[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vr / w / wg] [i / ic] [r9k / s4s / vip / qa] [cm / hm / lgbt / y] [3 / aco / adv / an / asp / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / qst / sci / soc / sp / tg / toy / trv / tv / vp / wsg / wsr / x] [Settings] [Home]
Board
Settings Home
/3/ - 3DCG



Thread archived.
You cannot reply anymore.



File: ue4.png (31 KB, 254x207)
31 KB
31 KB PNG
It's time to start a new thread.

Some good new videos are now available on UE's YT channel. Check out this one about lighting, I needed something like this so much a while ago, but it is still extremely useful.

https://www.youtube.com/watch?v=ihg4uirMcec&t=1231s

I also have one question, maybe somebody will know the answer - how can I make my deferred decals have smooth and faded edges?
>>
Do you think Unreal will ever be good at rendering realistic forest environements? It has some of the most realistic interiors along with foxengine but the outdoors are trash (compared to Cryengine and Frostbite). I wish they would put more effort on that.
>>
>>586088

If you havent, check out koola:

https://www.youtube.com/user/koooolalala/videos
>>
Why can't I tessellate my ground plane properly? I followed a tutorial, but it doesn't want to make more triangles after about 0.2 and that's way too low. I don't get it why it suddenly stops and doesn't want to tessellate any more. I'm tessellating a single-polygon plane from 3ds max, is that proper way of doing it? I even tried with turbosmoothed version of it, but it doesn't make any difference.
>>
>>586437
Nvm, figured it out. The plane did need more polygons and it worked when I reimported it again, because every polygon has a tessellation limit.
>>
I'm trying to make a rain with particle system, but no matter what bounds I set, in the scene my rain stays in the same position and always falls in a tiny area. I also can't see it from every angle for some reason, only from one side and when I look directly at it. Wtf is that and how to fix it?
>>
>>586825

This was always a problem for me too. The trick is to attach the particle system to the player camera so that you can always see it. In previous UE4 versions you could resize the particles to make them cover your whole level, but it stopped working starting from 4.14/4.15
>>
>>586825
Change your sprites' material to two-sided in the material editor
>>
>>586088
That would defeat the purpose of lightmass. I dont see how they could do really decent lighting for big outdoors scenes without making a totally new lighting system, or somehow combining both VXGI methods with lightmass.
>>
>>586920
>>586947
Oh, cool, I'll try both of these. Yeah, totally forgot about double-sided material, although that still won't fix the fact that it keeps falling on 5 square meters of land. Attaching to a camera seems good, yeah. Didn't know it stopped working for the latest versions and ofc, there no new tutorials for it so I'd never figure that out.
>>
>>586985
Although when I think of it - what would happen if I go into some house and watch the rain from the inside? That method wouldn't work in that case, and even then, you would still notice that it is raining just around you, and far away there would't be any rain.
>>
>>586153
https://youtu.be/E-FpKzD9_uM
the snow one always got to me, read his workflow
takes alot of time but its well worth it
wonder when we will see this implemented in games.
>>
>>587024
Damn, didn't know he provides that scene for free. Gotta check it out, thanks for sharing.
>>
Working on a toon shader that doesn't use Post Processing, doesn't modify any engine code, and has support for shadows despite using mostly Emissive properties.

There's still bugs that need to be fixed, but I hope to get them fixed soon.
>>
File: FB_IMG_1508281998316.jpg (19 KB, 720x405)
19 KB
19 KB JPG
>>587072
Shadows required me to use the forward renderer, because of course Deffered doesn't have support for grabbing just shadows.
>>
File: FB_IMG_1508282020054.jpg (16 KB, 720x405)
16 KB
16 KB JPG
>>587073
Only downside with this method is the artifacts with it, which could probably be fixed with some tweaks to the material function that I am using to handle shading and shadows.
>>
>>587072
how optimized is it
>>
>>587085
also, when is that fuckin arnage shader beeing implemented into the base engine, jesus fuck
>>
>>587085
On the level which the shadows are shown on the shapes, it runs at 300-500FPS with a GTX 1070 and an i5-7600k. At least that's what I could determine with the unlocked FPS counter flashing between different numbers.
>>
>>587087
Due to the way shader plugins work and how you can't replace any code in binary builds, probably never.

Also, Epic Games seems to be making the engine more modular, and came up with the excuse of the toon shader by Arnage bloating the engine.
>>
Currently working on a pretty large game. The main focus on it is having about 6 AI's that will always react to the environment, or the player.

How many assets can I have in the game before performance starts to get really bad? Each floor has hundreds of assets, and there's about 6 floors. There won't be much scripting, not as far as I know, except some fairly simple AI's. Most visual stuff like lights and particle effects can be turned off when the player is somewhere else, but the navMesh has to be usable for the AI.

Most assets will be walls and floor, which can range from 10 to 100 vertices.
>>
>>588275
>typical ideaguy
>>
>>588276
Doesn't most projects start out with an idea?
>>
>>588277
>The main focus on it is having about 6 AI's that will always react to the environment, or the player.
>There won't be much scripting, not as far as I know, except some fairly simple AI's.

is this satire?
>>
>>588278
I didn't say the AI's are going to be advanced
>>
>>588275
>How many assets can I have

About tree fiddy.
Come on, what kinda question is this? It depends on so many things.
>>
>>587074
>https://www.youtube.com/user/koooolalala/videos
any eh tutorial ? or printscreens ?
>>
>>588393
I used a modified version of this for my shadow rendering: http://www.tomlooman.com/disneyfaciliershadow/
>>
>>588275
it reminds me of how at the beginning of my learning phase i made a crater out of about 500 rocks with ~12k verts each

assets and draw calls are not the only thing that affect performance you should watch some vids on it
>>
I have a question about lightmaps.
Does every texture set have its own lightmap as well? Or do they all share the same?
If I make a model with 3 texture sets / slots, do I have to make one lightmap unwrap for the entire mesh or for each texture set separately?
>>
>>588891
one for the mesh
>>
i made my future room
>>
It's amazing how I keep failing to create a photorealistic exterior in UE4. I watched so many tutorials, but faggots almost never explain anything in-depth - still, I even setup my lighting the same as they do, even use the same textures for grass and such - and yet it still looks like a piece of garbage some 13 year old kid made in his free time while working on a tiny MMORPG project.

Why is there not some 50 hour long autistically detailed course about this shit?
>>
>>589943
Post pics
>>
File: shit.jpg (2.02 MB, 1920x974)
2.02 MB
2.02 MB JPG
>>589945
I'm ashamed to post this trash since I haven't worked on it a lot, but it still looks the same as one other project I spent more time on so whatever.
>>
>>589948
Where is the rest of the scene?
>>
>>589953
This is just for testing purposes, doesn't matter. I don't know if the grass is good or bad. It's bad to me.
>>
How well can this engine be made to perform on older and or low end machines? How can that be compared to unreal 3?
>>
>>589948
if it's not a big scene, sculpt a land outside of ue4.
i just messed around with blender's buggy dynotopo and got a good result exporting it later on ue4 and used vertex paint insead of the landscape thing (you will have to do some material trickery tho)
>>
>>589948
You're not using any ref, are you?

Grass is coming in thick, shrub-like clumps. While this is actual behavior of some desert type grasses, the transition is quite jarring, especially in the mid-ground, because it becomes apparent you have nothing but a tiled terrain and one grass type.

You won't make things fully photorealistic in a game engine, so you gotta spice things up to make them look good. It could help to have one or two rock types that blend with the terrain while adding variation to your ground meshes. Some taller grass (of a different color, or maybe with flowers) could also be used sporadically to create accents.

Your trees look really fucked up. What sort of tree are they supposed to be? Primary branches that jut out from the trunk are really damn straight, almost look like the default spikes in speedtree. Trunk thickness and the relatively low starting branch height also hints at those being some pretty large and beefy trees, which means your leaf size is way too big.

Lighting seems fine. You're missing a background but you obviously already know that.

My advice would be to pick a reference, or a piece of concept, and try to recreate it as an environment. It'll give you a solid goal to work towards, which is really all you seem to need judging from your level of skill.
>>
File: TStE6IGbR1iq7b4HOtyqdw.png (773 KB, 1202x547)
773 KB
773 KB PNG
Remaking SF3 Third Strike's NY stage. Alex version.

What do y'all think so far? not done texturing yet, gotta fix the front of the train, and i have to add the subway driving by in the background. make fix some of the beam textures too, and the lighting while im at it
>>
>>590053
SF doesnt make sense in 2D, because people wouldnt just be attacking / moving on the x axis.

Try tekken. Much better series.
>>
>>590055
im just remaking the stage my man
>>
>>590057
Stupid and pointless
>>
>>590068
thanks for your input
>>
>>589609
>>
>>590053
Judging by how the colors on the mannequins seem fairly normal, your textures are way too saturated.
>>
>>590057
Don't listen to that rude anon. Keep it up, friend
>>
File: _1.png (1.09 MB, 979x868)
1.09 MB
1.09 MB PNG
Making a shield pack. Getting materials ready before creating final variations of different types of shields.

Any criticism?
>>
>>590181
Looks like shit but could just be the lighting. Price for how many shields?
>>
>>586987
Make the rain particle bigger so you can see rain further away from the camera, also you can put a volume in every doorway that toggles the visibility of the particle effect when you go in or outside.
>>
File: ao bake.png (1.08 MB, 979x868)
1.08 MB
1.08 MB PNG
>>590181
shouldnt you bake something where i marked in red?
i dont know too much about this stuff, but can detect when something looks a tad simple. the center of the shield looks nice with that inward bulge
>>
File: _2.png (1.23 MB, 1107x636)
1.23 MB
1.23 MB PNG
>>590401

Yeah that's def. the worst. part. It only looks good with very specific lighting. I am trying out something else, separate boards, what do you think?

