[UDK] [CE3] [TB2] Modular Sci-Fi Scene
polycounter lvl 12
Prologue: Still in the very early stages of building out this Sci-Fi scene but I'm struggling with a few things, such as picking an engine, so I thought I'd just get my thoughts down here on this thread and get some community feedback while I post some WIPs etc.
Planning & Thought Process: I wanted to create a Sci-Fi environment for a few reasons. One is to show off my versatility as an Environment Artist but the other is simply because I haven't done one since my time at Full Sail University.
I really wanted this environment to be highly modular. I'm using Vertex 2 to help guide me along the process as there are some techniques that directly relates to this Environment (type).
Engine Discussion
So one of the biggest reasons for this thread is to try and figure out what engine I want to use. Here I will outline some of the strengths and weaknesses
UDK:
Strengths: Global Illumination, Keyframable Object, Camera, and Particle animation*, Vertex Color/Texture Blending, Decal Placement, Grid Snapping, Fog, extremely flexible and customizable node-based shaders
Weaknesses: LIGHTMAPS AND LIGHTMAP RELATED ISSUES.... AND LIGTHT BAKING. No PBR.
TB2:
Strengths: PBR Shader (best shaders around!), Awesome Reflections, Dynamic Lighting, No lightmaps
Weaknesses: Does not have Global Illumination, Keyframable Object, Camera, and Particle animation*, Vertex Color/Texture Blending, Decal Placement, Grid Snapping, Fog, No node based shaders
CE3:
Strengths: Great shader, Environment probes, Dynamic lighting, Dynamic GI, fog, Decals
Weakness: A whole lot of crap to deal with for maya users, Somewhat complicated, no full PBR, complex vertex painting for maya users,
In-Engine Tests:
TB2:
UDK:
CE3:
UE4:
I WISH
What now:
Going to keep messing around with some lighting and keep developing my modular pieces. If you would like to chime in on what engine I should use and why, please do! I love having these types of discussions!
Planning & Thought Process: I wanted to create a Sci-Fi environment for a few reasons. One is to show off my versatility as an Environment Artist but the other is simply because I haven't done one since my time at Full Sail University.
I really wanted this environment to be highly modular. I'm using Vertex 2 to help guide me along the process as there are some techniques that directly relates to this Environment (type).
Engine Discussion
So one of the biggest reasons for this thread is to try and figure out what engine I want to use. Here I will outline some of the strengths and weaknesses
UDK:
Strengths: Global Illumination, Keyframable Object, Camera, and Particle animation*, Vertex Color/Texture Blending, Decal Placement, Grid Snapping, Fog, extremely flexible and customizable node-based shaders
Weaknesses: LIGHTMAPS AND LIGHTMAP RELATED ISSUES.... AND LIGTHT BAKING. No PBR.
TB2:
Strengths: PBR Shader (best shaders around!), Awesome Reflections, Dynamic Lighting, No lightmaps
Weaknesses: Does not have Global Illumination, Keyframable Object, Camera, and Particle animation*, Vertex Color/Texture Blending, Decal Placement, Grid Snapping, Fog, No node based shaders
CE3:
Strengths: Great shader, Environment probes, Dynamic lighting, Dynamic GI, fog, Decals
Weakness: A whole lot of crap to deal with for maya users, Somewhat complicated, no full PBR, complex vertex painting for maya users,
In-Engine Tests:
TB2:
UDK:
CE3:
UE4:
I WISH
What now:
Going to keep messing around with some lighting and keep developing my modular pieces. If you would like to chime in on what engine I should use and why, please do! I love having these types of discussions!
Replies
You seem to know UDK quite well, light mapping is a pain but provided you know how to do it just make sure you allow enough padding and use a high enough resolution. once you apply textures & lighting most issues are minimal.
I dont think it changes anything, its just a Renderer aimed at the quality of what games engines produce. it takes screenshots (With alpha Masks for selection, if TGA) & Turntable Videos
Thanks for the explaination, I might as well just use Xolilul then.
