I'm currently working on a highly interesting and promising project focused on procedural texture damage/weathering.
The goal of this project is to create a UE4 blueprint system and a set of textures that can apply damage to the tex' in runtime, depending on a set of environment sliders such as heat, humidity and radioactive radiation.
The target is to create an environment that combines that quality of handmade maps, with the versatility of procedural parameter based damage over time.
To do this I've been working on a system that manages environmental factors as editable parameters.
With the usage of a couple of sliders in a simple UI the user will be able to adapt the type of weathering and damage to the scene.
In the meanwhile I'll be posting my WIP shots here, any and all feedback is highly welcomed!
[MOST RECENT WIP]
Replies
Seems to work quite well fading in over time so far!
Right now it's hard to tell how the performance of the final product will be, seeing as how there's no actual textures in there.
Final product shouldnt be too heavy on the fps though, everything behind the scenes is rather straightforward.
I'm also working on a complete scene to showcase this in, I'll be posting some WIP shots soon.
[VIDEO]
I'd be glad to hear any feedback while I continue to work on this project though!
I'll be sure to post the WIPs and results in here.
Currently working on the performance issues, for a moment I feared it was my huge level blueprint that was clogging things.
Luckily, the issues were all caused by the moveable lights, and the post process supersampling (was 200% to 1080, is now 150%)
Fixed it, fps while recording is now a stable 45fps!
Next up, the elevator. Doors start and end in the same position and thats just a bit weird.
Stay tuned for more updates!
(sidenote, what is the song name )
nevermind, found it,
Esben and the Witch – Wooden Star
Quick update, working on tiles that fall from the ceiling, animations for both lift doors, and more physics props in the scene.
Cool stuff, stay tuned for more vids!
I'll also be posting all of my project files on gumroad afterwards (for free), so that anyone who would like to research it, can.
Depending on how much wind this project catches, Ill go for more of a tutorial type of video, in which I explain how to do certain things.
But a basic tech breakdown should be included either way!
Covering some of these corners with furniture or props such as the sign is a great way to give the scene a more interesting shape. This would also suit the kind of atmosphere you're trying to achieve. Too many obvious right angles such as these can make it look boring.
Here's another update, lot of dynamic props (too many y/n?) , got the basetex for the ceiling done too.
Next up, making the roof material dynamic, swapping blockout materials on props for legit ones, more polishing!
I'm still not sure what caused the issue, but I've managed to find a workaround.
Fog and a few small particles have been added!
Currently working on the ceiling blending, shouldnt take to long.
Once that's covered, I'll move on to the rest of the particles, and then by popular demand, the frozen groundpuddle!
[image deleted by puush.me]
Going to do something about the light coming from the door on the right too, feels like its coming on too strong at the moment.
I'll probably fade it out as time progresses, going to do some tests.
I've completed and implemented the ice material and blending, here's a few shots of what it looks like.
Unfortunately my desktop motherboard died and my laptop doesn't record it well, so it's just stills for now.
I'd be glad to hear your feedback!
Looks pretty cool fading in too, the first spots to freeze start at 0.05 emmisive strength and gradually strengthen to 0.2, depending on their location.
The Day After Tomorrow has a good example, it's a terrible movie, but there's a part around 20 seconds in that has the frost at that opaqueness I'm talking about:
https://youtu.be/E87raPj9m0A
Not too hard to implement either, so I should be working on it pretty soon.
That, and the still unfinished props, about time they got done correctly.
Here's another update, currently polishing the floor material!
https://www.artstation.com/artwork/yKVL3
Research footage and my progress in a quick vid
I'm wondering why the floor tiles are much cleaner compared to the wall tiles. I see decals but the tiles themselves looks very clean and sterile. The one clean room version of your scene from one of the images looked perfect with it too. Really awesome stuff.
Anyhow I really like how you achieved the weathering effect! Could be neat in some cinematic sequence
The problem is that I'm using actual light in the spot can't have shadowcasting enabled.
The reason for this is simply the falling light. It'd look really weird if the lamp cast shadows on the wall next to it, even when it was supposed to be a major light source.
I might do some tryouts in which the lamp doesnt fall, and the actual light source DOES cast shadows.
Either way, I'll be sure to keep you posted!
Here's a quick gif showcasing it nice and quickly!
(Made by 80.lv)
Thanks for the comments guys.
Still, the final product could still be way better, so I'm continuing to work on it at last.
Upcoming priorities are the ground shader (you might have noticed that the roughness switch is broken in the final version) and the unfinished props.
Currently working on the liftdoor moss that wasn't moving with the doors.
Up next; dynamic moss meshes that grown based on the input parameters.
I've done some tryouts with Max' blendshapes, and things are looking quite good at the moment.
Right now it's a weird base shape, but now that i've got the implementation down, It'll be switching to moss meshes soon.
Also did some work on the bottles, meshes have been changed, and they're hollow now, which translates to way better physics fragments.
Up next, probably changing some more props.
Still not entirely sure on what i'll do with the table and guidesigns on the left.
Update; started applying force to props that should fall over, massive visual improvement from having to tilt the props far enough so that they'd fall over when gravity was activated.
It might have given me more control over the fall, but the still objects looked weird in their pre-fall weightless rotations/positions.
The force pushing over time is great, and subtle enough to make it look natural too.
Only 'issue' here is the inconsistency in the props' fall and shatter.
Gonna be doing a ton of new blueprints soon, seeing as how every prop has a different force tumble threshold.
Made a door, trashed it completely.
Also made a a table, trashed it aswell.
Got rid of the door, feels better in the overall image, less focus on the side of the image at the moment.
Still some edges that are waaay too hard, anyone have any ideas on what to do about em?
Just the frame though, the actual door would probably look out of place and attract unwanted attention.
I should probably also fix the blueprint that adjust the light strength in the grunged up version, way too dark as it is.
Up next; the rightern hallway-whatever-thing's hard edges, and then an upgrade to 4.14 for fancy cables and contact shadows.
Anyone use either of those two new cool features yet?
Came up with a doorframe as filler, works great, it's subtle enough to not stand out while breaking the edge.
Also annoying was how the ceiling tiles were falling in structured waves, something that was quite obvious (to me at least).
fixed it by dividing each activation wave into 3 new ones with randomised drop times.
Comparison gif below:
Next up; Radiator, some sort of satchel (optional), better tubes/pipes, the terrible lift doors, decent plints between the walls/floor.
The thing with this way though, is that i need a specific colour blend sequence for each colour to each different colour, and ways to check them.
Probably going to look into a 3-state enum that can tell me which transition to choose from.
It's blending without going radioactive, so the hardest part of that is over! :-)
[Fog strength was boosted 10 fold to facilitate testing]
Also, my 4 different LUTs seem to not cycle as they should (not at all to be specific) so i'll be looking into that soon.