>>590200

I was going to release for free if people like it. This is just going to be my first "serious" project for my portfolio if it turns out good. I will be making many variations such as round shield, buckler, war door, greek style, scutum, heater, kite, etc...

Also will have "addons" for shield like the metal front part, but different styles. Also, I'm gonna make the shield "paintable" so a person can add a mask, choose paint color to have "clan" shield or whatever.
>>
>>590428
Wood looks good but metal frame seems a bit low res imo
>>
>>590471

I once asked about texel density while watching a, at the time, popular 3d streamer. I got laughed at and everyone joked about it for 10 minutes.
>>
>>590478

Does that refer to UV space? I try to keep my UVs the same size relative to each other, is that the correct way?
>>
>>590478
But why?
>>
>>590481
it depends on how detailed each uv shell needs to be. generally speaking, the answer is "yes" but there are times when you don't need some uv shells to have the same detail level as others
>>
>>590536
it also depends on if you use a tileable texture or not

for tileable textures, UV shells should be the same relative size, otherwise the texture will obviously be of different size
>>
>>590181
is the red paint a separate layer i can customize both the color and design of?
>>
So, I have a landscape master material with 4 different layers. When I generate my 8km by 8km landscape with a heightmap, there I can choose my master material.

Now - the problem is that I'd like to have one additional texture for a whole landscape, which I got from World Machine, because the textures are different from my master material layers, and I want the mountains in the distance be textured properly and just like I saw them in WM.

What am I supposed to do here?
>>
For a procgenned top-down game where I intend on generating stuff with a tile grid -
UE4 or Unity? I get the impression UE4 might handle procedural stuff worse than Unity but I've got nothing to back that up aside from a hunch.
>>
File: whattheactualfuck.png (800 KB, 1209x531)
800 KB
800 KB PNG
I realized my draw distance is small. As I fly over my terrain, hills and shadows start appearing once they get in the range. You can clearly see the edge. How can I fix this/make it be visible from far away?
>>
>>590714
My mistake, nvm. Looked like an actual geometry is not rendering, but I just increased dynamic shadow distance on my light source and now I can see everything just fine.
>>
Is there a way to disable dbuffer decals and not have the engine completely hang next time it loads?
>>
>>590699

Very simple, just setup a distance based blend
>>
>>589948
secret outdoors pro tip: match foliage color to ground texture
>>
Whenever I undock a window from the UE editor, the frame rate craps itself and goes to <30fps.
First it only happened on my laptop, and I chalked it up to maybe the graphics chip rendering some things instead of the 1070.
But now it's happening on one of our main rigs as well - i7@4,2 GHz, 1080ti and 32GB ram.

Anyone knows this problem or what to do about it? Forums turned up nothing. This shit even happens in entirely empty projects, and we can't work around it because doing everything on one screen just plain sucks ass. Sometimes even hovering the mouse over a window will drop the engine to 8fps.

We're on 4.17 right now but it also happened on 4.16.
>>
>>590744
>dark brown ground
>oh now my grass also has to be dark brown!
Great tip.
>>
Can I add some sharpening to my scene? Everything looks good in realtime, but when I want to take a high resolution screenshot, everything actually looks somewhat blurred, not even close to what you would see from AAA games. Before you ask for my settings, everything is on max, from world and project settings to postprocess and lighting, it's basically a tech demo at this point, I only have just enough FPS to move around the scene.
>>
>>590776
Other way around, your ground ought to have grass texture on it unless they're tufts sprouting out of dirt.
It'd also be more convincing if the dirt wasn't a color that makes it look bone-dry and a texture that makes it look like a bunch of rocks.
>>
>>590776
>hurrdurr

obviously don't put grass on a dark brown texture, that will never look good in any circumstance, in realtime at least
>>
>>590810

You need to use supersampling (screen percentage, in the settings of your camera). Also, open the console and write these lines

r.TemporalAACurrentFrameWeight 0.1
r.TemporalAASamples 4
r.Tonemapper.Sharpen 0.5
>>
>>590810
You probably have temporal AA turned on, that's why everything's blurry
>>
>>590846
>>590844
Thanks guys, didn't know about those.
>>
>>590859
>>590844
Btw, do I have to enter those commands every time I open the engine? If yes, shouldn't I just change .ini settings then?
>>
>>590860

Nan you just have to do it once, unless you create a new project of course.
>>
>>590860

And btw, don't turn Temporal AA off like >>590846
said, the console commands I gave you will sharpen it and remove the blurriness.
>>
>>586083
Is there any way at all to get rid of the really weirdly plasticy lighting model that ue4 uses? It's seems to be almost omnipresent in every UE4 project I've seen and it's the one thing about the engine that I absolutely hate. No other game engine has lighting this weird, not even udk. Even unity's base lighting model looks more visually appealing than this.

I feel like I'd have to modify the source code to do accomplish it but i'm not a coder, though I'd be willing to learn if it will let me fix this shit.
>>
File: diffuse_indirect.jpg (232 KB, 1918x1130)
232 KB
232 KB JPG
>>590892
image for reference
>>
>>590892
You could, y'know, not use the default lighting? Look up how to set up proper lighting in Unreal and set up materials to use things like PBR. Post processing effects might also help
>>
>>590897
>not use the default lighting?

are you retarded or something? I'm talking about the basic way light is calculated by the engine. It's not anything to do with the actual light settings in the editor. As far as i can tell, modifying the source code is the only way to eliminate it. There's nothing you can do in the editor to get unreal's basic lighting to look like unity's.

>and set up materials to use things like PBR.

okay so you are retarded then, got it. It's almost impossible to NOT use PBR in UE4
>>
>>590892

It's unreal engine 4's trademark look. Almost all unreal engine 4 games have this plastic look, I hate that too.
>>
>>590901
yep. Just wondering if there were any coders on here that could help me figure out how to fix it. Glad someone else sees it and it's not just me going crazy. The worst part to me is that i'm trying to figure out exactly what weird math is causing it. It seems like the gradient from light to dark is not quite as harsh as it needs to be.
>>
Anyone have any experience attaching morph targets to a widget slider?
>>
>>590892
Yes, brilliant. Absolutely great shitpost.
https://www.youtube.com/watch?v=h5AR3Wf1QJg
https://www.youtube.com/watch?v=bXouFfqSfxg
>>
>>590893

Well with that kind of scene of course everything is going to look like plastic

I'd also look at any post processing options that might be active.
>>
>>590921
Are you seriously accusing ME of shitposting while you post scenes using vxgi? No shit the plastic look will be mitigated if you implement ultra expensive, ultra-advanced real-time GI that is too taxing to reasonably use in a game.
>>
>>590921
looks horribly noisy anon
>>
File: HL2-UE3-vs-UE4.jpg (510 KB, 1195x1372)
510 KB
510 KB JPG
>>590935
>>590935
>Well with that kind of scene of course everything is going to look like plastic

it shouldn't, and in any other engine it wouldn't. That's my point. Here's a comparison of the same scene in UDK and UE4 to illustrate what I'm talking about better. Sure the UE4 scene has cleaner lighting but also notice how the ue4 seems to have a weird gradation in the lighting that causes hard surfaces to look somewhat off compared to the UDK image. It's especially noticable on the trash can and the bench.
>>
>>590945
the only diff in lightning i can see on the trashcan and the bench are the ao is greatly reduced, making it look unironically "unreal"
>>
>>590945
Fuck having shadows am I right?
>>
File: RRbAMGP.jpg (115 KB, 1918x1041)
115 KB
115 KB JPG
>>590950
you're missing the point. The point is that the shadowing looks wrong pretty much invariably, making hard surfaces not look as flat as they should. here's another UE4 image that clearly shows what I'm talking about.
>>
>>590952
What's the fix around if any?
>>
>>590952

Don't waste your time with unreal fanboys. Unreal looks like plastic, and it's damn impossible to get rid of that look. I wish Lumberyard was easier to use and had actual tutorial videos for script canvas, it looks miles better than Unreal.
>>
>>590955
that's what I'm here trying to figure out. From what I can tell the only fix is messing around with the lighting source code. I can pretty much invariably tell when something was made in UE4 because of this way the lighting works and it honestly bugs the hell out of me.
>>
>>590958

Yep,Unreal engine 4 games are so easy to spot because of this. It's a shame because it's an amazing engine to work with, but the shiny plastic graphics are annoying. The only workaround I have found so far is to add a bit of noise effect to make the picture look a bit rougher.
>>
>>590956
>Don't waste your time with unreal fanboys.

That's the issue. I honestly love what ue4 does in a lot of ways, it tends to be more of an engine for artists while unity is more of an engine for coders. but Epic either seem to not understand that this issue even exists, or they're too busy making tools for developing games in VR that nobody will ever use. The new Skylighting is an improvement, but it's still not quite there. And it's only Unreal that looks like this. Unity doesn't have this issue. Lumberyard doesn't have this issue. Even UDK didn't have this issue.

