[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] [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: romain-rouffet-final-50.jpg (1.09 MB, 1920x2145)
1.09 MB
1.09 MB JPG
https://www.artstation.com/artwork/QJGgr
>this was done in 15 min
Is this real life?
>>
>>564139
Yes,now try to actually use the model.
oh wait you cant because 5 billion polygons and mesh holes.
>>
>>564140
It's still unbelievable, though.
>>
>>564140
What do you mean mesh holes, anon? Looks consistent to me. And there must be a way to reduce polygons using topology or something.
>>
File: C6vAA_fWcAIWa4S.jpg (66 KB, 720x918)
66 KB
66 KB JPG
>>564139

>first find a bike (minimum 2weeks searching in srap dumps)
>get a bike owner to allow the scan(another 2 days)
>scan (15minutes of penis range)
>pay the owner fees for using his shit in your game
>hire 3d guy to make the variations of this bike.
>>
useful for a diffuse map / normal map, and maybe you could pull some other maps from it, but as far as the mesh goes it's useless.

>>564143
spend 2 days finding bike owner, tell them their bike can be in your game if they allow you to scan it
>spend 15 minutes scanning it
>pay him nothing
>>
you cant scan it like that in photogrammetry.
He must have taken 500 photos +
Which never took 15 min to make
a pc would need a day at least to calculate the model.
>>
>>564148
>its russia
>buy the scan, the bike and the owners friendship with one bottle of Vodka
>>
File: facepalm.jpg (35 KB, 600x600)
35 KB
35 KB JPG
>Reconstruction with 652 images in capturing reality. Rendering in Marmoset toolbag 3 ( 40Millions tris and 32 textures in 8K)

This is hilariously impractical. There is also no way this only took 15 minutes to make.
>>
Why did he not post a wireframe though, probably looks like shit. Though the close up shots look really good
>>
>>564139
It's a re-constructed mesh for a point cloud, in other words it's massive and has no topology. It's 100% useless.
>>
>>564153
Can somebody explain why is bad topology automatically useless? Can't it be used as a non-movable prop in a scene at least?
>>
>>564154
because it takes processing power
go to sketchfab and try to rotate a model with 200k tris, the response time will be abit slower (but noticeable) than a model with 2k tris
>>
>>564154
> why is bad topology automatically useless
I recommend you read Digital Modeling William Vaughan.
Its on cgpeers.com for free.
>>
>>564152
>40m tris
>wireframe probably looks like shit


you think?
>>
my bud makes a lot of 3D scans and only reason he does it is cool textures and easier modeling via retopo. scans are garbage for anything other than cool rendering to show off.
>>
>>564160
>and easier modeling via retopo
so they aren't
if you retopo
>>
>>564139

Mesh is a dense mess of course, but i'm very impressed at the amount of detail you can capture using photogrammetry. I guess the software has become very advanced since i last looked into it! To think that a fancy camera costing less than an expensive laser or DLP scanner can produce much better results...
>>
>>564155
Ok, that's true, thanks.

>>564156
Didn't hear about that, will download it now. Thanks.
>>
>>564157
I still want to see it
>>
>>564139
>"Muh realism" 3D modelers BTFO
>>
>>564139

All this guy does is scan shit and he has 4k likes and nearly 100k views. This is the equivalent of an instagram whore.
>>
>>564194
>All this guy does is scan shit and he has 4k likes and nearly 100k views. This is the equivalent of an instagram whore.
looks like he's replacing your ass and you're already SALTY heh heh heh
>>
>>564139
so his 3D skill is...taking photos?
>>
>>564200
>Everyone else's 3D skill isn't is literally modeling from photo reference.
WHEW
>>
>>564204
So you actually don't understand the difference between modelling and just taking photos?
>>
>>564206
its the same end result, so theres no difference. Its like when the factories in china are now almost 100% machine automated and every gook who worked there previously is now on the streets
>>
>>564207
>its the same end result, so theres no difference.

