[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] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/3/ - 3DCG


Thread archived.
You cannot reply anymore.



File: 1541539863226.jpg (10 KB, 305x301)
10 KB
10 KB JPG
I think it's an incredible elegant way of creating hard surface objects without having to care about fucking subdivision shit and polygon fuckery.
If you need it for animation, you can still retopo later. It's a bit like sculpting for hard surface so I wonder why they cream over sculpting but hate NURBS.
>>
I don't know man, I use solid modeling followed by retopo for hard surface. It gives me freedom from the technicalities of polymodeling, just as sculpting gives freedom to character artists.

To be fair I'm not very good at direct subdiv modeling. Maybe an expert in it wouldn't find it more cumbersome to do exploratory modeling that way.
>>
at this point I'm sure not even 20% of /3/ know what nurbs are
>>
>>707081
Only retards cream over hard surface sculpting, it's always painfully obvious when a machine part has been sculpted. NURBS are superior to even traditional subdivision modeling for most hard surface applications, but they are big brain only so not many people know how to use them.
>>
>>707081
They take time an skill to master, but there is nothing better in hard surface modelling if you aim for quality and precision. They are also pretty much the only thing allowed in modern industry prototyping. And let's not forget the automotive and aerospace industries where NURBS are a requirement.

For game-quality assets it's of course total overkill.
>>
N.U.R.B.S. aren't for you regular /3/ peabrain
>>
>>707755
>Only retards cream over hard surface sculpting
I can't understand how it became so popular. Then somebody comes in doing hard-surface with CAD tools (Fusion 360) and everybody loses their shit for a couple weeks, only to forget about it and go back to sculpting again.
>>
>>707763
too late, already have a NURBS gf
>>
>>707081
It's mostly for engineering and manufacturing (aka the boring stuff), not artistic and animation work.
>>
File: 14468578775.gif (1.82 MB, 241x265)
1.82 MB
1.82 MB GIF
>>708154
>It's mostly for engineering and manufacturing (aka the boring stuff)
>(aka the boring stuff)
>>
>>708154
>artistic and animation work
(aka the 99% useless stuff)
>>
>>707081
It's not that bad for rendered animations. Pretty bad for games, because you tend to have too many polygons after converting.
Cleaning up the topology takes more time than NURBS workflow saves during modeling.
>>
File: 1568425043655.jpg (136 KB, 592x726)
136 KB
136 KB JPG
>>708180
Fuck game content. This topic is for grownups.
>>
MOI should go FOSS
>>
>>708191
THIS
>>
>>708180
There's basically no point in using NURBS for animation because subdivision surfaces became standard. Subdivision surfaces are like NURBS but better because they support arbitrary topology. Verts at lowest subdiv level are like control points, the limit surface is like the NURBS surface.
>>
>>708317
my thoughts
>>
What are some cool things created with nurbs?

No cars/memeoplanes
>>
>>711266
I think some of the early 3D animations were done with NURBS instead of triangles/polygons
>>
>>711266
Polylets GTFO
>>
>>711266
When I was in art school there were several Koreans doing their MFA's making characters in NURBS, highly detailed, beautiful models, and very light on computer resources. Extremely difficult to do though. As we all know eventually Zbrush won out as it was more accessible.
>>
>>711301
You can't use Zbrush in architectural or manufacturing work. NURBS have real world applications, the other stuff are toys for entertainment.
>>
>>711322
Daily reminder that "real life applications" are what makes life livable, but "entertainment" is what makes life worth living.
>>
>>711322
>those are just for gay BABBY toys, not REAL men's work
Why do you feel the need to wank over your choice of technical application.
>>
File: 1568574496797.jpg (9 KB, 236x236)
9 KB
9 KB JPG
What exactly would nurbs be more efficient for? Hard surface concepting?
>>
>>711322
and now complain about the sculpt of some designer you had to rebuild in solidworks.
>>
>>714037
If you have a simple form that needs to be precise use it. Think of it as vector graphics in 3D.
>>
>>707081
I never understood it either...
the ideal method is define shapes using nurbs and then choose your subdivision accuracy for a usable model