I know I'd probably have to fuck with the source code, I was just wondering if anybody could actually help me know what values to change in the code to fix it.
>>
>>590961

Nah I suck at coding, I'm a blueprintfag. I thought about asking Unreal devs about this on the forums but I know that I'll be murdered by unreal fanboys.
>>
>>590962
>wasting devs time when you admitadely are a blueprintfag
>>
>>590965

What's wrong with being a blueprintfag? How does that make my point any less valid?
>>
>>590965
Lol, why would he have to fuck with the source code to get decent results? What, now you have to be a senior programmer to have the right to work with UE4?
>>
>>590968
i remember bacc in the day when people would say that Unity was for people who made 1 type of cookie cutter mobile game, and unreal was the real shit. Now unreal has been exposed for what it is - a bunch of idea guys putting together visual "nodes" with spaghetti wires and knowing nothing about programming and not having a degree in STEM
>>
>>590977
I hope you're joking, but - are you aware on what board you currently are? This is not /g/ or /agdg/, there is no "ideaguying" going on, we're creating visual art here (and maybe some basic mechanics) and that's it.
>>
>>590942
>you continues to shitpost
LOL. The state of your life!
https://www.youtube.com/watch?v=epCvJsyQMbw
https://www.youtube.com/watch?v=yxdp0icxcC0
>>
>>590892
Just fuck off back to Unity retard. No one cares about your dumb shitposts and the majority of game developers who are utilizing free engines for graphics are using UE4 because it's superior. The fact that you can't make your environments look non plastic means you're just shit tier in lighting.
>>
>>590977

I'd reather be a idea guy with an artistic vision than one of these unity asset flippin' coders who spend their time making bad 30 minute long horror games because they have no artistic vision whatsoever. Some people hate programming, and Blueprints are a great way for us artists to make games. By the way, you're on a 3D art imageboard, not a coding board.
>>
File: fortnite_br.jpg (270 KB, 1920x1080)
270 KB
270 KB JPG
>>590981
okay mr "you're a shitposter" If I'm wrong then how is it than I can invariably tell when a game was made in ue4 just by how weird and plasticy the lighting looks? Also the first example you posted shows what I'm talking about. The mushrooms look like plastic toys. Hell, not even epic can seem to make a game that doesn't look like shitty plastic. Fortnite looks this way. Paragon looks this way. Even gears 4 (which wasn't actually epic but still) looks this way. Are you saying that all of these professional devs don't know how to use the engine properly, not even the devs who actually made the engine?
>>
>>590982

But he's right, Unreal looks like plastic. That's the reason why Unreal Engine games are so easy to recognize (and the reason they all look the same). Unreal is superior to unity (which engine isn't anyway), but it can only dream of looking nearly as good as Frostbite and Cryengine.
>>
>>590982
> The fact that you can't make your environments look non plastic means you're just shit tier in lighting.

Then I guess epic just can't use their own engine right because their UE4 games look the same way. Also stop trying to claim that I'm shitposting when I'm just looking for an honest solution to a real issue. I think UE4 is a great engine, but it just has this one thing about it that annoys me to death.

Don't be an asshole.
>>
>>590985
You tell me with your concern trolling. It's literally nothing yet you continue to shitpost as if you have any validation in what you write. Give me the direct link to the Unreal developers commenting on the UE4 looking like plastic, and I'll at least be obliged to agree that UE4 does take a bit more to set up so you don't have something mediocre. But other than that, you're in reality just concern trolling and have no skill whatsoever to setup a good lighting and it shows with your blatant concern trolling stupidity
https://www.youtube.com/watch?v=FzoY062kY1s&t=117s
https://www.youtube.com/watch?v=ATX7kmET4zE&t=5s
>>
>>590989

Here's the same scene made in Cryengine.

https://www.youtube.com/watch?v=U60m7RRuuFU


UE4 version looks fake as hell in comparison.
>>
>>590985
>he lists names of games instead of providing evidence to his claims
Provide proof, because from what we're seeing here this is just concern trolling like >>590989
says.
>>
>>590990
>same scene
>different times of day
>different moods
Yes brilliant. Absolutely brilliant shitposting my friend.
>>
>>590992
This.
>>
>>590986
even unity can look really nice when it's being used by devs who know what they're doing.

https://www.youtube.com/watch?v=yD758VaYeY0
https://www.youtube.com/watch?v=R8NeB10INDo
https://www.youtube.com/watch?v=5ePB_bQT-iI
https://www.youtube.com/watch?v=AT29OGk_Byc
https://www.youtube.com/watch?v=cKLPe2RFPDw

UE4 tends to look better by virtue of a few things mainly

A. Lightmass
B. really nice models for skin shading right out of the gate
C. general better performance and ability to handle larger scenes
D. comes with a node-based shader editor.
E. Way better terrain tools.
>>
>>590994
>he doesn't post the developers comment about UE4 looking like plastic when asked for it even though he said they did
Give the comment or stop posting. You say they said it, so prove it.
>>
>>590989
>Give me the direct link to the Unreal developers commenting on the UE4 looking like plastic

If the engine devs don't acknowledge an issue with their engine it doesn't exist? Literally all you have to do to see what I'm talking about is play Fortnite. Or Paragon. Or the new Unreal tournament. Compare the new UT to doom 2016 and tell me with a straight face that the new UT doesn't look more plasticy. Fortnite is probably the worst offender I've seen and it's literally developed by Epic themselves.
>>
>>590996
gears 4 looks like a cartoon. Not realistic at all, but in a bad way. I dont play the game tho, because im not an ideaguy
>>
>>590995
>even though he said they did
>You say they said it, so prove it.

Literally what the fuck are you talking about? Are you high? I never said Epic ever acknowledged the issue.
>>
>>590992

You unreal fanboys are even worse than blenderfags. It's not just about the mood. It's about the shaders, GI, AO, Cryengine always looked amazing for outdoors. Even the old Cryengine 3 looks better than Unreal Engine 4.
>>
>>590996
>>590998
LMAO YOU ACTUALLY DELETED YOUR POST WITH THE QUOTE. Holy fuck this is rich. Yeah I'm out, I'll let the shitposter continue with his concerned trolling. Too much fun though.
>>
>>591000
delete yourself from life
>>
>>591000

I never deleted any posts dude, you're just delusional.

>Yeah I'm out, I'll let the shitposter continue with his concerned trolling. Too much fun though.

All i did was post a thread with a legitimate issue that I wanted to figure out how to fix. You're the one who's flipping out and calling everyone a shitposter.
>>
>>590996
This is a strawman just so you know. I saw you post saying they stated that it was an issue and they were looking into it, but now that post isn't anywhere to be found in the thread. He wants you to post proof so you can prove you're not just talking out of your behind. If you just make up hearsay and falsities it doesn't bode well for your arguments. Why did you delete your post? Not saying you're wrong or right but this is looking like you're just trying to not lose any ground you potentially made in the thread. Deleting your post significantly makes it look like you screwed up.
>>
>>591000

Unrealfags are definitely more retarded than Blenderfags.
>>
>state of the art PBR CG with Unreal Engine 4 running on Xbox One X at 4k 60
...t-they fucked up the skin too, didnt they lads?
>>
>>591003
> I saw you post saying they stated that it was an issue and they were looking into it

That never happened. I stated literally the opposite.
>>590961

>Deleting your post significantly makes it look like you screwed up.

I never deleted anything. You're just nuts.
>>
>>591006
Multiple people saw the post, and you're now denying you posted it. It really doesn't make me too care much for your arguments if it's backed up by falsities.
>>
>>591007
>and you're now denying you posted it.

Because I didn't post anything that said that? I literally just linked to a post I made earlier that stated the exact opposite.
>>590961
>>
Will we ever see another big engine become free for everyone, like Frostbite or such?
>>
>>591008
It may not have been you specifically who posted that response saying the developer knew that their lighting cause a type of plastic aesthetic to everything. However from what it appears there's only one or two anons agreeing with each other about this. Seems to reason one of you deleted your post containing the comment about the quote. That post was legitimately there and now it's gone, so it certainly reminds me of shitposting.
>>
>>591011
>It may not have been you specifically who posted that response saying the developer knew that their lighting cause a type of plastic aesthetic to everything.

because it wasn't.

>However from what it appears there's only one or two anons agreeing with each other about this.

or it could be one anon pretending to be multiple people.

>That post was legitimately there and now it's gone, so it certainly reminds me of shitposting.

Why would a shitposter delete their shitposts? Doesn't that kind of defeat the point? Either way if anybody did, I never saw it (and I've been in this thread all day) And it wasn't me: AKA OP
>>
>>591009

