NURBS or Polygons or SubDs?

[Howitzer]

Their use is never consistent. My guess is that they are each used depending on what is being modelled. When modeling scenematic characters or organic objects you should use subdivision surfaces. When modeling game assets, you should not subdivide and just rely on polygons. When modeling organic objects such as cars or ships, you should model with nurbs or patches.

Just a guess. I'm not in the industry. Yet.

[wetworx]

Their use is never consistent. My guess is that they are each used depending on what is being modelled. When modeling scenematic characters or organic objects you should use subdivision surfaces. When modeling game assets, you should not subdivide and just rely on polygons. When modeling organic objects such as cars or ships, you should model with nurbs or patches.
Just a guess. I'm not in the industry. Yet.

each methodology has their respective benefits in a pipeline of course, but I am questioning whether SubDs have become the norm for all (film) character models or whether most studios still use NURBS (or even Polys)?

On a related note, the Alias Hyper-realistic masterclass is taught utilizing a polygon model and the latest Alias Learning Maya books only touch on SubD models (never actually modeling a character just set pieces). So I'm left to wonder whether those resources are representative of where the industry is at/headed.

[vfxdude2]

Well, there's at least *one* major CG feature studio which uses only NURBS, so they're still around. But, it's all in-house software. They have special techniques for stitching and what not.

I dunno.... I think if you're a modeler, you should probably know how to use all the various techniques. Different shows use different methodologies.

In other words, use the right tool for the job...

-vfxdude

[wetworx]

Well, there's at least *one* major CG feature studio which uses only NURBS, so they're still around. But, it's all in-house software. They have special techniques for stitching and what not.

I dunno.... I think if you're a modeler, you should probably know how to use all the various techniques. Different shows use different methodologies.

In other words, use the right tool for the job...

-vfxdude

Thanks for the input, at Siggraph this year I heard mention of NURBS being the favored technique for most film prodution work and Polys for games. Nothing on Subd tho. If anyone from other studios can confirm this, please speak up ;-)

[vfxdude2]

Thanks for the input, at Siggraph this year I heard mention of NURBS being the favored technique for most film prodution work and Polys for games. Nothing on Subd tho. If anyone from other studios can confirm this, please speak up ;-)

Games use polys because they have a predictable complexity and render time. In games, everything has to render in a known, finite amount of time.

In film, we have the luxury of variable render times :-) So, you can use whatever suits your needs.

I'm not a modeler, so I'm not completely informed on the subject, but I believe there are ways of dynamically subdividing polys or subds at rendertime. This is definitely true of NURBS since they represent a continuous surface which isn't tessellated until render time.

The best renderers (e.g. renderman) use a technique knows as micro-polygon rendering. This algorithm works by dynamically subdividing a surface (at rendertime) until each polygon is sub-pixel size in screen space. This produces better renders than traditional anti-aliasing techniques.

NURBS lend themselves easily to this technique, but I think there are probably algorithms for doing it with polys or subds as well.

Anyway... how much of this do you have to think about as a modeler? I don't know; I'm a TD so I think about this sort of stuff all the time ;-)

-vfxdude

[newellteapot]

Sub divison surfaces are not the norm in many studios and nurbs are not used very much in the TV/Film industry, as for my personal experience.

[vfxdude2]

Sub divison surfaces are not the norm in many studios and nurbs are not used very much in the TV/Film industry, as for my personal experience.

Well, not be be contradictory... but as I said before: One of the REALLY BIG studios which makes CG features uses NURBS exclusively. I can't say which one 'cause I signed an NDA.

But my point is simply that if you're intending to pursue a career in modeling, you should definitely be familiar with ALL the methods. Different studios use different techniques. You might be better at one technique than another, but you should at least be capable of picking up the others.

-vfxdude

[bblanar]

I'm still a noob, but...

I have been debating this myself too, I started my current project with subd's then i started remaking stuff in nurbs. Then I thought I could do the whole thing in nurbs, but I cant. Well, I probably can but the stitching just gets real ugly and crazy.(XSI). However, NURBS let me output poly meshes that are MUCH cleaner than just starting with polys and subdividing. I also made things that I thought were finished, smoother than ever and much less polys.

I'm setting up nurb surfaces and polishing their adjustments, then the final stages are outputting the poly meshes and 'capping' open edges or sides that I just cant seam to with stitching. Basically, I create a 'mold' out of nurbs first. Ouput Poly Mesh, Hide the Mold, and Fine Tune.

[newellteapot]

Well, not be be contradictory... but as I said before: One of the REALLY BIG studios which makes CG features uses NURBS exclusively. I can't say which one 'cause I signed an NDA.-vfxdude

yes you are right some studios use nurbs and subdivs, but as far as I know most here in London don't. I think it's sure better, as you said, to know all techniques.
:)

[CGTalk Moderation]

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.