Main difference between engines for me:
1. You have to f**ck with light maps:)
2. You don't have to.
Second variant is better
Marmoset Toolbag for interiors... Well, i'd like to see your result. I didn't see a lot of interiors rendered in Toolbag.
And what about Unity?
If I think I understand your question completely, the answer is no. TB2 will not effect your maps and apply changes to them- which actually isn't a really dumb thing to think becuase when I was in uni I thought that was exactly what it did. But yeah, it's basically a render engine just like any other render engine. You can make changes to the scene via post process etc, but it effects the scene not the maps.
Hope that helps!!!
@CarlK3D
I used CE3 for my last Environment and I pretty much got it down pat (exporting and shader creating) but there was something about it that just rubbed me the wrong way. I also don't like that I would have to create a "vis-ob" and light portals in order to do the interior properly.
As far as UDK goes, its pretty much the shaders that I hate. I mean, the customizability is GREAT, but the non-PBR workflow sucks for the real-ism I am trying to achieve.... if only I could get my hands on UE4
@Bigglesworth
I'd rather not complete each in each engine and just pick one... but I wouldn't be completely opposed to it as a study.
I agree with you on the light maps haha
I'm trying to be one of the first using TB2 for full interiors. With its awesome shaders, I completely see the benefit of working around its downsides.... or waiting for new things to change (or just waiting for UE4)
oh and as far as Unity goes, I'd have to drop cash in order to get shaders that don't suck... and in order to get light baking and lighting options that are actually valuable I'd need to get unity pro. Not worth it in my opinion.
Tomorrow I'll work on getting this blockout in CE3 just so I can compare all of these and go from there. Gonna push myself to create the vis-obs and all that kind of BS
I still might do a few more engine tests before I pull the trigger but I do like the result CE3 is giving me even with the "wrong" set-up.
The more I think about it the more I'm thinking I will stick with CE3 myself. I really just wish UE4/UDK4 would just release though :poly127:
Tonight I will start to move forward with setting up the scene for CryEngine and see how that goes. I'll also post a few concerns that I have when using CE3 up against UDK that maybe some people can chime-in on. There are definitely a few things I'm concerned about that possibly some light could be shed on.
edit: Actually, just thought of the first issue I was having. When I assigned the glass shder to my glass panels, the light was not passing through at all. the polys were blocking the light. I eventually quickly had to hack it and use the "water" shader just to get the result that I wanted but obviously I'd like to know how to get the light to pass through the glass. Any info would be appreciated!
This is the result I am getting so far. I like the shading. It feels more realistic to me than any of the other options.
At this point, I'm fairly certain I will continue with CE3.
Right now my biggest concerns are:
I'm not certain what the workflow is for importing modular pieces. I'm used to importing large scene in the engine, but I've never used it as a modular kit so to speak. More R&D will be needed on this and if anyone wants to share their methods PLEASE DO!
Second, I have no idea how to do object animation (sort of like kismet in UDK). I am sure I can figure that one out on my own, but again if anyone wants to share their methods PLEASE DO!
Third, is vertex painting. I remember it being a gigantic pain when I tried to do it with my last scene, although I don't remember exactly how it was supposed to be done.
Other than those things, I think the project will be fairly successful in CE3 not do I think it will be THAT necessary or a least a secessary as it would be if it were outdoors.
Now that some R&D is done, I will hit asset creation really hard. I'll elaborate a bit more on my blockout and then just start going to town with modeling and texturing. I wouldn't be surprised to see this thread go dormant for a week or two with me working 9-6 but once I've done a considerable amount of more blocking out, I'll back.
with just light from outside a couple of lights and a environment probe and had no problems. For the walls and railing and such they where all Modular. I usually use decals to break up the repeating.
With the vert blending you can set it up like this http://freesdk.crydev.net/display/SDKDOC2/Blend+Layer
you just have to do your vert painting in your modeling package.
