[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: pbrcomparison.jpg (622 KB, 1920x1963)
622 KB
622 KB JPG
wow all this processing power and years of research to get a render that doesn't even look better than what we had in 1995.
>>
>>599809
DELETE THIS
>>
The point is that people usually didn't know what the fuck they were doing or how materials look, and made everything look like plastic. See: most video games.
>>
>>599809
Spoilers: technological progress is overrated, people are retarded and/or clueless
>>
>>599809
What am I looking at exactly?
>>
>>599829
Yeah, I'm not going to lie, but I don't see much of a difference in OP's picture aside from the rust's normal map being more obvious in the bottom right.
>>
>>599809
Yet, the multimillion dollar industry couldn't figure out how to pull a fuckin' specular map. So awesome, can we suck the 90's dick yet?
>>
>>599809
>traditional
>dude numbers
>pbr
>generate a map that hooked up by any pbr app plus blend them all together
pbr is great, fuck off
>>
>>599835
yeah funny coincidence all vfx after 2000 have looked like garbage.
>>
>>599837
Kek those textures remind me of dark souls
>>
Does anyone know why Uncharted's foliage looks like plastic? Did they do it on purpose every time? I mean, I don't understand it, I know it's using PBR, but the vegetation doesn't look realistic at all.
>>
>>599850
>foliage looks like plastic
>it's using PBR
There's your problem, even "professionals"don't give a fuck about properly utilizing the tech.
>>
>>599809
PBR is about simplifying the workflow. The principled shader Disney developed has 11 basic controls for determining various surface properties, and everything else is derived automatically based on assumptions made about how a material with such properties should behave.
The shader should be usable for any type of material, so that every object in any scene would use the same shader, instead of having a dozen different shaders for everything and figuring out how to make them work together.
The issue is that that to get the best result on traditional shaders you can't just get away with dialing in a value or controlling the amount with a map and calling it a day, because certain characteristics like specularity and glossiness change properties depending on the viewing angle and intensity of light, requiring you to set up falloffs, blends, and maps that are sometimes grayscale but sometimes have color in them, and none of it really makes sense unless you're into color science. With each setting a PBR shader has for you to tweak, the only thing that matters is the amount, not the how or why.
>>
File: steak.png (1.09 MB, 1280x720)
1.09 MB
1.09 MB PNG
>>599809
>PBR
>>
>>599860
And that's a retarded way of thinking, considering not every material works that way. Especially organic and translucent materials.

PBR looks fine for most inorganic materials, like plastics and metals (and even there, it's lacking certain options I feel). And yet skins and other stuff that require SSS still look rather fake-y. I feel like those who originally made the node should have put in some more effort and options when it comes to the shader itself.

>>599850
>uncharted
>realistic
>naughty dog
Well there's yur problem mate.
>>
>>599857
>>599865
Wow, I can't believe how I can never get a normal serious answer here, only greentexts... I want some explanation because I don't understand this, Uncharted is obviously stylized, but why are they saying they are using PBR?
>>
>>599870
2 things:

- this is 4Chan, serious answers are hard to come by. So don't get your hopes up.
- this ties into the first point, but I don't think they ever gave an answer as to why beyond "it looks better".
>>
>>599865
>(and even there, it's lacking certain options I feel)
>I feel like those who originally made the node should have put in some more effort and options when it comes to the shader itself.

What, do all PBR shaders use the same code or something? Am I the retard here or is that you?
>>
>>599837
>all vfx after 2000
>godzilla 1998
Really jogs my peanuts
>>
>>599874
They share the basic setup- A set amount of options that are rather broad and not always specific. Uncharted with its plastic-looking leaves being a good example.

So yes retard, you are the retard. Now go talk to mommy and she'll coddle you.
>>
>>599809
>PBR
>thinks, it's about the looks and not about workflow
Faggot.
>>
>>599809
PB shaders are just objectively more realistic because they accurately utilize physical equations and their approximations for light bounce, reflections, energy conservation and so on.
That's why they're called Physicall Based. They weren't made for the sake of little mongoloid brains like yours going crazy over supposedly "superior graphixx",
but rather to optimize the texturing pipelane and to make the materials look more consistent.
Of course PBR materials can still look like dogshit, and in some cases worse than their non-PBR counterparts, but that depends entirely on the artist.
>>
>>599878
my favourite post of the year so far right here
>>
File: 1488194354787.png (158 KB, 816x754)
158 KB
158 KB PNG
>>599887
THIS
fucking this

PBR doesnt mean that all the work is done for you, you still have to know how to shade and create texture maps properly, it's just a way of streamlining the workflow based on real-world physical values and properties that materials have.

>>599865 I won't argue that it indeed works best for metals, plastics, and whatnot vs organic materials, where you still have to cheat the SSS. Still, PBR brings some standardization regarding that, and real-time shaders are become more and more able regarding translucency. It's never gonna be 100% physically correct anyway, but the process is eons better that what it used to be.