>>707756
>For game-quality assets it's of course total overkill.
how?
lets say you make a racing game, you define the car geometry using nurbs, now you can output that car using any poly count you want
imagine the gran turismo guys redoing every car in every successive game with more detail, when they could just define the car once and choose what detail level to export with
>>
>>714044
That makes sense, thanks
>>
>>714046
if you machine a part irl the machine can follow the path more precise than a vert.
>>
>>714046
I'm 100% sure that machine learning will do the retopo in like 1 second and maybe hardsurface modeling will go back to some sort of NURBS modeling again.
Sculpting for organic stuff, NURBS for hard surface and AI for the monkey chores.
>>
>>714060
*the result of machine learning
>>
nurbing is the thinking man's modeling
>>
>>714044
This is important to understand.

Subdivision surfaces are approximations to a "limit surface" obtained by an infinite process of subdividing your model. In practice the render engine will only subdivide maybe 6-7 times maximum, so there's an approximation error. It's also computationally expensive though that is less of a problem these days.

NURBS are closed form, exact surfaces, defined by the control points, like vector graphics.

The approximation error in subdivision doesn't matter when rendering 3d animations but it does matter for CAD and other applications that require infinite/arbitrary precision.
>>
>>707764
Because "hard surface sculpting" is easy so even the retards who need college tutoring for something like 3D modeling can do it.
Even worse, to them sculpted models and subdivision/NURBS look identical.
>>
>>707081
NURBS is the best method to make cars. The smooth curves you get from nurbs can never be matched with a poly model.
>>
>>717167
not to mention that nurbs are a fucking god-tier tool when it comes to rigging
>>
>>707726
well i guess i know and it feels like there are 4 more people on this board
>>
File: 1572150176933.gif (3.89 MB, 200x200)
3.89 MB
3.89 MB GIF
>>717404
>>
>>717499
yea well he's right you know
nurbs make into 10/10 controllers, any female animator would wet herself the moment instant you present those
>>
I see NURBS will be the new (old) 3d hipster meme of the 2020s.
>>
>>717640
Were they a meme in the first place? I remember starting 3D when no professional was using polygonal modelers, only NURBS stuff. Starting from the early 90s more and more subdivision surface solutions started to appear in professional tools, first spline patches, later the much more comfortable cage method, but NURBS never went away for all the high-end design and manufacturing stuff.
>>
>>717167
I'm pretty sure the limit surface can be solved for analytically these days, and can even be formulated as a linear space to transform into/out of for interpolation, making computation a non-issue. At least, I read a recent paper with such a result. I don't know if the larger 3D modeling community has implemented these augmentations yet, but I'd be surprised if they haven't/don't. For example, I know that blender 2.8 now projects all subsurf levels to the limit surface, so internally they may be already using more efficient algorithms that skip the iterative subdivision step.

If I remember correctly (big if), the limit surface is actually gaussian in nature and so mathematically might be cuter than NURBS because of C-infinity continuity. I doubt the CAD industry would ever move to it though, because of non-familiarity. NURBS behaves a little more predictably and intuitively than SDS modeling when it comes to defining manufacturable surfaces, at least with today's technology.
>>
>>718453
disregard this nonsense I had no idea what I was talking about. Catmull-Clark subdivision is C2 continuous at the limit surface and I only thought otherwise because you can solve for it via a matrix exponential.
>>
File: 1580411693166.jpg (119 KB, 398x644)
119 KB
119 KB JPG
nurbs workflow for blender?
>>
>>723033
orthographic
>>
>>723033
There is none. Nurbs in Blender is literally only the most basic implementation. You can make a surface and that's about it.

It's not really the same thing, but "Curves to Mesh" might be an addon to check out.
>>
>>723232
Newfag here, what software should I use instead
>>
>>723259
This already intersects with CAD. I worked with Rhino, Alias, SolidWorks, CATIA, Siemens NX and they all do NURBS pretty well.



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.