Unlikely so. That would mean that these game companies would have to provide support for the engines on forums. I can't imagine Electronic Assassins (or any game companies who aren't into engine business) doing that.
>>
>>591012
Considering if he posted that the developer stated it and they actually didn't, it would be detrimental to his shitposting. If he gets called out for hard facts that he himself admitted but can't submit, then this would bring light upon his shitposting. He can continue the long game by denying he never stated this. This is common sense.
>>
>>591012
Also to add on to >>591015, one or two anons are agreeing that UE4 makes the lighting look plastic, which I would say yes, there could be some samefagging going on in here. But we know that's not what you're responding about.
>>
>>591012
This just means you're the guy who was shitposting.
>>
Does unreal really not allow you to code your own shaders without modifying the source code? Does that mean ggxrd and borderlands were put together using nodes?
Jesus. I can understand offering a node alternative, but to completely disallow the ability to code shaders seems kinda shitty.
>>
>>591015
I am the only "he" who denied making a post that said that because I didnt. maybe someone else made a post like that but I didn't see it. And you constantly crying "shitposter" is more shitposting than someone asking for a solution to a legitimate issue.
>>
>>591023
There is only one answer to a shitposter continuously shitposting over a nonexistent "problem"; to ignore them or otherwise it gets out of hand as such in this thread. Reading through your posts I see a few arguments you make solved by just replacing certain elements of each video you critiqued with another sources materials/assets. For example the shrooms were too plastic in one video and yet another video showed their assets to be quite illuminating and organic within theirs. Irregardless of whether or not you are "concern trolling" it does seems quite fishy (again) to see somebody delete a post which would make their arguments seem weak and poorly managed. I'd suggest pointing out the errors within different UE4 games which specifically pinpoint plastic lighting in non relevant situations and explain why they seem to be this way. All the argumentation you've made is mostly just hearsay/subjective viewpoints and has very little in providing solid evidence towards UE4 being consistently plastic. Not to mention you've made the argument that Unity has better lighting when Unity is one of the most disparaging in its cases of having "sameface" (which is also contributed towards in partial by Unitys lighting). With this said I'll have to stop replying to you and just ignore your responses because this thread has gotten extremely off topic. For now, just consider that I see you as a shitposter and agree with the anons above accusing you of it. Doesn't mean you are, but I'm chalking this up to aforementioned "concern trolling".
>>
>>591025
Maybe if you actually would contribute to the actual discussion rather than throwing around the label of "shitposter" left and right, the thread wouldn't have gotten off topic
>>
>>591025
All you have to do to see how plastic ue4 is is just open fortnite.
>>
>>591025
My argument is not that unity has better lighting overall, but that it's base model looks less plastic. But it's not just unity, it's basically every engine outside of ue4.
>>
>>591020
They were able to implement some custom code due to the shading not interacting with the lighting. Same with dbfighterz. You can have custom shaders, you just can't define custom shading models that interact with the lights without modifying the source code.
>>
>>591027
>>591028
>>591029
holy shit, this is the most blatant samefag I've ever seen. can the samefagging stop please.
>>
>>591031
Samefagging is when someone is pretending to be multiple people. Never in any of those posts did I pretend to be multiple people. Please stop shitting up the thread. Thank you.
>>
>>591033
>three posts all within repeating minutes of each other
>same amount of posters
>on /3/
topjeje.jpeg heres ur complimentary reply.
>>
>>591034
Again please stop shutting up the thread. I never pretended to be multiple people so what are actually accusing me of? Making several posts in rapid succession? Wow sorry I guess I'll stop.
>>
>>591035
Shitting*
>>
>>591035
>he replies with multiple posts to one person in order to seem like multiple people are responding to him
>he admits this
>proceeds to cry about shitting up the thread
>this is the power of autism on /3/
really makes the nogging go a joggin lads
>>
>>591037
>in order to make it look like multiple people responding

No, because I thought of something else to say. If I was Samefagging why would I admit they were all me? How about disussing the topic rather than being autistic about how many posts I made in a row.
>>
>>591038
>so weak he admits he tried samefagging
maybe if you weren't such a adhd loser you'd be able to just respond once.
>>
>>591039
I just admitted to making multiple posts in a row without trying to appear like multiple people. Some of us actually want to discuss unreal engine rather than just shitfling. You seem really angry.
>>
>>591031

>>591031

Listen you little autistic unreal fanboy. There is no samefagging going on here. If you cannot accept the fact that more than one person agree that unreal looks like plastic, then get the fuck back to to the unreal engine forums, there you'll be surrounded by your fellow fanboys who think unreal looks the best. In the meantime, we'll enjoy freely talking about how overrated UE4's graphics are. Faggot.

Here's another discussion about that matter

https://www.gamefaqs.com/boards/691087-playstation-4/65840596

"UE4 looks like plastic toys."

See, we are not the only ones agreeing on that.
>>
File: 1510099599741.png (703 KB, 757x580)
703 KB
703 KB PNG
>>591040
>ywn be so autistic you think the person you're replying to is angrily replying to you with memes
wew lad. take some Ritalin alright junior?
>>
>>591041
didn't read lol. also
>he cites a gaming board full of gamers (TM) for support
>>
>>591041
ur an idiot
>>
>>591042
Can you do anything but shitfling? How about, I dunno, discussing the topic?
>>
>>591045
>ywn be this autistic
can you do anything except continue to be autistic and post as if your pooper isn't peeved? at this point you shitpost as much as me.
>>
>>591046
I've done nothing in this thread but try to have an honest discussion and it's just been derailed by constant accusations of shitposting.
>>
>>591041
I personally liked that "plastic", or more like a "clay" look, especially in games like Gears of War and Batman. I always knew it was not perfectly realistic, but I found it pleasant. Maybe I'm the only one.

But now that I work with UE4, I don't really notice it - maybe I got used to it, or maybe I know how to get a good lighting and PP so I manage to get rid of it, but anyway, I don't see the problem. Probably just don't know any better.
>>
File: Capture.jpg (60 KB, 1199x740)
60 KB
60 KB JPG
>>591050
Here's a still from a major UE4 game. Horrible, but "state of the art" PBR
>>
>>591050
Heck, I notice more in ue4 than in ue3. If you like it then that's cool, I just wish there was a Ray to get rid of it since it sort of grates on me.
>>
>>591043
>>591044


lul didn read ur an idiot

Retarded Unreal fanbase in a nutshell
>>
>>591054
stop. Lets just discuss UE4. no flaming.
>>
>hes still samefagging this hard
>even after everyone had left
lol
>>
File: 20160521201116_1.jpg (326 KB, 1920x1080)
326 KB
326 KB JPG
>>590996
>Compare the new UT to doom 2016

Are you really trying to say this doesn't look like plastic?
>>
File: 20160711211214_1.jpg (456 KB, 1920x1080)
456 KB
456 KB JPG
>>591103
Because Doom 2016 REALLY owns the plastic look, here's rocks in Doom.
>>
File: 148923971726099274.jpg (314 KB, 1000x562)
314 KB
314 KB JPG
>>591104
Here's rocks in Unreal.
You want to tell me with a straight face that Doom looks *less* like plastic?
>>
If you can control all the parameters like specularity, roughness and Fresnel, why is it problematic to create a non-plastic models? What am I missing here, what else defines a material?
>>
>>591107
I don't understand it myself, I if you understand how the PBR lighting model works you should be able to tweak the textures and constants to get the material to look exactly the way you want. Of course, the key thing is "understand". You can't just simply plug in the maps and expect it to look the way you want it to.
Also, it's totally possible to write your own shaders for UE4 https://docs.unrealengine.com/latest/INT/Programming/Rendering/ShaderDevelopment/index.html

Does nobody read docs anymore?
>>
>>591107
There's plenty of textures in existing Unreal games with parameters that make them not look like plastic, and this dude's relying 100% on posting >>590945 images >>590893 of scenes >>590952 with a single plastic-like default material on them like oh fuck I wonder why this default soft grey pbr material looks like plastic.
>>
>>591109
Because the entire thread was turned into a "UE4 PLAESTIC" shitposting and samefagging poop zone to throw off the entire discussion of topic. Though it's not like /3/ is productive anyway so who cares.
>>
Does anyone have a guideline on how many vertices/tris UE4 can handle on a high-end GPU like a 1080, and still maintain about 60fps? My modeling rig is a potato or else I'd test it myself.
I've started doing some content for a sci-fi themed project that'll mainly be geometry-heavy, so I'm employing medium-poly modeling (all edges chamfered, one smoothing group), but I can't help and go back and optimize the crap out of my meshes time to time, wondering if I'm cutting back on too many polys and sacrificing the smoothness of my surfaces.
The current modular sections I have for the levels weigh in at anywhere from 1500 verts for a section of bridge to 14k on the largest single piece, and a massive (as I see it) 24k verts on a really complex assembly that'll probably bloat to 55k once I propagate the individual pieces into the complete shape, although it's a fairly prominent structure within the level. The individual chunks are all below 4k verts each, it's just that there's a lot of them.
>>
>>591138
Archviz guy here
We did an entire shopping mall, interior and exterior and a bunch of the surrounding buildings, cars, trees, people, you name it. Scene had over 100 million tris. We set up a little bit of level streaming but not much, and we didn't bake any lighting.
Still got 30 FPS on a 1080ti, 4.2GHz i7 and 32gb ram.

What brought our FPS down was not polycount tho, but shadercount. We bought some storefronts from Evermotion to put inside the mall, and those were highpoly models for offline rendering. We reduced the meshes and remade the shaders in UE4 so it was all instances of one master shader, but it was still up to 30 shaders per storefront, and we had like 90+ of them. Without those, we were sitting at upwards of 60 FPS, and the scene was still not optimized in any significant way.

I don't think there's really any limit to polycount in UE4. Other things hit the performance a lot harder, like shaders and texture size.
>>
>>591145
Holy shit, well that changes things... not only can I go and make all my rounded surfaces fully round now, but also model out all the detail I was saving for normal decals. Due to the large scale of the things I'm making, unique textures and normal maps isn't a thing anyway, just tiled materials and decals.
>>
>>591153
Don't throw every care out the window tho. In Archviz the camera is mostly stationary or moving really slow, if we had zoomed around like in a shooter, we would have had a little less FPS. But mostly due to shader stuff I think.

And of course we set up LODs for almost everything using UE4s auto lod system.

