Anonymous at Sun, 20 Jun 2021 09:58:43 UTC No. 834086

Anonymous at Sun, 20 Jun 2021 10:00:29 UTC No. 834087

Anonymous at Sun, 20 Jun 2021 11:26:09 UTC No. 834097

Might as well use AI to generate doors

Anonymous at Sun, 20 Jun 2021 11:35:45 UTC No. 834098

It is nothing but a disgusting flex of a technically minded artisn't for other artisn'ts to lose their minds over it and give him fictional internet points on a slave picking marketplace website so industry overlords notice HIM instead of another rat racing artisn't and give him the opportunity to work on a next blockbuster bideo gaming hit shooter game for a McDonalds-tier salary, while he will still be happy about it anyway because his name has been stamped in credits together with other 4000 artisn'ts, and that makes him "truly" fulfilled knowing that he was a part of something great and totally influenced the project with his amazing one of a kind, irreplaceable art skills, and will now brag about it for the rest of his life.

Anyway, I'm gonna take my meds now

Anonymous at Sun, 20 Jun 2021 11:49:39 UTC No. 834101

>>834098
You don't matter and your opinions are irrelevant. If he wants to use nodes to make his work then why not. Want to say that this applies to me? Doesn't work buddy as I'm not delusional into thinking that my drivel is important.

Anonymous at Sun, 20 Jun 2021 12:06:05 UTC No. 834103

>>834098
>It is nothing but a disgusting flex
Correct, but the rest of your post is a disgusting seethe and cope.

Anonymous at Sun, 20 Jun 2021 12:16:22 UTC No. 834105

>>834085
yup, if its hard surface I would just model it.

Anonymous at Sun, 20 Jun 2021 12:41:08 UTC No. 834107

>>834101
>>834103
Triggered artisn'ts found themselves in my post.

Anonymous at Sun, 20 Jun 2021 12:45:07 UTC No. 834108

>>834087
If this has been made with regular modeling instead of spaghetti, it probably wouldn't get even third of those likes. This only proves that, in fact, end result is not the only thing which matters.

Anonymous at Sun, 20 Jun 2021 13:07:01 UTC No. 834114

it's something people do for practice and because you enjoy SD it's not that deep.

If you don't experiment you don't find new solutions to problems, I'm sure doing trim sheets or foiliage in SD sounded insane in the early days too and it's now commonplace.

Anonymous at Sun, 20 Jun 2021 13:12:50 UTC No. 834118

>>834114
Isn’t foliage still sculpted a lot?

Anonymous at Sun, 20 Jun 2021 13:18:03 UTC No. 834122

>>834114
>artists "problem solving"
No, bad artist! It puts the nodes in the graph or else it gets the hose again.

Anonymous at Sun, 20 Jun 2021 13:22:16 UTC No. 834123

>>834118
both are popular yes.

I'm not great at SD so I sculpt then bring the bakes into painter but you can envision the advantages to doing it entirely in SD I'm sure.

Anonymous at Sun, 20 Jun 2021 13:25:32 UTC No. 834124

>>834107
Not claiming to be an artist, I fully accept my fate as a craftsman

Anonymous at Sun, 20 Jun 2021 13:44:23 UTC No. 834130

> it looks dumb as the video of that guy painting normal maps instead of generating them

https://youtu.be/oJ0vrDTA0sA?t=261
lul

Anonymous at Sun, 20 Jun 2021 13:51:11 UTC No. 834133

>>834123
Oh I know, for sure. But yeah, it seems too complex for me to do it in SD as well. I even followed some long masterclass on it and still got shitty results. I've been working on a simple template for Substance Painter though, so I can just paint the shape and get the height, opacity and some albedo info procedurally. That way I have more creative control, but still have to do more stuff manually.

But still, from what I've seen, that only works for leaves and some simple plants like ferns, but if you need a plant which can't be fit on a single plane, you would still need to either use SpeedTree to create a plant from your procedural leaves and then bake again, or just sculpt everything from scratch.

From what I've seen in CGMA foliage courses, they sculpted everything. I mean, I like it as well, but it absolutely sucks if you realize you messed up something when you drop it in the engine and have to do a large part of it from scratch.