But of course the mongoloids of /3/ will fuck that up and claim it's a fraud; that's what clueless amateurs do.
>>
>>599913
There was a thread on /v/ where they proved that graphics haven't gotten better since like 2010
>>
>>599918
>graphics
could you be even less specific please?
>>
>>599878
godzilla looks great
>>599841
you know it's a photo right?
>>
File: 1466474807488.png (165 KB, 639x462)
165 KB
165 KB PNG
>>599920
>you know it's a photo right?
>giant lizard monster in a blockbuster movie
pick one
>>
File: waxy.jpg (72 KB, 617x838)
72 KB
72 KB JPG
>>599887
>PB shaders are just objectively more realistic because they accurately utilize physical equations
are you serious? I could tell you at least one major physics flaw with the common implementations just off the top of my head.
>>
>>599919
Video games by AAA
>>
File: 1485718797894.jpg (19 KB, 426x304)
19 KB
19 KB JPG
>>599924
Humm.. but don't the better performing consumer products (mostly PCs and next gen consoles) allow for more/better real-time rendering? Number of objects, texture sizes and material definition, particle effects, light/shadow quality, etc. I mean, today's graphics clearly have the upper hand over what we were seeing 7-8 years ago.
>>
>>599922
>I could tell you at least one
>Doesn't proceed to say any argument
Well?
>>
>>599925
Not according to them. I hardly play games. Surely the in game cutscenes are better like in the StarCraft LOTV game
>>
File: physically btfo.jpg (58 KB, 924x684)
58 KB
58 KB JPG
>>599874
yes. all pbr basically are implementations of the cook-torrance model. you can plug various different models in like GGX or beckmann for the specular as long as it's energy conserving. There's even a blinn term modified to be energy conserving which would qualify as PBR. There's also different models for the shadowing function and the diffuse term.

what disney did that is unique is set the whole thing up to have input values which are "plausible" making things easier on the artists.

but it's really not much better than alternatives, just more expensive to calculate. which is good for the renderfarm and gaming gpu businesses.
>>
>>599927
Yeah I don't play much either, but I looked at 2010's top games (Red Dead Redemption, Bioshock 2, God of War 3) and it's fair to say that we've come a long way from that. Maybe techniques haven't changed THAT much, but visually it has. Even the Switch renders better stuff than back then. Hardware probably makes a huge difference.
>>
>>599931
>2010 console games
running on middling 2005 hardware.
try PC games.
>>
>>599926
the fresnel term is not evaluated per microfacet. but naively across the surface normal.
and there is no integration over the microfacets for snell's law.
>>
>>599933
ah yeah good point, I'll check
>>
>>599809
left tree is less blured
>>
>>599936
honestly 2011 is probably more the year, where you see games like battlefield 3 catch up to crysis. and even crysis 2 and rage on those same consoles which is probably the most impressive optimization work you'll ever see.
>>
PBR is a nice concept, Disney is just so far ahead they don't even put effort, like NVIDIA.
>>
File: maxresdefault.jpg (135 KB, 1920x1080)
135 KB
135 KB JPG
So this is not an improvement?
>>
>>599995
going to more powerful hardware and not cherrypicking screenshots helps
>>
>>599809
>creating materials that react inaccurately with lighting.
PBR is not a meme. PBR gives Devs the ability to create photo accurate material/ lighting and standardizes the process so all materials behave consistently.
>>
>>599995
>"Nope i love flat lighting and lower res textures"
>>
>>600003
you can do "photo accurate" just as well or better with phong. it's 99% artists and 1% tech.
>>
>>600008
>better with phong
so out of all the things you could be irrationally fanboyish about you picked a fucking math equation from 1973? is this the new blender shitposting?
>>
>>600027
Fuck off, kiddo.
>>
>>600027
I picked it because it was first and most primitive. And it still gets good results all the same.
>>
File: 1484507221501.png (27 KB, 477x387)
27 KB
27 KB PNG
>>600003
>>600027
this anon is correct

>>600008
>>600028
>>600035
this is some retarded blenderfag vomit
phong is not physically accurate
sure you can do realistic stuff in a non-physically-based workflow, but all your parameters and tweaks and settings and whatnot will be case-by-case trial and error bullshit, which is not suitable -acceptable even- in modern day professional workflows

why the fuck would you not wanna work in a physically correct manner? it's not even restricting you to realism, you can still do kickass cartoon stuff and cheat past true PBR when needed.. only a dumb amateur would argue any of that.
>>
File: hqdefault.jpg (10 KB, 480x360)
10 KB
10 KB JPG
>>600168
>arguing on /3/
Mate, your're literally retarded
>>
>>600170
what a newfag
/3/ is 99% arguing, hopefully someone learns something out of it fucking once in a while
>>
>>600173
nobody learns anything here lad, we all just die slowly of old age but some by the self inflicted gunshot. rip little bro
>>
>>600168
>phong is not physically accurate
neither is "pbr". being "physically accurate" means nothing.