But yea if you optimise a bit you can get away with a lot of geometry. Large scale arch projects come with tons of geometry anway simply due to their size. It's often cheaper to model a few more details rather than splitting everything up into modules and using tons of normal maps and shit. You can also use vertex painting for large stuff.
>>
>>591164
I see... The poly count won’t go up a lot, I was already using up to 8 segments per 90° as a guideline for round objects, but some geo is so complex that even that many edges is barely enough, face-weighted chamfers tend to make round edges look jagged if there’s not enough detail. I’m expecting a difference of maybe 50% extra tris, so it’s still not that bad. Camera-wise, it’ll be a third-person platformer a wide view of the surroundings, so it won’t move fast.
>>
>>591168
>third-person platformer
Sounds like you can use a lot of level streaming, LODs and even proxy meshes if need be
>>
File: compare1.jpg (1.13 MB, 1900x2152)
1.13 MB
1.13 MB JPG
>>591105
It's hard to describe and maybe "plasticy" is a bit too simplistic a description but it's extremely easy to tell when a game was made in UE4 simply by virtue of the way the lighting looks. It's something that seems to be essentially universal over every game made in UE4. If it was just in my imagination, then how can I invariably tell when a game was made in UE4 and not some other in-house engine?

>>591111
You're missing the point. I used those because since it's an issue with the lighting, obviously textureless scenes would best show what I'm talking about. I showed that train station scene to show how UDK doesn't have the same issue, it's a comparison image.

If you want textured scenes here's two games with pretty much the same artstyle. One made by epic themselves in UE4 and one made by a smaller dev in UE3.
>>
>>591109
> if you understand how the PBR lighting model works you should be able to tweak the textures and constants to get the material to look exactly the way you want.

PBR isn't a lighting model, it's a workflow. PBR just means that it uses real-world properties in setting up your materials (i.e. having metalness as an attribute whereas past shading models didn't). It's something totally separate from the illumination model (which is how your lights work) and the BRDF (which is your base model that defines how your materials react to the light).

So if there's an issue with the accuracy of the illumination model, messing around with the textures and constants of a material isn't going to change that.
>>
>>591183
Jesus. Ue4 looks like absolute SSHIT.
>>
>>591183
I wasn't really decided on the UE4 "plastic" look since I thought the latest Unreal Tournament was cool, though I ain't played it, but then I opened this JPEG.

I thought for a moment this was supposed to be a Toy Soldiers game and the house was far away. The giant pick axe didn't help.

What the fuck happened?
>>
>>591192
PBR happened
>>
>>591199
it has literally nothing to do with PBR. Again, PBR is just a rendering pipeline based on real-world properties like metallicity and roughness. it isn't a lighting model, nor is it a shading model. Pretty much every modern engine uses PBR.
>>
File: days_gone2.jpg (1.23 MB, 3840x2160)
1.23 MB
1.23 MB JPG
>>591199

What's wrong with PBR? The problem here is Unreal. Cryengine doesn't have these horrendous plastic/playdough aspect, same with foxengine. The problem with unreal is that it looks stylized.

I mean come on, look at this pic. You can immediately tell this game is made on unreal just by looking at how everything is rendered. It looks thick, oily, almost cartoony.
>>
>>591200
>PBR is just a rendering pipeline based on real-world properties like metallicity and roughness. it isn't a lighting model, nor is it a shading model
> it isn't a lighting model, nor is it a shading model
lmao, go study cg
>>
>>591203
>lmao, go study cg

PBR is literally neither of those things. One PBR pipeline might use a totally different lighting or shading model than another.
>>
>>591205
kill yourself. Its one big multistage shader and they all are basically the same
>>
>>591206
>Its one big multistage shader and they all are basically the same

No, it really isn't. And no, they really aren't. One engine's implementation might be totally different than another's. This is why when you import something from say, marmoset, to UE4 it will look different. Hell, even unity has switched their standard PBR shader from using normalized phong shading for specular reflections to GGX. But guess what? In both instances it was still PBR! Cryengine uses physically-based shading as well but for Ryse they used an oren-nayar model for calculating diffuse rather than the standard lambertian model they used to use. Both were still PBR.

stop telling others they don't know what they're talking about when you have zero clue. PBR implementations can and do vary in loads of ways.
>>
>>591206

What a clueless little jackass.
>>
>>591210
>No, it really isn't. And no, they really aren't.
yes, it really is.

>Hell, even unity has switched their standard PBR shader from using normalized phong shading for specular reflections to GGX
welcome to 2013.

> But guess what? In both instances it was still PBR!
whoever is feeding you this stuff needs to retire from the industry and work at walmart

>Cryengine uses physically-based shading as well but for Ryse they used an oren-nayar model for calculating diffuse rather than the standard lambertian model they used to use. Both were still PBR.
lambert isn't energy conserving and thus isnt PBR. Again, whoever tells you this is retarded.

PBR is oren-nayar / ggx direct lighting and HDR IBL indirect * ssao with bloom post processing and tonemapping on top. Period.

>>591211
kys
>>
>>591212
>PBR is oren-nayar / ggx direct lighting

Then I guess by your logic ue4 is actually NOT PBR then because UE4 uses lambertian diffuse shading by default. Just stop please. Unity also uses lambertian diffuse for its PBR rendering.

>lambert isn't energy conserving and thus isnt PBR.
http://www.rorydriscoll.com/2009/01/25/energy-conservation-in-games/
https://seblagarde.wordpress.com/2012/01/08/pi-or-not-to-pi-in-game-lighting-equation/

>with bloom post processing and tonemapping on top

lolwut. If you don't have bloom or tonemapping it's not PBR? You think that PBR has anything whatsoever to do with post-processing? Are you high? There is also nothing "physically-based" about bloom.
>>
>>591213
ue4 uses ggx. I've read papers on lightning by epic themselves. Just kys please.

> If you don't have bloom or tonemapping it's not PBR? You think that PBR has anything whatsoever to do with post-processing? Are you high? There is also nothing "physically-based" about bloom.
bloom simulates a HDR display on a non HDR display. If you dont understand that, there's no saving you.

Stop wasting my time.
>>
>>591214
>ue4 uses ggx

Yes, it does. I never said it didn't.

>bloom simulates a HDR display on a non HDR display.

That is not even remotely what bloom does.
>>
>>591216
...it "blooms" the pixels around a "hot pixel", simulating what you get in real life when you look at something similarly "hot" like the sun reflecting off a smooth reflective surface. This is not possible on a regular display and barely possible on HDR.

Are you just pretending to be retarded?
>>
>>591220
Bloom kind of does that but not very well and usually just ends up making things even uglier. Its also not at all part of the shading model. Also, just ssao? Not hbao or dfao or vxao? Ssao is the only kind of acceptable ambient occlusion?
>>
>>591225
>Bloom kind of does that but not very well and usually just ends up making things even uglier.
You're just using awful unreal and unity stock blooms. You dont even know what you're talking about and making a blanket statement.

>Also, just ssao?
our target is 4k 60 on console or higher for VR. SSAO only.
>>
I lost your train of thought guys, but I need to know where to learn all this technical stuff you're talking about. Please give me some directions.
>>
>>591228
And yet you still think bloom has anything to do with pbr shading in games. Bloom has nothing whatsoever to do with the shading model. Its purely postprocessing.

But ue4 does use lambertian diffuse shading by default. Does this mean ue4 does not use pbr?
>>
>>591232
bloom is an essential part of pbr and every paper, video, and breakdown I've read in recent years supports this.

I dont give a fuck about what ue4 uses by default
>>
>>591233
>bloom is an essential part of pbr and every paper, video, and breakdown I've read in recent years supports this.

bloom is totally independent of PBR. It's a post-processing effect. PBR is used to describe the general type of shading pipeline that uses physical properties like metalness and roughness. If bloom

>I dont give a fuck about what ue4 uses by default

So by your logic ue4 isn't PBR unless you actually go into the source code and change the diffuse BRDF?
>>
>>591237
>bloom is totally independent of PBR. It's a post-processing effect
its a key component of making a final render look based in the physical world that we humans look at every moment of every day. What amount of pixels it works on is irrelevant

>So by your logic ue4 isn't PBR unless you actually go into the source code and change the diffuse BRDF?
i've read lengthy papers and presentations by epic games on using ggx, and this was years ago. I dont use the engine, and i dont think much about it now.
>>
>>591239
>its a key component of making a final render look based in the physical world that we humans look at every moment of every day.

again. PBR describes the SHADING PIPELINE. It has nothing to do with post-processing effects.

>i've read lengthy papers and presentations by epic games on using ggx, and this was years ago.

Yes it does use GGX. It also uses Lambertian diffuse shading. Do you think they are mutually exclusive or something?
>>
>>591242
>PBR describes the SHADING PIPELINE.
you execute your shaders in a sequence every frame. This is your pipeline. You have no idea what you're talking about.

>Yes it does use GGX. It also uses Lambertian diffuse shading. Do you think they are mutually exclusive or something?
They are each direct diffuse shading models so yes they are mutually exclusive, like gold and platinum. Are you this dumb?
>>
>>591201
looks exactly like the last of us, the nondyke version
>>
>>591243
>You have no idea what you're talking about.

says the guy who thinks post-processing effects have anything do with whether or not it's PBR. bloom has nothing to do with the actual shading model. It's a post-processing effect that generally isn't very realistic in the first place.

>They are each direct diffuse shading models so yes they are mutually exclusive, like gold and platinum. Are you this dumb?

UE4 by default uses lambertian diffuse shading and GGX for specular. Please learn what you're talking about before you accuse others of being dumb.
>>
>>591247
says the guy who doesnt actually know a pipeline is

>UE4 by default uses lambertian diffuse shading and GGX for specular.
my mistake, i was distracted, working on my work. GGX is direct spec, Oren-nayar is direct diffuse.
>>
>>591248
>says the guy who doesnt actually know a pipeline is

I do. And you can have PBR pipeline or a non-PBR pipeline. And whether one is PBR or not has zero to do with which post-processing effects it uses.