The end result in this case is a model with 40 million tris that uses 32 different 8K texture maps. In practical terms, it's almost completely useless. Showing it off in Marmoset Toolbag is pretty much the only thing you can do with it.

Conventional modeling could easily cut the poly count down by 90% or more, with only a negligible loss of visual quality, and on top of that you'd end up with a mesh that actually has proper topology and could be animated.

So no, the end result most definitely isn't the same at all.
>>
>>564208
>The end result in this case is a model with 40 million tris that uses 32 different 8K texture maps. In practical terms, it's almost completely useless.
zremesh it. Its non deforming, so topology wont matter. I take it you havent been doing CG long
>>
>>564207
For starters, he can only do things that exist in real life, so any artistic input or imagination goes into the trash. Secondly even a monkey can take photos of things from random angles and upload it to a computer which does the entire model by itself, so anyone can do his job.
>>
>>564211
mate...have you seen any movies or games lately? 99% just inserting modelled extra cars and props into the bg with after effects, like in the Walking Dead. As an industry, you are done, just like those gooks in china car factories. Sure, you might have a few hundred making custom made shit, but the boom days are long, long gone.
>>
>>564214
>99% just inserting modelled extra cars and props into the bg with after effects
Do you even know what you're talking about? Do you think those characters, creatures, giants, elves, orcs, CGI zombies aren't modelled by modellers?
>As an industry, you are done
lmao. We can fire hundreds of people working for the Game of Thrones CGI team then, since that CGI is just gonna make itself right?
>>
>>564217
>lmao. We can fire hundreds of people working for the Game of Thrones CGI team then, since that CGI is just gonna make itself right?
we've clearly been talking about the modelling industry, which is done like i said, not the after effects / nuke industry which although on its last legs, will be on its way out soon too as software improves
>>
>>564219
What is even your point?
>>
>>564221
i started posting around here, you figure out my point

>>564200
>>564206
>>564207
>>
>>564223
Those posts clearly indicate that you don't know what you're talking about. I gave you another chance to properly explain what your main argument was and you managed to avoid it. I think you should educate yourself on these subjects before posting.
>>
>>564225
my main argument is clear for all to see
>>
>>564226
go away
>>
>>564228
I've been here for over 6 years m8, post all the time, lurk all the time
>>
>>564229
yet you can't just lay out what your point was...
>>
>>564230
>
>>
>>564148
>not already owning a bike
>>
There is difference between medical or historical imagery and artistic modeling.
Say you have a 1000 years old artifact from the time we wuz and you want to preserve it digitally, then point clouds or voxels make sense. From an artistic standpoint you're just applying a fancy video filter.
>>564210
It does matter. Topology influences the continuity of surfaces. Surfaces created from point clouds are not continuos and can not be made reflective. Low-pass filtering and automatic re-meshing don't solve that problem.
>>
>>564237
>It does matter. Topology influences the continuity of surfaces. Surfaces created from point clouds are not continuos and can not be made reflective.
do you realize how retarded you sound? The resulting meshes are full of triangles, sure. That doesnt mean they arent contiguous. And "cannot be made reflective"? You have lost it.
>>
>>564240
I said "continuos" not "contiguous" you big retard. Continuos means surface normals have no sudden jumps. A simple triangle mesh is in general not a continuos surface. It's a mathematical thing you don't have the mental capacity to understand.
>>
>>564241
the normals are averaged and fine. Any tool ready for mass release will give you good normals. This isnt 1990.
>>
>>564242
You still don't understand. Think of a cylinder. It's a continuos surface in one direction but, because of the caps, discontinuos in another. If you generate the normals of a cylinder by averaging the triangles, it'll mostly work but it will fail along the caps.
>>
>>564210
Decimate then zremesh ya noob.
>>
>>564245
Then clone and project detail
>>
>>564243
its up to the team of programmers to solve whatever small problems you can think up. They'll probably use some sort of depth solution.
>>
>>564248
Or just re-make the model from scratch correctly instead of bothering with 3D scanning nonsense.
>>
>>564249
YES
>>
File: flat,800x800,075,f.jpg (18 KB, 248x189)
18 KB
18 KB JPG
>>564249
WETA would like a word
>>
>>564248
You do a bad job and then somebody else will fix it along the line. I like your way of thinking.
>>
>>564254
>hurr durr programmers dont exist in current year
>>
How does this only have 40 million tris? Decimation?
>>
>>564140
just zremesh it bro
>>
>>564241
'continuos' isn't even a word you big retard.
>>
>>564296
continuous
adjective
1. forming an unbroken whole; without interruption.
>>
>>564156
>I recommend you read Digital Modeling William Vaughan.
I can't access it :(
It's been blocked in mu country
>>
>>564139
Useful for static rendering maybe, but useless if the mesh has to move or be rigged.
Now if someone finally creates a functional remesher with decent topology out, then it would be a different story.
>>
>>564148
That is actually illegal. You don't own the right to use the bike in your game.
>>
>>564433
No it's not.
>>
>>564453
no no he's right. Its like if you had a pack of M&M's and you took photos and scanned it in using OPs method. You could never use this in game unless you paid the company behind M&M
>>
>>564454
That's a bad example. If you take a picture of something, wether by using a normal camera or an expensive 3D scanning system, the picture if yours. However, say you are making a 3D pgay porno and you make the virtual actor perform acts against nature which involve M&Ms, the company may sue you for slander, not for copyright.
>>
>>564454
What is getting rid of logos?
>>
>>564456
You can take a picture and keep it in your hard drive, you just can't put it in your game without paying. For example you couldn't take a picture of Kanye West, do ops method, and put him in your game. You would get sued until your company was out of business forever.

