, crits, and patience in advance. I'm sure I'll have plenty of questions.
..oxy..
Replies
Suggestions? Feedback?
but if it is a single barrel then that seems fine, never seen it done before through zbrush, so should be an interesting workflow.
No workflow is the best workflow, its all down to the artists abilities and strengths
That said, don't hesitate to try different workflows and see what happens. Just be mindful of time, quality and flexibility. Eventually you'll find what works best for you.
First, I think the barrel is looking good but doing the rust in zBrush is going to make it fairly difficult to get a realistic looking rust pattern and paint chipping. My preferred workflow for something like that would be to push the model further in Max/Maya and focus on getting the construction of the barrel on point, particularly where the metal seams/folds are and the opening(s) on the top which could even be a separate object. Trying to do that in zBrush, while certainly possible, could easily result in getting your details melted together, which appears to be happening on the top. In zBrush, I'd focus on getting your dents and major scratches but don't go overkill with rust or paint chipping. In fact, exaggerating the dents at this stage will probably make them read better in a normal map and make for a stronger prop in the end. After doing all that, I'd bake the map down to my low-poly and head to Photoshop to finish the texture and do all the rust work here rahter than through polypaint. Photo overlays will be immensely helpful in getting a natural rust pattern. Note how the paint wears and rust appears at the exposed edges and where scratches have developed. Focus your rust detail into particular areas while leaving others more bare to make it appealing (your bottom right barrel reference is great on this). From there, I'd put a fun graphic and colors on it and just tweak shader/specular settings to my heart's content.
As far as shader stuff goes, I think vertex blending on this asset would be overkill unless you were planning on texturing it through tiling metal/rust textures that would be reused across a number of assets. Something that might be of more value for a prop like this in a real-world application would be a color variant shader that allows you to swap out the color of the barrel as you please while using the same shader. Of course, that'd only really be useful if you wanted a variety of barrels in your final scene, which doesn't appear to be the case (though it could just be fun to do). Vertex blending is hugely important these days, I believe you should save it for props that should utilize tiling textures, such as a tree trunk or all of the architectural pieces you'll need for the house.
SaboR1996 & Chase: Thanks, as per Axios' suggestion I think I will save the vertex painting for rocks or tileable textures further down the road. I gotta get the basics down first.
jStins: Yea, going to do all the texturing in Photoshop now. It is more of a traditional workflow that I should prob get down first.
Axios: Thanks for all the suggestions and feedback. I have decide to follow most if it. I added more detail on the lid in Maya then moved to Zbrush. I think I gave the holes and rims enough sharpness that it held up when dividing it in Zbrush. As you suggest, I will save the vertex painting for later on.
Here is where I am at... I am pretty happy with the basic sculpt. I made a (super) low poly barrel and tried a bake. I am coming up with some baking issues. I realize that I have made the low poly WAY to low. What would be a ball park for an asset like this for a current gen game? 800-1000 would be my next guess. I can tell that I am getting some issues because the details and roundness of the high poly don't have anywhere to go. I am pretty new at this normal mapping thing. I decimated down 2-3millon polys then did the baking in Xnormal. I am still not sure how I need to tweak the cage though.
Any feedback and suggestions would be great. I will try to do some more research while I'm at work.
I wouldn't model the details on the lid/top of the barrel at all, you can have that in the normal map only, but your overall shape of the cylinder needs a lot more subdivisions.
The top detail is basically on a flat plane and it doesn't show up in the silhouette, so no need to actually model it. Spending those tris on the round shape of the barrel would have a bigger impact on the overall look of the prop.
I still don't trust my normal map baking skills enough to give you feedback on that though
It appears to be some sort of waviness, might be caused by your super low poly cylinder.
Basic info on the topic and possible errors: http://wiki.polycount.com/NormalMap?highlight=%28\bCategoryTexturing\b%29
And another one: http://www.polycount.com/forum/showthread.php?t=107196
I really hope you'll pull it through and end up with a full environment (;
Here were go.
Thanks again for all the tips and suggestions. I ended up using zRemesher to get a basic low poly then cleaned it up further in Maya. I think textured everything in Photoshop. The main reason I started the texturing in Zbrush the first time was because I didn't know I was going to get normal information for the rust and smaller details. Well I took the time to research and found NDO. That program is fantastic. So I just kept all my texturing layers separate then created a detail normal map in NDO and overlayed it on my baked normal. I'm pretty happy with the results. The spectacular is pretty subtle as this is a pretty wore barrel but you can def notice it when moving the camera/light. I'm going for a realistic look and didn't want to over do it.
I'd love some feedback and nit picking on where or what I could improve or techniques to keep in mind moving forward.
I am off to tackle the next prop, a old wooden box/crate.
When creating assets or content do you guys suggest keeping maybe one master file for props in Maya with everything in there so you can manage and tweak assets in one Maya file or should I keep separate files for each asset? I like the idea of being able to see and edit multiple props at once but have noticed that editing and updating textures is a pain if UE4 is looking for the textures in a FBX which then requires me to re-export from Maya. My other thought is, instead of importing materials from Maya, is to just export without materials in Maya and build them from scratch in UE4 as to make for quicker adjustments.