> i was distracted

Yeah, you were "distracted" that's why you kept posting that using GGX and lambertian diffuse were mutually exclusive.

> Oren-nayar is direct diffuse.

Yes and UE4 cannot use it unless you mess with the source code. So by your definition UE4 doesn't do PBR?
>>
>>591250
you STILL dont know what a pipeline is. You think a "post-processing effect" is not part of it. If this was a course at gnomon you would get an F
>>
>>591251
I admit i made a stupid slip-up there. Yes post process effects are a part of the shader pipeline, they're just not part of the actual material shading model. I knew that, I just used the wrong word. Either way bloom has nothing to do with PBR and you still refuse to answer whether or not UE4 uses PBR since it uses lambert for diffuse shading.
>>
>>591254
>Either way bloom has nothing to do with PBR
smmfh

--------

is lambert physically based?
>no
well then.

why do they use it?
>perf

I rest my case
>>
Please tell me where can I learn about this. Btw I see you mention Gnomon, I already asked this before, but if you are/were a Gnomon student, I'm interested where could I find some more information about the coursework and see how the assignments and exams look like.
>>
>>591257
buy and read books, papers, and presentations about CGI
>>
>>591256

So all these devs saying that they're using PBR without oren-nayar shading are just lying? Or maybe perhaps PBR doesn't refer to exactly what you think it does.

>performance
>I rest my case

was it not you who posted this?
>>591228

So you're just a hypocrite then. It's alright for you to not use the 100% best type of AO and have it still be PBR but for them to use a somewhat cheaper diffuse model suddenly makes it not PBR. What engine are you using for VR that's using oren-nayar for diffuse? Also by your logic that we have to 100% the most rendering techniques for it to be PBR then nothing will be PBR unless it uses VXGI or some equivalent.

In reality PBR just describes a shader workflow that uses real-world properties to try to achieve more realism (i.e. metalness, roughness, energy conservation, etc.). It doesn't mean they're using necessarily the most expensive or accurate BRDFs or illumination models.

https://de45xmedrsdbp.cloudfront.net/Resources/files/2013SiggraphPresentationsNotes-26915738.pdf
>>
>>591256
>smmfh

find me a single technical paper or presentation that says that bloom is a necessary part of PBR
>>
>>591263
find it yourself.
>>
>>591264
I thought so.
>>
>>591265
ok?
>>
>>591266
You you claim bloom has anything to do with whether an engine has PBR or not but you won't provide anything to back that up. You also refuse to admit that PBR doesn't necessarily mean you HAVE to be using the best of the best most expensive graphical techniques.
>>
>>591269
You're just wasting my time now.
>>
>>591270
okay. But you're the one who claimed shit without being able to back it up.
>>
>>591271
>non sequitor
>>
>>591275
>"bloom is necessary for PBR"
>evidence?
>"lol look it up yourself"
>okay you can't back up your assertions then i guess
>"lol non-sequitor"
>>
>>591277
0/10
>>
>>591259
I asked "where", no shit I can learn from books and presentations, come on. Why is it so hard to recommend some useful sources you use, I wouldn't even ask otherwise.
>>
How to properly edit my normals on foliage so it receives lighting correctly? I did it by following some tutorial months ago, but I forgot how to do it and now can't find anything on YT. People don't seem to edit their normals at all, but half of my grass is almost black.
>>
>>591336

You need to use a two-sided shader. UE4's material editor has a built-in two-sided shading model specifically for foliage. Just select it in the shading model dropdown menu in the material editor.
>>
>>591336
You can use normal thief if you're using max. Not sure about other software
>>
>>591336
Basically you create a half-sphere over your mesh, scaled up until it covers your tree, and then find a way to transfer it’s normals over. In Max you need Normal Thief script, in Maya you can transfer attributes and leave just normals ticked.
Also you may want to make sure your polys are two-sided first, because the normals will get reset if/when the engine generates new faces on the backside, you can do this by shelling your billboards by a value of 0 and welding the verts.
Finally, make sure when exporting the mesh, ensure that all options pertaining to normals or smoothing groups are checked to preserve the normals.
>>
I think I have a bit of a problem with baked lighting.

I'm trying to recreate a lighting setup I had 2 days ago. I made a mistake of not saving that version and adding some tweaks over it. But now I can never seem to nail it again, because I can't remember if I baked my skylight in the meantime.

You see, every time I bake my stationary skylight, it affects the numbers and the same tweaks start giving different results, based on what kind of lighting I baked on a scene. Even if I put it back to movable, some of that information still stays and it messes everything up.

E.g. if I changed between a cubemap and captured scene in my skylight options, the difference between them was once so huge that I had to reduce the intensity of captured scene to 0.3 from 1 in order to not completely blind myself. But after baking, the difference can be barely noticed, and now e.g. if I set the intensity of captured scene skylight to 5, when once it was extreme, now it's even underexposed still.

So, I don't know if it is somehow possible to remove any bake I made, so it looks just like when I imported the lights into the scene. Basically, I don't know how to deal with this, because I think I understand how a lot of parameters work, but I can't tell what will happen after baking something, because everything constantly changes and gives different results.
>>
File: yd9KWxf.png (3.48 MB, 1920x1080)
3.48 MB
3.48 MB PNG
>>591105
And here's rocks in modified source.
>>
>>591612
Why do rocks in games generally look like shit? Is photogrammetry the final answer?
>>
>>591628
Artists tend make them look wetter than they should and fresnel in video games is generally wrong to begin with.
>>
>>591573
Nvm, fixed it. Dynamic GI somehow got disabled.
>>
>>591644
>Dynamic GI
explaim
>>
>>591652
You can enable GI for movable lighting through config files. It's hidden by default. Check out "tutorial UE4 dynamic GI" by PolyPixel on YT. Should be the first result.
>>
>>591655
oh sweet, thanks
>>
>>591628
-People make them too glossy
-Texture resolution rarely spared for rocks
-Filtering reduces sharpness, which has an effect on a few kinds of rock formations where soft edges in the texture really aren't suitable
-Compounding effect when realtime polycount rounds off edges without adding the extra polys necessary to make the transitions look natural, leaves the rock looking like a rock texture that has been shrink-wrapped over something

All this stuff is visible in photogrammetry.
>>
File: 2016-01-20_00003.jpg (908 KB, 2715x1527)
908 KB
908 KB JPG
>>591811
>>591612
>>591628
I think these are acceptable
>>
File: 2016-01-20_00004.jpg (880 KB, 2715x1527)
880 KB
880 KB JPG
>>591828
>>591811
>>591628
>>591612
>>
>>591828
>>591829
photogrammetry only works for existing environments and is thus incredibly limited. A dead end pretty much.
>>
>>591828
>>591829
Even this still has the obvious unnatural shine and oversoftened edges to it. Photogrammetry, most of the time, looks kind of convincing until you stick your nose into it, then current tech absolutely looks like a cloth has been draped over it.
Honestly prefer "stylized realism" action-figure looking stuff to that. It has the "wow, this looks visually appealing and HD" effect without my brain stopping to go "Wait, this looks seriously wrong and weird" the moment I get close to anything. I can deal with CG looking like CG.
>>
Is there any way of creating some nice skies and clouds without having to buy/pirate TrueSky plugin? I don't think I even saw the torrent. Just placing an hdri image won't be enough. I'd gladly learn how to make my custom skies in some program, but I didn't find anything. People either use TrueSky or shitty UE4's sky sphere.
>>
>>591830
You can turn photogrammetry into tilable textures with displacement mapping included. You can do loads with it.
>>
>>592432
There's a pretty killer library of scanned tiling textures called Real Displacement Textures, I have the first three volumes and they weigh in at about 36GB, for about 80 surfaces with 7 map types.
>>
>>592432
>>592439
Go away. Nobody cares.
>>
>>592439
Can I use them professionally for free? I've been trying to make my own textures with photoscan but each one takes a long time to do
>>
>>592432
Oh found it. These look pretty great though a bit pricey. A big studio could make good use of this
>>
>>592413
Take look at how the unreal skysphere is constructed. It has has multiple UV channels to display the various textures of clouds, stars and the horizon gradient. also look at their material setup, you can find it by looking in the engine content.
>>
>>592462
Go away. Nobody cares.

>>592473
Not sure about that... you can always blend them with other textures, and people might also think it’s Megascans, but I don’t think there’s anyone so autistic that they would look at a material, know what library it’s from, and then investigate whether it’s under license or not.

>>592474
The first three volumes are on CGP if you want to try them out. Quixel also has scanned materials and even a dedicated program to go with them, but it’s a subscription-based service, so cheaper but limited downloads. It also means there’s no single torrent file available, as it would require a lot of people to compile the stuff they can get over several months.
>>
In some of my foliage (be it grass or trees), if some parts are very thin (usually branches on the top etc.) they look like they are just a bunch of unconnected dots floating in the air, and it creates that blurry effect when you look at them from some distance. What can I do aside from increasing their width?
>>
Can you guys give some quick rundown on what causes biggest fps drops in UE4?

>4k+ resolution textures?
>number of polys?
>amount of transparency in foliage cards?
>number of assets?
>size of a landscape?

I don't know about the rest, but I see that my foliage literally decreases my fps by about 2/3. I did try to reduce the amount of transparent area on my cards, but still a lot of it remains. Should I just add even more cuts and reduce it, or should I do something else? (No, I won't reduce the density).