a well made phong material will be just as consistent as any other material. the idea that pbr is somehow different just because it enforces common sense limits is daft. only used by tech illiterates.
>>
>>600170
You know you're on 4Chan right? Arguing is all we, and every other board does.
>>
File: 1408734841743.jpg (33 KB, 350x401)
33 KB
33 KB JPG
>>600180
like i had said: only a dumb amateur would argue
i guess that concerns you, cause you're literally arguing with the fundamental mathematics of the different BRDFs

https://pdfs.semanticscholar.org/023a/5c184b8cb9e9c5f2ac61fa679fedc1a478f8.pdf

>the Phong model is a computational convenient method to analytically approximate the reflectance properties of a small set of materials.
>In general however, not too many materials
have reflectance properties of this limited form. > it is not actually physically plausible
>the lack of physical validity is problematic

this is introduction material

>being "physically accurate" means nothing
it actually means that computation is done to be as faithful (ie. not biased) as possible to real-life light and material behavior and physical effects

you can thank me later for doing all the goddamn research for you
>>
>>600182
inb4 you disregard all the information only because copy/pasting fucked up the greentext
>>
>>600182
Your arguments are invalid and you are a doo doo head.

There, back to the old form. Now, everyone, neck yourself.
>>
>>600182
call me when you can write a PBR algorithm faggot. it's complete bollocks. it's basically wrong. and it provides no benefit to the informed artist over older, faster and more convenient methods.
>>
>>600188
You first anon.

What? you said everyone. I don't see how you can get a free pass out of this.
>>
File: amerifat.png (178 KB, 1000x1000)
178 KB
178 KB PNG
>>600196
tried and failed, picture related
>>
>>599850
They use a slightly stylized look intentionally.
>>
>>599809
Wow, incredible, you've managed to compare two free-floating orbs and one looks different to the other. PBR is truly owned now.
>>
Jesus christ does this thread show no-one here actually works in the industry. PBR is not a technological advancement by itself. It's supposed to be a set of rules that quarantees that the materials created with the principle end up looking the same in all renders.
>>
>>600269
nobody knows why pbr was created but some on /g/ and /v/ say it was as another bulletpoint to sell software after cgp cannibalized all their sales
>>
>>600270
PBR was made >>599887 for the sake of little mongoloid brains like mine going crazy over supposedly "superior graphixx". and then appropriated as a marketing term by 3D vendors who started making stuff up about workflow improvements to sell it to tech illiterate artist monkeys who are paid by the frame.
>>
>>600182
Fucking this
This is the post

>>600188
>doo doo head
Kekd
Pretty much sums up the state of most /3/ discussions
>>
>>600269
I do, and you're correct.

>>600270
>nobody knows why
>I don't know why
fixed that for you

Shading/lighting/rendering was a huge mess before physically-based workflows and renderers came along, just listen to John Carmack:
https://www.youtube.com/watch?v=IyUgHPs86XM
>>
>>599809
I always assumed it meant that one shader can be used in mostly all lighting situations.
>>
>>599922
>and their approximations
>>
>>600347
not an approximation. straight up an error.
>>
PBR is nasty. I didn't think anyone drank it any more. Why are you all discussing beer anyway?
>>
wef
>>
daily fucking reminder that PBR is a streamlining of content creation, not an improvement in rendering
>>
>>599860
THIS SO MUCH.

If you think PBR is about "looking better" you need to get out. PBR is about making texturing intuitive, straightforward, and cross-platform. (across realtime/raytraced, across different rendering engines, etc.)
>>599934
That weak sauce shit is all you come up with? Fresnel implementations differ by rendering engine, nothing specifies which approximations engines are and aren't are allowed to use.
>>
>>600416
it's not "weak sauce". it's very important if you want to claim to follow physics. most even multiply the specular directly with the lambert term which is downright retarded.

PBR is about being more accurate, but it's not really much more accurate. the "muh workflow" shit is marketing garbage.
>>
>>600379
I drink PBR because it's tasty and is an excellent source of carcinogen ingredients
>>
File: 1485234345617.jpg (34 KB, 480x522)
34 KB
34 KB JPG
>>600421
>the "muh workflow" shit is marketing garbage
mfw

>So, why do you think you'd be the perfect candidate for this 3D position?
>Well faggot you'll find that my beginner status is overly compensated by the fact that I'm a blender user, which is toptier; as you know autodesk is cancer. Also I exclusively use the phong shader, since being physically-correct is a meme and obviously a scam. 3D hasn't evolved much since 1996 anyway, so why bother educating myself or questioning a workflow that fits MY needs. Everyone on the internet forums think I'm right and talented.
>...please leave

Keep being unemployed m8, it's as simple as that
>>
>>600431
>get btfo
>start appealing to job opportunities rather than logic
>wagies think their opinions matter
>>
>>600431
Thanks for the insight. Now back to your cubicle.




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.