Would really appreciate some insight here.
Thanks!
For UE4 workflow, I find it best to work on the scene as a while in Maya, and then export the assets individually from that scene as .FBX files.
If I am baking normal details then I tend to keep the high and low versions of the model in the exact same location, and I have them both in a group in the outliner.
I will have for example a group called "door", then within that group is "door_High" and "door_Low". I just use ctrl+h or shift+h so show or hide the particular one I am working on.
This way you can see how the assets compare to each other without having to do it in UE4.
This is just my workflow though, I'm sure everyone has their own methods ^_^
I haven't used UE4 myself, but I'd encourage you to just export individual assets into the engine and do all your material work in there. Do as much as you can in the engine if that's where you intend to have your final presentation as how it looks there is all that really matters. Depending on the asset, you may want to make some slightly more advanced materials like vertex blending for things like wooden walls, floor, posts, dock, or door that could share the same textures.
I have been pulling my hair out trying to fully understand normal map baking and all the details and best practices. I love 3d art but this normal mapping stuff is driving me crazy. I am still having trouble learning it. I was making nice progress on the crate and in the middle of texturing I saw my AO map not lining up as well as I wanted/expected on the UV seams of the planks so I went back and revisited my baking/low poly. I tried hardening the UV edges in hopes of that helping but I am still getting issues.
I have read and re-read Earthquake's posts on normal mapping (http://www.polycount.com/forum/showthread.php?t=107196 and http://www.polycount.com/forum/showthread.php?t=81154 ) and each time I learn more but they are really technical and mind stretching for me.
I would love somebody to help me understand what I am doing wrong and how/why another way is better. I put the bevel on the low poly as to try and make the edges smoother but I haven't really seen others do that so I am not sure that's the best method.
Can this be improved or is this about as good as it gets?
I'm using Xnormal / Maya.
Thanks in advance.
Your bevels on the low poly do look a little extreme. It may just need to match the high poly more. There's nothing wrong with having beveled edges but make it match the high poly as close as possible.
Are you going for PBR or just staying with diff, spec and norm?
@luyhyn. Agree about the noise on the painted parts. Went back and compared to my ref. I also want to bring out more spec on it. I did think about modeling some nails might do that when I revisit props. I plan on using PBR still learning it but that crate uses Dif / Roughness / Normal right now and just a constant for reflective since it's basically all wood except for the rusted nail heads. Didn't figure it was worth a metalness or reflectivity map.
Off to learn the basics of sculpting/texturing rocks.
http://seblagarde.wordpress.com/2012/04/30/dontnod-specular-and-glossiness-chart/
I'd suggest having a spec map, it honestly wont take too much time because you'll probably have a flat value.
I think giving this an accurate spec map (and giving ALL the spec to the nailheads) would just make this look a little less... dated..?
PBR is all about being physically accurate, so you can't leave out a part of that process or it's kinda moot.
I would suggest roughing out the environment and laying things out before you jump into creating assets for it. Get a feel for the space, get some preliminary lighting in there, and place some roughs for your props. Then you can have more information up front for what you need out of your props. For instance..your crate. I don't see one in the concept..do you know where you are going to place it? Is the size you are making a good size? Or will you find out you should have made a larger/smaller/wider crate to fit the space? You may find that you are putting too much, or the wrong detail in your props.
If you are primarily worried about making some awesome props, then you can ignore what I said and focus on awesome props, but if you are worried about the whole environment, I would suggest thinking about the big picture before you start heading right into the small stuff.
@direWolf: I love the concept but I think it lacks a little in department of hue variation. I mean everything is green, brown, or grey. It's almost monochromatic. My thought was to add just a few props/assets that had some complementary color to the scene. Just concerned that it will feel a bit to bland otherwise. I'm open to being convinced otherwise though.
@gsokol: You got me rethinking here.
I initially started with props for a few reasons;
1) I have been out of 3d for almost 2 years and am rusty.
2) There are so many new techniques and pieces of software that I wanted to get a basic grasp on (zBrush, NDO, Marmoset ect).
3) I could make mistakes and get some general workflow feedback on props that I could learn on, thus getting some mistakes and growing experience out of the way before starting on textures and modular stuff that would be seen everywhere.
My plan was then to go back and replace the props pretty easily because I would be not only up to speed with the process/software but would also have the visuals nailed down. I initially chose those four props because they touched on some of the main texture that I would be required to make, wood, stone, and metal.
You bring up some good points though and maybe I am starting to feel comfortable enough with the basics of the game art workflow that I can get started on the actual environment. On top of that, I want to become an environment artist, not a prop artist. With that said, I am going to follow your suggestion and do a block out and try to break down the concept by doing a modular texture paint-over.
Here's a rock test sculpt I was in the middle of.
Thanks so much for the feedback/thoughts guys!