Also, I heard that the resolution of textures actually doesn't affect the fps, it just increases loading times - is that true?
>>
>>592855
It’s because UE4 implements a variant of TXAA by default. Temporal AA smooths edges by comparing the last and next frames, so it’s mainly of benefit for a still or slow-moving camera while employing little to no oversampling, but in fast motion where the AA becomes useless, motion blur is typically utilized to hide the lack of definition.
Since TXAA doesn’t oversample, stuff like hair, fur and foliage will shimmer, and sub-pixel details even get “eaten” by the AA algorithm, which is why there’s a switch for excluding particles from the AA solution.
The only real fix is using a different algorithm like FXAA, or disabling AA entirely and super-sampling the image from a higher resolution like 4K.
>>
>>592863
From what I understand, Unreal doesn’t give a shit about geometry, it’s mostly down to texture size and shader complexity. Anything that can be instanced in some way or share part of its design across multiple assets is the best way to optimize, but the actual number of assets doesn’t matter. Chances are that there’s something that isn’t very well optimized in your foliage.
Also I’ve heard that any elements dynamically constructed from BPs, such as those from splines, need to be collapsed into a static mesh for best performance. I don’t know enough about UE to say whether this applies to foliage in some capacity as well.
>>
>>591051
that's unreal 3... tho heavily modified from nrs
>>
Does anybody know how to install VXGI? I'm surprised by how little it is talked about on the forums, and I can't find almost anything on YT. Is it even available to everyone?
>>
>>593440
There are forks on GitHub.

https://github.com/MaximeDup/UnrealEngine

You have to be signed in to see it I think.
>>
>>592863
dynamic shadowing
>>
If I'm using dynamic GI and DFAO with my movable lighting setup, and I switch my directional light to stationary in order to bake my scene - do I have to manually turn dynamic GI and DFAO off, or will they automatically get disabled once I make my light stationary?

I'm asking this because I can literally see no difference between baked stationary directional light (and movable skylight) and fully movable lighting, and I don't know why, but that doesn't seem right. Is my baking even working?
>>
>>593828
Have you enabled depth fields in your project settings?

Also, DFAO has nothing to do with directional light, it will only get enabled when your skylight is set to movable, and will automatically get disabled when set to stationary.
>>
File: 1510341792212.jpg (104 KB, 750x864)
104 KB
104 KB JPG
I'm trying to take screenshots (using the HighResShot command) and no matter what I do, the screenshots are dark as fuck. I tried taking a screenshot with the print screen button on my keyboard and pasting it onto photoshop and export it from there, but the problem persists. I can't take any good screenshots because of that.

Any way to fix this?
>>
File: wtf.png (17 KB, 1170x546)
17 KB
17 KB PNG
Can anyone tell me what i'm doing wrong?


When i import the texture i get pic on the left

When i save the texture in Unreal, it changes to pic on the right

how do i fix it?
>>
>>593854
>depth fields
What is the full name of that? I don't know what you're talking about and I can't find anything.

Yeah, I know that about skylight, I also put it on stationary, but the difference after baking is just non-existent.

Actually - what do you guys recommend for lighting setup for exteriors? Fully dynamic or baked (static, stationary?) lighting? I'm having a hard time with tons of different parameters that change everything, and it's complicated to balance them.

>directional light intensity, color
>dynamic GI intensity (if I increase or decrease directional light intenity, this will also be affected)
>skylight intensity, color
>cubemap
>sky sphere cloud opacity (e.g. I want softer shadows, but if I add more clouds, lighting intensity will go through the roof)
>sky sphere color
>exponential height fog density, falloff, color, distance
>dfao intensity
>shadows, mids, highs in PP volume

So many things affect the lighting that I don't even know what to do anymore. I can play around with them for literal weeks.
>>
>>593949
And I have to mention that I have barely seen any good exterior scenes in UE4. Almost everything looks like shit, even project with hundreds of likes on ArtStation The only guys that got some decent results IMO were Koola, Eoin O'Broin and Art by Rens. Those first two seem to use dynamic setup, but I still have no clue how they get so good results, while Rens is a literal magician and my mortal mind can't even comprehend what is happening in his scenes. Maybe texturing is even more important, but I still feel that my lighting is never good enough and can be improved.
>>
>>593911
Unreal's texture compression is S-H-I-T. This is something everybody who works with UE4 has to deal with, there's no real "fix" besides changing the texture group to "UI texture" or something like that, itll turn off compression but you'll get huge files/memory usage.
>>
>>593949
>depth fields
I believe the option is "generate mesh distance fields"
>>
>>593953
Oh, yeah, I have that turned on.
>>
Help a brother out please!

I have created a skeletal mesh with a basic walking animation and I want to create an infinite group of characters spawning, crossing the road on a single line and despawning).

>How do I spawn the charaters lets say every 5 seconds?
>how do I put a simple straight path on them?
>how do I make them despawn after reaching he end of the road?

I will donate 5$ to a charity of your choice if you manage to help me.
>>
>>593952
Fuck. so i guess just change to UI Texture for really noticeable objects?
>>
>>593950
>Maybe texturing is even more important, but I still feel that my lighting is never good enough and can be improved.
What people never do enough of is grayboxing, also known as a clay preview. The theory is as follows: If, in reality, you had a perfectly white room, with all white objects in it, and even the table with the vase and the flower in it were all white, or at least some flat neutral shade, how would the scene look given real, natural lighting?
People often get distracted by textures and other effects, never paying attention to how the lighting affects the scene on its own. Try setting up a scene, and with just a basic off-white shader for everything, try to make it look as good as it possibly can as-is. Begin to add in things like AO, and other effects which affect only the tonality of the image, and eventually everything else you add will look good as well, possibly requiring only minor tweaks along the way.
>>
>>594017
This is great advice.
I would say that a good texture can save a bad model. Probably in that same line of thinking, good lighting can save a bad scene.
>>
File: witch_20.jpg (900 KB, 1600x1597)
900 KB
900 KB JPG
>>593950
I'm currently baking a static skylight and keep the sun as a movable directional light, with that setup I can manage a dynamic time of day by adjusting the indirect lighting color in the post process settings, baking a static skylight works only if your levels aren't too big though and 4.18 added volumetric light maps which also help a lot.

>>594017
You can just switch to detail lighting view mode (alt+5) in the viewport to get your scene without colors.
>>
>>593955
go ask /agdg/ or google, these sound like pretty basic shit you can look up
>>
>>594029
When's the alpha fgt? We've been waiting forever.
>>
I have no experience with previous versions of UE, but I'm interested when could we expect UE5 and what needs to happen for them to give it a new number in the name?
>>
>>594243
UDK released in early 2000s, I believe 2003? Something like that. UE4 released in 2014. So 7-12 years from 2014.
>>
Can somebody give me a quick explanation on best practices to import Megascans assets into UE4?

I tried using megascans bridge but it imported the highpoly into UE4 and didnt create any shaders. Am I supposed to just import the mesh+ all LODs manually and set up all the shaders manually too?
>>
New Adam episode made in Unity.

Opinions?

https://www.youtube.com/watch?time_continue=46&v=tSDsi2ItktY
>>
>>594364
cool. Now imagine if unity actually came with all of the tools they used to make that out of the box.
>>
>>593950
>, while Rens is a literal magician and my mortal mind can't even comprehend what is happening in his scenes.

He mainly uses photoscanned objects plus VXGI.
>>
>>594523
well they did post the assets at least
https://unity3d.com/pages/adam
>>
>>594524
Is VXGI that much better than baked/LPV? How much worse is the performance compared to fully dynamic setup with dynamic GI?
>>
>>594660
>that much better than baked/LPV

Yes.

>How much worse is the performance compared to fully dynamic setup with dynamic GI

Wut? VXGI is a type of dynamic GI.
>>
>>594660
1080ti should be fine as your target for 55 fps.
>>
>>594732
Oh, I know, I just heard it's very expensive, and thought it might be even worse than LPV.

>>594752
I'm pretty sure I'll get around 20 fps on my rig, but I want to try it out.

But where can I download it? GitHub link gives me 404 even when I'm logged in. There's also this page, but it hasn't been updated in a while - is there a newer version somewhere or could I download this?

https://developer.nvidia.com/gameworksdownload#?search=vxgi
>>
>>594756
Check the UE4 forum thread for updated versions built by users.
>>
>>594759
Found a tutorial and followed the steps, but I'm getting way too many errors and wrong paths, don't want to invest so much time in this. It's impossible for me to get clean installations and builds for anything. Fuck it.
>>
>>594768
Well, that's the price of early-adoption and cutting edge tech, just imagine what these people had to go through to get it to work. A lot of the stuff that you're even using now was created by people with top degrees in their fields, and just using their shit practically requires you to have the same knowledge base, it's only through trickling down from more experienced users that most mortals can get it to work. Consider it a luxury that you can enable a whole wealth of effects just by checking a tick box in your editor, years of blood sweat and tears might have gone into that box.
>>
Is it normal to have to tweak materials when lighting changes? I don't know if I'm doing something wrong, but I instanced my materials and in my "day" scene, I tweaked their brightness, SSS amount and such until they looked good. The problem is when I decrease my light intensities, rotate my sun and do all the other things to make a night scene, a lot of materials don't look right anymore, and some stay very bright. But how would that work with dynamic environments, then? How to make sure everything will look good under all lighting setups?
>>
>>586984
There used to be a UE4 with Nvidia's VXGI - Lightmass was built to negate that, because its a Nvidia baised feature. To build on your answer, never ever. They won't replace lightmass until they absolutely need it when producers want to make a open world game.