>>564457
You can't deliberately rip something off like this. For example if you scanned something iconic like hd800 headphones and changed the logo. This is illegal copyright infringements and again you would be sued and your company would lose everything
>>
>>564459
You're wrong.
>>
>>564460
I'm not and in fact I used to work for a lawyer.
>>
>>564461
Maybe that's why you don't work for a lawyer anymore. If you really worked for a lawyer you should know that giving people legal advice without being a lawyer is a bad idea.
>>
>>564459
>Covering, removing and replacing logos is done all the time with real life movie props but 3D scanning is where the line is drawn.
Quit being retarded.
>>
>>564463
No need to get jealous mate. I know the law. I have a degree from a real school, not your artsy fartsy gnomon.

>>564464
You're describing copyright infringement 101. Hire actual designers to design from the ground up, not literally photograph and blur out in PS.
>>
>>564465
>I know the law.
I am the law.
>>
>>564466
t. millennial
>>
>>564467
if you tried to insult him it didn't work. he could be 37 years and still be a millennial kek
>>
>>564465
Just change the logos, the meshproportions and texture a bit and you're good to go.
Everyone uses real life objects as base for their models, and even big studios get away with blatantly stealing shit by just changing it a bit.
>>
>>564433
>This is illegal copyright infringements
>I used to work for a lawyer.
>I have a degree from a real school

as if your constant confusion between criminal and civil liability wasn't proof enough that you're full of shit, you're also wrong about scanning objects. the 3d shape of utilitarian objects is NOT eligible for copyright in the united states. the producer of a mug does not have copyright for the shape of the mug, although he might have copyright on the logo on the mug.

so you would be ok scanning everyday household objects as long as you remove the logos. as for vehicles like a bike, the shape is, once again, not eligible for copyright, but it might be protected by a design patent. by removing the logos and altering the design enough to not be identical to the patent you are pretty much ensuring you will be fine. notice how even giant video games like grand theft auto are full of vehicles that are recognizable real world products and yet by slightly changing the name and shape they can get away with having a "faggio" scooter that is blatantly a piaggio vespa with slightly different proportions.
>>
>>564499
wel not because its copying the bike. but because copying the bike makes it easier to sue you for design infringement.
>>
>>564499
you copy the bike with your photos, you get sued. Plain and simple. You infringened on their work and now you claim it as your own because you changed one or two logos. That'll never hold up in court and history has shown over and over that it hasn't. If you want to argue semantics take it to reddit.
>>
>>564514
sure, fuck every movie and show that features piles of commercial products with logos obscured, fuck every gta game having barely altered real-world cars, fuck every fps that has real guns (with textures literally sourced from photos of the real guns) but with fake names.