---
Here is an update on where I am so far. I still have a LONG way to go as the Floor and Ceiling have had absolutely no love yet, the pillars on the window walls are super blocky, and most other assets here are in their early blockout stage. The exception being the two pillars as they are quite further along as I got carried away modeling yesterday. =P
Here are the pieces I have created so far.
I have also decided that I am going to make this scene a bit of a DOUBLE FEATURE.
The interior is going to be mostly of my design/craft but the set extension outside will be mostly derived from this concept art
The hard part of this melding is going to be choosing where the mood and lighting is going to end up, but I may just adapt the feel of the concept and adapt it to my idea.
Overall, still super excited about the project and will continue pushing forward!
I'm looking forward to seeing more progress on this though , it was very nice to see some proper comparison shots
Hey Chris!
I'm finding that once you go through the terrible BS of setting up the exporters, the actual exporting goes really quickly. As far as the shaders go, you definitely lose A LOT of the flexibility that UDK offers, but what you lose in Flexibility, you gain in visual fidelity with CE3. Couple that with what you mentioned about the cool lighting features and there are just a ton of trade-offs all-around.
Oh, and thanks for the compliment! I just had an amazing discussion with a couple of my Infinite Crisis co-workers about how I can really push the scene a lot more in both look and feel. I'm really excited to get home and continue working on it!
For those following the thread, expect a steady stream of updates through the week! Also, I appreciate all and any feedback very much! Whether it be about the scene, the engine, or anything really, it's much appreciated!
-Rogelio
To answer your question, the emissive rectangles do have lights associated with them, yes. This way it looks like the tubing is emitting light onto the floor and walls.
Not sure if there is a way to get that effect without fakign it with lights, but it'd be awesome if there was.
---
Change notes:
-Downloaded the newest build of Cryengine 3 that had a ton of new updates and seems to be giving me some more visual fidelity. Apparently My build was old (from 2011) so I really benefited from the update.
-Figured out how to do a more modular workflow that works better for me. I used to just import a giant mesh from Maya with different sub-materials. That method is cool, but I'm finding that re-building this in CE3 with a more "instanced" workflow is turing out to suit me more. Feels a lot more like UDK, which is a good thing for me.
-The new build came with an exporter for Maya 2014. I used to have to import into 2012, THEN into the engine which frankly kind of sucked. Now that my workflow is seamless, I'm definitely having a better time with CE3 than I previously had. From maya, updates are almost instant.
-Changed the lighting to further suit the mood I will be going for. The ambient lighting will change a bit, but the interior lighting is a lot closer to the final feel and tone.
--
I still plan on making the exterior dark, cold, and grungy as a stark contrast to the interior. However, I wont really even start the blockout of the outside until I am a lot further along with everything in this hallway. Once the hallways is close to done, I will start blocking out and prototyping the exterior scene (since it is almost like a separate project anyway)
Otherwise, things seem to be going well. I do plan on having some decorative props in there, I just wanted to have the structure more-or-less finished. Probably what I'll do next is choose a few key story telling props to block out and see where it goes.
Thanks for checking me out!
-Rogelio
Everything looks pretty nice so far. I'm looking forward to seeing some more models! I like the red lights much more than the blue right now. The red and yellow lights look great, and perhaps a dark, cool gray for the interior metals? I'm not sure about the light blue/white lights though, but that could just be because the interior is still a little bright. Can't wait to see where you take it :thumbup:
@leleuxart I completely agree about the workflow and upgrades to the SDK. Really enjoying myself now that I'm rollin!
I definitely can see dark cool gray for metals as well as some panneling a-la deux ex and some painted metal areas as well. Thanks for the suggestions! The blue/white lights are placeholder right now for what I believe will be monitor images (that will vary in color, most likely). Thanks for all the feedback though ^.^
@e-freak Yes! I'm realllly diggin them so far! Much more improved than my older build.