But there is cryyengineeeee
>>
>>594879
>cryyengineeeee
but it is a hassle to use; even to import an fbx into the engine is a pain in the ass
>>
>>594889
Got two questions:
1. What's a good way in UE4 to fade geometry that's between the player and camera in a 3rd person game?
2. What's a good way to generate thick cloud cover?

To expand on point 1, from what I understand, you can't stack transparent geometry in a deferred engine, and since I'll already have water in many cases, fading via transparency is a no-go, right?
If I do end up creating a fade of some sort, will it be dependent on the geometry having to be broken up into discreet pieces that have capped faces on the inside?
I'm already building relatively modular components to construct the environment, but some of these modules are fairly large/tall and are currently in one piece from top to bottom, if it ends up that I can only hide individual chunks, I'll have to slice them vertically to avoid having to hide the whole thing, correct?

And to expand on point 2, I mean clouds that you could, for instance, see from above as if you were in an airplane, which would obstruct the view of the surface, so not just "background clouds" but actual physical clouds in the world.
>>
>>594889
definitely if you aren't using maya or whatever. how about lumberyard? I've never tried it.

and CE could be worth it if you want that look, but in exchange for many hours of pain.
>>
>>594893
LY is just a fork of CE, right? Based on what I've seen of them, I think they're both in a state of transition where they want to get away from the obtuse nature of the legacy build and have a solution that's more like UE4.
I think that, unless you're willing to eat the pain right now for the potential benefits, it's best to wait until their next major updates when more stuff becomes ironed out. As it is, you have to consider whether you can get a better result out of them given the time investment.
>>
>>594841
Bump. If I have to always tweak materials, I can't make timelapses.
>>
>>594897
That shouldn't happen Anon. Those materials which stay too bright break energy conservation.
The whole essence of PBR materials is that they work under all light conditions. If they don't you somehow broke the rules.
Check your node soup, you don't want to feed numbers above 1 into your albedo color. Use a normalizer node or something like that.
>>
>>594891
1- There is no 100% trick for 3rd person cameras, specially in closed interiors. What some games do is have the camera move closer to the character when it collides with a wall/object.

If you can't move the camera closer, or you don't want to, you can use a checkerboard/dither node in you material so you can make it transparent. It would still be a masked/opaque-ish material, but it could have a translucent effect: take a look at this video https://www.youtube.com/watch?v=ieHpTG_P8Q0

2- https://www.youtube.com/watch?v=CWzPP5FAYAg <--- this should clear all you doubts toward volumetric simulation
>>
>>594907
Yes, every time you do anything involving changing float values around, always put a 0.0/1.0 clamp on it before it reaches your material’s input node (or even the next blend in the chain), the exception being, of course, for emissive surfaces.
>>
>>594910
>>594907
Okay, thanks. One more thing - under what lighting conditions is the best to tweak all those color and brightness values? I also realized that my SSS color gets too strong in my night scene. I've had it set to only about 0.5 and it looked good under that kind of lighting, but here it doesn't look right anymore. Should I also disable PP volume while I'm doing it? I don't know, a loooot of things affect the overall look, and I kinda tweaked everything together, added some desaturation on PP, some on my textures, increased brightness there, reduced it there etc. It looks good and fairly realistic, but maybe I fucked myself up now since they don't work under every lighting.

And yeah, I'll put clamps to be sure, but I never increased values over 1 anywhere, so I'm not sure yet if that's what is causing the problem.
>>
>>594911
You know what - never mind. Yesterday I probably made an accident somewhere. Everything looks fine now.
>>
>import height map from world machine
>want to texture it

>apply landscape materials based on alpha textures
>looks like shit, not even close to what I see inside WM

>try to literally apply a single diffuse texture over a whole landcape
>at least all those sand, rock and vegetation colors are applied correctly
>gray and bland as fuck though

God damnit, I can spend days on the stupidest shit. Still didn't find a good tutorial which gets it right.
>>
>>594909
1. So, I would have to implement this fading effect for every single material for every object that can theoretically be occluded? Seems a bit heavy-handed as I'd have to do this with basically all the materials.
In the video, the person says that Unreal composites the translucent object from a forward render pass, but if everything is going to be set up to potentially be occluded, doesn't that mean it simply makes more sense to not use deferred rendering at all?
Considering this is such a ubiquitous feature in most games, I figured there'd be a simpler implementation for it.

2. Seems a bit too high-end for what I want, I just want overhead cloud cover, not necessarily full on volumetrics.
>>
>>594918
If you are not going to release on mobile, stay in deferred. And yes, you have to dither every material the same way you give textures to every material; everything has to be the same if you want the same stuff to happen.

Switching to forward will make everything perform and look worse.

The dithering process is so you DON'T use translucency. You still get all the benefits of deferred, and you get a translucent-like effect that is in fact a masked pattern, so no translucency involved. In fact, this is the same shit they do for TW3 vegetation when you put your camera behind a grass plane for example.
>>
>>594916
haha loser.
>>
>>594996
I bet you're the funniest guy in the high school.
>>
Need some advice. I have an object that's kind of a cross-beam similar to pic related; essentially a center junction surrounded by four struts that end in anchor points. I have a modular piece for the center, the strut, and the anchor, and they all neatly attach at 45º angles.
The thing is that the length of the struts are going to vary depending on where they are, and placing all of these manually is going to be a waste of time, so I know that I have to set up a construction script that can do it for me, but I'm not entirely sure how yet.
I know how to use splines to draw out stuff like tubes, and so far I figured out how to spawn the anchors some arbitrary vector locations, although as it is, it's an ugly set of Add Static Mesh Component -> Make Transform -> Get Anchor Location pasted four times, with the make transform scale used to flip the anchors around to face the center point. But I'm not sure how to actually repeat the strut between the two points, so that it maintains its length until it needs to add a new strut instance to make it longer if you drag the endpoint far enough.
Am I doing things backwards, or what is the simplest way to do this in general?
>>
File: StrutFunction.png (304 KB, 1711x725)
304 KB
304 KB PNG
>>595328
Ended up figuring it out myself, with the help of various resources and videos, although it's probably horribly inefficient, being my first BP and all that.
I packed it as a function that takes input from my end-point vector, gets the distance, turns it into an index, and uses that index to multiply by my grid snap amount.
Since I couldn't figure out how to repeat the function for each strut automatically, I simply exposed booleans that ask me which strut I want to create, and select the distance stepping in the axis combination that makes sense for the one I need.
The two extra inputs are the scale of the endpoints and struts, so that I can flip them in the axis I need so that they point in the right direction.
But at least it works, I drag in my center junction with the anchors at their minimum distance, and as I pull out the anchor widgets it creates sections of beam behind them that line up every 20 units.
>>
should i be using the film tonemapper?
>>
>>596069
do you need to?
>>
Can you explain what exactly makes "baked" scene better and more photorealistic, compared to a scene with a fully dynamic lighting? Softness of shadows, gradients of light on surfaces... What? I certainly don't know.

I'm also interested if that has anything to do with the reason people usually use dynamic setup for natural exteriors. I wonder if it can be because the surfaces of foliage and trees are so thin, that you can't really notice any improvements in baked lighting anyway (maybe I'm talking bullshit, but as I said, I don't know the theory behind it). I mean, I know the reason is because of the dynamic weather, but maybe baked lighting wouldn't improve those types of meshes anyway? From my experience, I tried it many times and I literally cannot see any difference when I bake my lighting in nature exteriors, I don't know why.

And a third question which is also somewhat connected to the topic is - why the fuck are UE4 exteriors so shitty? I never saw a really good looking one. I'm trying to make one, but the amount of tryharding to get something good is crazy compared to what I can achieve by just pressing a render button in Cycles or Arnold.
>>
>>597187
Because static/stationary doesn't have to be rendered in real time, lots of extra calculations for smoothness, multibounce, avoiding-banding, etc. can be done. This is at the expense of staying completely static, being unable to move them in any fashion, and not being able to be casted on movable objects.

Because dynamic has to udpate every frame, it has to be dumbed down to be able to be rendered every frame without fucking up player's gpu and getting 16fps. It doesn't mean that dynamic looks necessarily bad, it simply means that it will be more expensive and won't look as good as stationary in exchange of being able to vary among time for certain purposes (such as night/time cycles, weather, flashlights in player gun, etc.) and to be casted in movable objects.

The reason most people do dynamic for exteriors is because night/time cycles and weather. If your exterior won't vary in any way, you can go with stationary and volumetic lightmaps for wind-affected vegetation and the characters (deus ex mankind divided doesn't change prague, and when it does, it is after a cutscene, but doesn't change in real time).

If your game does vary weather and/or time (such as most open world games), you will have to go with dynamic; but this doesn't mean it will look like shit, as you have a variety of tools (such as DFAO, height based fog, etc...) that you can use to improve the general look of the scene to "compensate" for dynamic lightning.
>>
>>597191
Thanks. I'm still not sure why I can't see any differences when I bake stationary light. I noticed that if I use LPV for my dynamic directional light, even if I change it to stationary and bake it, that damn LPV is still active. How is that possible if it should only work with movable lights? Is that the reason I don't see any changes? Does it even bake if I don't turn off LPV manually?

Also - what happens to dfao if I bake my lighting? E.g. if I keep skylight on movable and use stationary directional light. And is it better to have both lights on stationary or to keep the skylight on movable?
>>
Are there any Unreal 4 games out with custom level support aside from the titles accessible in the Epic launcher? (UT4, ARK, Robo Recall, that side-scrolling one etc.)




Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.