in other word, fuck the real world and whatever happens in it, what really matters is your vague instinct about how you think the law should work. so what if you don't know what "copyright" even is, so what if you can't even spell "infringed": you are the expert here because you were once a janitor in a law firm before they fired you for masturbating in a closet.
>>
>>564514
Nigger, you could make an iphone, the edges slightly rounder, the back flatter, and call it iPonee without getting sued. There are thousands of 3Dfags who do this. Changing a few measurements and the logo is all you need to do. You can always do some additional edits after 3d scanning too. Even if companies could sue shitload of 3Dfags for this, why would they? Not only would they waste millions of dollars for cases with a very uncertain outcome, 3Dfags are even indirectly advertising their products for free.
>>
>>564520
>>564522
You two are living in a fantasy world where intellectual property laws, copyright and patents don't exist. Neck yourselves.
>>
>>564559
Bitch I've seen several movies and tv shows with Crown Victorias with the Ford emblems removed. Explain.
>>
Bitch copyright is not the reason why they obscure brand names.
>>
>>564559
please explain how countless video game companies have used almost exact copies of real cars and guns in their games with only the names changed. why does every other shooter have a gun that looks almost precisely like the ithaca 37 but it's called "pump action shotgun"? why is the vespa in all gta games since vice city but called the faggio? in fact, why is gta, one of the best selling game franchises in history, filled with slightly altered but clearly recognizable versions of real products with a sex joke as a name?

or am i living in a fantasy world and in reality rockstar games has been obliterated by lawsuits a decade ago? are there even video games in the world you come from?
>>
File: 1468621107919.png (187 KB, 409x409)
187 KB
187 KB PNG
>>564139
>3Dscan fags think they are relevant in the industry
How are those 6bil poly meshes working out for you? :^) That poly reduction working out fine on those organics, amirite? :^)
>>
>>564618
>or am i living in a fantasy world and in reality rockstar games has been obliterated by lawsuits a decade ago?

no, rockstar pays money for almost everything you see in the game.
but for shit like a toyota car design, they can't be sued because its a popular design (much like fender stratocaster is a popular guitar design).
there is a certain limit on what you can sue people for
>>
File: sexyCyborgScan.png (135 KB, 1366x768)
135 KB
135 KB PNG
>>564174

not a bike, but some chick. still, this is the kind of topology you can expect
>>
>>564140
just remesh and re-topo by hand
i think this might be the way of the future for extra-realism in the future, they already use it for character models
>>
>>564139
>>564142
>>564153
>>564154
>>564237
>>565129
>>565678

> muh topology
> b-but the mesh holes

It takes only a few hours to do cleaning and retopo after the scan is finished. The result is photorealistic geometry (either high poly, or low poly + displacement) plus a basic diffuse texture.
>>
>>565695
have fun retopoing a posed scan. The only thing it's going to be good for is rendering, which you can do with the high poly scan anyway.
>>
>>565678
Well, a human is easier, actually. Less separate meshes.
>>
>>565695
whats the point of being a 3d artist if you're just a monkey that weld verts all day
>>
>>566960
i mean if there are situations where you can use photogrammetry to achieve results in faster time than it would take to model then go for it. there are obviously limitations, since you need to photograph something IRL. it's also an opportunity for traditional sculptures to import their work into a 3D program.

no reason to be scared of new technology, you'll just end up being left behind.

https://www.youtube.com/watch?v=7R1glUbnYEo

not bad use of photogrammetry from a group of students.




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.