----
Here is a kind of small-ish update. Started really thinking about the story of this place and all of those sort of design elements. This updates focuses on some feedback I got from @JeffParrott about the floors being barren and boring. I went with this sort of installation because it was something I was thinking of doing anyway- having a nice 90 degree slope down into a glass walk-way with a cat-walk like structure supporting it. I quite like it so far and am debating putting some sort of lighting down there or not.
I spent all night and part of today optomizing and converting the scene. Obviously everything is very WIP and the lighting is very blocked out and rough; still lots of tweaking to do!
It feels great to finally be back in Unreal though, I must say. So far the experience has been amazing, and I know it will only get better!
Oh and for the record, I havent gotten much done on this due to an art test and my birthday. About two weeks with not more work accounted for- but I do feel pretty fresh and I'm ready to pick this back up!
Hopefully more updates this weekend!
-Rogelio
Happy belated b-day too :thumbup:
OOooups ! i haven't slept in 58 days ...
How are you finding the new UE4? Have the shaders or lighting changed much? Hows the workflow?
AlexCatMasterSupreme: I hope I get the chance to as well.
leleuxart: Thanks for the birthday wishes!! But yeah, Alex is definitely right, however I just know if I'll have the time. =\
peanut : That's quite a long time sir >.>
Mordin: Thanks! I am absolutely in love with UE4. The shaders are much better being fully backed by PBR, the lighting is a ton better as well. The fidelity of the lightmaps is a lot stronger, the lightmas bakes are A LOT quicker, and the inclusion of the skylight and environment probes really just raise the bar a ton. The workflow is CONSIDERABLY better as well. Overall it has just been a complete joy working with UE4 so far. Can't wait to do more!
Congrats on the job, looking forward to more art updates and UE4 experiences.
Might just be me! Looks awesome though!
[Redacted]
[Redacted]
Thanks! Yeah as far as the pillars go, I might try to axe them in favor of a new asset but I'll defintely keep that in mind and go from there
Also, in case anyone is intersted, here is my Mask and Layered shader for for the floor material
I'm working on a very similar modular project now, and Ii'm having a hard time learning how to build optimal UV and texture sheets for such complicated low poly pieces (I'm coming from a weapon and single-asset background.) I'd love to see a few screenshots of your UVs and the objects they're attached to.
Cheers
@Thaiauxn When I shake some time out I will try =]
----
Looking for thorough C&C for this. Want to put it to bed and move on to something else, but also want it to be portfolio worthy. Am willing to put in some more work to get it to that level, but not willing to really overhaul it at this point.
KNOWN ISSUES:
*Doors are flat planes; they have no detail. Will address.
*Lack of wires. Yes. I realize there are no wires. Bu wires are cliche, and frankly I just don't want exposed wires. I'm still thinking about it.
*Removing the floor lights because I feel they don't fit.
Please feel free to address anything else! Bombs away!
A few things to try:
- Think about how things would animate. The floor lights could have a gradient running through them towards the door. In a still frame, you'd see the further light in each strip lighter than the ones closer to you, leading the viewer towards the door.
- Break up the symmetry. Cables are cliche, but can serve a designated purpose. Could be anything else just lying around or hanging down that would spice the composition.
- Balance & push the color scheme: maybe a touch of stronger greenish-cyan would work well vs the strong red (or maybe not). Try to shove one such pipe among the darker ceiling pipes.
- Focus the composition: it all looks great, but what's the important stuff? it looks like a pre-final-boss, lavish top-floor corridor to me, although the doorway is small and unremarkable. Look at all the pure-white areas you have in the scene, how they attract attention and how to balance them.
Thanks for the C&C !!
-The idea of thinking about animation and leading the player towards the door is a great idea!
-You might have sold m on wires =P But overall breaking up the symmetry with props is something I absolutely should do.
-I was actually thinking about making that far back wall more of a point of interest. Maybe making it a really nice large important sci-fi door would suffice? Either way, making it look like the whole scene is leading towards that big important area might be the direction I want to head.