I'm staying away from scanned textures though, that's boring.

Anonymous at Sun, 20 Jun 2021 14:03:31 UTC No. 834135

>>834133
yeah and likewise there's a whole shitload of trim sheets you'd never want to attempt in SD. Never one answer for anything with 3d eh?

Anonymous at Sun, 20 Jun 2021 14:45:29 UTC No. 834148

>>834107
I'm a CAD monkey trying to feed his family, not a beardless "artist" or autistic coder

Anonymous at Sun, 20 Jun 2021 16:41:33 UTC No. 834163

>>834130
Not watching that but pixel perfect normal maps have to be painted manually

Anonymous at Sun, 20 Jun 2021 16:48:10 UTC No. 834165

>>834163
Literally what I did for texture packs(normal/spec/glow) for both Doom and Doom II and for all the textures in Duke Nukem 3D, the later making it into the commercial release of Duke Nukem 3D: 20th Anniversary World Tour. Took forever. No pun intended.

Anonymous at Sun, 20 Jun 2021 21:26:34 UTC No. 834282

>>834085
>the video of that guy painting normal maps instead of generating them
Well to be fair, that guy was painting them for a very specific purpose and workflow. You might think it's unnecessary, but you're not seeing the bigger picture and just thinking of it as just painting normals because he's retarded.

For the most part, he was painting normals on top of already painted images, not 3d models, and using them as a way to change the lighting and stuff inside a painting. Sure he could generate a shitty normal/bump, but it'd still be shitty, and wouldn't have the look he was going for.

Anonymous at Sun, 20 Jun 2021 21:32:57 UTC No. 834283

>>834085
Also to add to something a bit relevant with the nodes...
It seems like a fuckton of bullshit nodes for a visually simple output, but if you're just looking at the number of them as a whole, then that's kind of stupid.

You kind of have to think of them as chunks doing small parts of an operation, like gears in a clock, or parts of a car.
Like an entire car may seem like a complex piece of machinery, but its basic parts still amount to simple machines and actions that add up to a single output of making a car go vroom.

It's absolutely a flex showing all your nodes like a smug douche.
But working with nodes, even simple things can ramp up to look visually complex when looking at the nodes, but they all break down to really simple things. Odds are there's a ton of the same repeated functions all over the place with it. Like making circles/squares/noise, fucking around with the placement, and doing all that. It's just compounded by the amount of shapes, and multiple materials.

Anonymous at Mon, 21 Jun 2021 16:30:20 UTC No. 834419

>>834085
>>834086
1. You can resize the texture
2. You can port it across different platforms
3. Can be used for film and gaming

Anonymous at Mon, 21 Jun 2021 17:18:51 UTC No. 834435

>>834085
>Why not just model it
Because the procedural approach means that it allows you to change some of the initial inputs to create a completely different result, meaning that you can reuse it, share it with other people, integrate it into larger systems, etc. Or simply because the guy doing it found it challenging or fun.
>Does this way feel like cheating, it looks dumb
The only one that looks dumb is you, honestly. I have no idea how something like this would be considered "cheating" ? Because you can't do it ? Who or what are they cheating ? Procedural generation has massive benefits and there is a reason why its used for so many things, from materials to volumetrics.
Personally I would prefer to code something like this instead of using this visual node system but I guess the later approach is more comfortable for some people.
>>834098
>Anyway, I'm gonna take my meds now
Try taking them before posting next time, anon

Anonymous at Wed, 23 Jun 2021 10:56:56 UTC No. 834824

>>834435
>meaning that you can reuse it, share it with other people, integrate it into larger systems, etc.
No one does this with a graph as complex as this. Learing which node is meant to do what and which inputs give good results or break everything takes far more time than just modeling the thing. In 6 months the artist himself will have forgotten 80% of it and he too will prefer just modeling it.
Reusable node trees are mostly base materials (wood, metal, plastic) that can be slightly "bent" via changing parameters to accomodate the taste of an artist.

Like >>834098 said, it's a flex to impress low IQ artists since even midwits will immediately realize how useless something like that is.