So I decided to go ahead and block out the area I am going to focus on. I will try to focus on getting the money shot to be the same as the concept while fleshing out the area enough to get a few more decent shots that capture the houses and props.
I have added a few props that aren't in the concept to give a area a little more color and lived in feel. Comments and crits welcome.
The two individual tasks that look like they'll take up the most real estate of your final product is your terrain plane and blending and the wood setup for all your structures. Personally, I'd dive into the buildings simply because its a more straightforward task and will be a good learning experience. You're going to want to come up with a tiling wood material or two that probably have some vertex blending capability to use on both wooden houses, the dock, and whatever else fits the bill. You certainly do not want to 1:1 those houses or dock like you did the crate or else you'll likely see the quality drop and the workload rise.
Your blockout and scene look good so far, but this concept does touch on a lot of different disciplines so it's a formidable amount of work (but I suppose that's good given your desire to cover all of it). Just take it piece by piece and don't get hung up on the little stuff. If you hit a wall on any one task, move on to another one and work your way back; there is plenty to be done.
It's taken me a bit to get reacquainted with UE4 but man it is a dream to work with compared to Source Engine. I moved my blockout over to UE4 and began tileable textures. Here are some textures made in zbrush/quixel and applied to the house. Left my 9-5 to pursue env / level art so should be able to make good progress and update this a few times a week.
Crits and feedback welcome.
[SKETCHFAB]159adea17da346ff87ecfc9f3be98ae3[/SKETCHFAB]
Here is a rock I made yesterday. I'm actually really happy with it. Spent some time going through the 'rock thread' and picked up some good tips. It's sculpted in ZBrush using mainly the Trim Dynamic and Trim Smooth Border brushes. I made a few custom rock alphas to ensure it had some uniqueness to it. I baked in Xnormal and textured it in Photoshop using various masks and tileable base textures from a small library I made. I can't believe how relaxing and fun it is to make rocks. I plan on making 5-7 of them in various sizes and then maybe make 1-2 tilable ground textures from them for the terrain.
Out of curiosity, how did you approach blocking out the terrain? I know it sounds a bit daft, but I've never done an exterior environment and I'm never sure where to begin with terrain. I start blocking it out, think it looks weird or that my method is stupid and give up in frustration.
It's ridiculous, because If I could just get past my mental block of having a Terrain plane that I'm happy with from step 1 I'd be much more comfortable continuing.
You seem as though you figured out a way, relatively painlessly, so would you be willing to share?
I am assuming you are talking about the blockout I did in Maya? I think it was easy because I am working off a concept. I started by setting down two blockouts that were my houses, then trees and other larger landmarks. I quickly threw down a locked camera and moved stuff around until I got the composition I wanted. I then tweaked the camera focal length / fov until I got a placement of objects that worked as a mini level and also met the visuals of the concept. Once that was done I then made a large plane with maybe 50x50 subdivisions and moved around vertices to get my roughed out terrain. I also had a flat plane with a 50% transparent material on it that was my baseline for the water. I then exported everything at a single FBX into UE4 and just that as my 3d reference so that what I do in Maya and UE4 matches in scale and translation.
I hope that makes sense and is helpful.
It does, thanks! And yeah, you're spot on. It was mainly the plane used as terrain that throws me, and It's likely a large part of that with me, is lack of fixed camera, which I hadn't considered.
Thanks, looking forward to seeing this finished!
I'm planing on 4 terrain materials at minimum, Mud(+ wet version), Rock, Forest Floor and Dirt w/ sticks. If I need more after those I'll cross the bridge when I get to it. Here are the solid rock and mud textures. The both use height map blending and the mud has a wet version.
I have moved on to foliage... Spend most of the past few days brushing up on foliage workflows and picking out the specific trees and plants I plan on making. Here are some shots of my first pass on creating grass... This is the short version, I will also have a medium and tall version to mix in. I hope to have at least all the grass versions done this week and maybe even some plants.
looking forward to this
As for the grass, the planes are to big, they are looking really bad. The grass is thick as hell. You have really nice terrain textures, and grass is sticking out. With thickness, plane size, and hue.
Rocks need more contrast. Its mostly white now. Altough sculpt is good enough.
Subbed: lets see where this will go.
Thanks guys!
@Mr Significant
Thanks for the crits. Yea, that was my first attempt at grass. I already have made the planes half the size and thinned out the grass a good bit as well (see wip in maya shot) Hope to have some ingame shots over the next few days. Fair crit on the rock, at least in some of those shots... See #29 or the sketchfab and see if you still think it needs more varity. I think it may be UE4 with it's variable exposure blowing it out a bit. Either way could def add more to it. Trying to bring this all together at once then go back and fix/tweak some of the weaker spots.
Isaw your sketchfab model, and still dont see the point of rely on skethfab viewer
Since then I have been working on foliage here is my updated grass and some poison ivy. More screen shots coming tomorrow.
As for the rock being too white. It is definitely the lighting. Unreal 4 has auto HRD/Exposure and it can blow things out. See the shot with the house in it for proper lighting.
Great point.. Fixed it.