Anonymous at Wed, 23 Jun 2021 11:38:24 UTC No. 834826

>>834419
How can you use it in games if it's a texture? You won't use a displacement for a door, that would be more expensive for no reason.

Although I do question if this could be an approach in the future with Nanite in UE5. Could we just use procedural textures with displacements to make somewhat flat things like ornamental walls, doors etc.?

Also, isn't Substance working on procedural modeling tools? I think I've seen them show something like that inside of Substance Designer.

Anonymous at Wed, 23 Jun 2021 13:06:43 UTC No. 834827

>>834826
>How can you use it in games if it's a texture? You won't use a displacement for a door, that would be more expensive for no reason.
Depends on the game and its graphic settings. You could easily use parallax occlusion + normal maps, tessellation for when you are close. Or you use it on a background asset where you wouldn't get close and just use a normal map alone or in combination with parallax occlusion.
Or you add a couple of nodes in Designer which create a simplified height map, then feed it into Houdini to shit out a simple mesh with extrusions for the rough shapes. Or you simply model it in like 20 seconds.
Basically you would have to add a tiny bit of work to optimize it for game engines.
>Although I do question if this could be an approach in the future with Nanite in UE5.
You could just create an detailed mesh from the height map and use that.
Stuff like this could be done while "installing" or loading the game for the first time in order to safe space on the disk.

Anonymous at Wed, 23 Jun 2021 13:16:11 UTC No. 834828

>>834827
That could be very cool, but then why is nobody doing it? I’ve never heard of a game using displacement for regular props, only for the ground and similar things. Also, that would only work for static meshes, right?

Anonymous at Wed, 23 Jun 2021 13:18:57 UTC No. 834829

>>834828
Oh, and not to forget, the resolution would be shitty for bigger assets where trims are usually used. Idk, it all sounds cool, but in practice you realize it just isn’t viable and only a boring and slow way of doing things actually works.

Anonymous at Wed, 23 Jun 2021 13:54:12 UTC No. 834838

>>834828
You do know that there is tessellation in game engines for years now and that games have used it for years?
You can tessellate animated characters, has been done before.
Consumer hardware is the limiter not the tech or engines.
>the resolution would be shitty for bigger assets where trims are usually used.
You do know that tessellation is camera distance dependent, just like Nanite?
>it all sounds cool, but in practice you realize it just isn’t viable and only a boring and slow way of doing things actually works.
Nope, I just explained ALL of its possible use cases for you, but you seem to have shit in your brain because you seem to be unable to process the data I just gave you.
Designer + Houdini gives you all the procedural power to get whatever the fuck you want as mesh out of this procedural approach.
Did you not think this through? If you have a procedural asset you can create hundreds of variations inclusive the meshes with all LOD's or a super high detailed mesh for something like Nanite or for an VFX pipeline.
>then why is nobody doing it?
Which retard says that nobody is doing it?

Anonymous at Wed, 23 Jun 2021 14:35:09 UTC No. 834844

>>834165
Nice, were you paid for that?

Anonymous at Wed, 23 Jun 2021 14:42:37 UTC No. 834845

>>834838
You still haven't posted examples of it, and I'm genuinely asking. Why can't you communicate without insults?

I've seen it on the ground, walls, roof tiles and similar, but that's it. I can't think of any examples where they would be used on hard surface props.

I'm interested in the subject because all I've ever been reading about these hard-surface SD materials throughout the years is that they are pretty much useless in production. That seems to be the general opinion in 3D communities, but it is a good practice.

And there are things like making whole facades in SD, but from what I've seen and been doing myself for other AAA studios, it's still pretty much manual modeling and trims + tileables, nobody used these kinds of materials in the production. I wonder if there are any examples of stuff like that being used in games, and where you can actually go close to it as well. It's still modulars + trims.

Anonymous at Wed, 23 Jun 2021 23:32:28 UTC No. 834937

>>834844
Paid and credited as staff.

Anonymous at Thu, 24 Jun 2021 09:16:35 UTC No. 835025

>>834283
This. In the realm of digital logic design this would be the equivalent of showing every logic gate in between the input and output