connectPolyShape vs. Actual Maya SubD's

[awnold]

Well for one, mayas sub-ds are hierarchical, I dont think there is one studio which uses hierarchical sub-d. The main reason being you cant export it to manipulate or render somewhere else.

Now on the other hand, cps is just a fancy way to model with polygons. Polygons are just about as versatile as you can get, they can be exported and manipulated in every program on the planet. Plus texture mapping them is comparatively easy.

So, I think if mayas sub-ds where non hierarchical they would be the tool of choice, but as of now cps is a great alternative.

[-wT-]

Originally posted by wgeddes
someone told me that they had used both Maya Unlimited SubD's and connectPolyShape polyproxy simulation, and that they preferred CPS!!! I found this a bit hard to swallow, so I decided to ask what you all think. Gracias in advance for your input!

Accidentaly pressed Post New Thread instead of Post reply?
Who's "they"?

[bigfatMELon]

I know of many instances where CPS was chosen over the real subs. CPS is offers a lot more than just smoothing, like multi-part management and smart symmetry handling.

I know that the folks at Big Idea at one point were using both; modeling in CPS and then funneling that through the real subs.

-jl

[wedge]

"they" is one of my teachers, why?

[svenip]

i think we shouldn't loose the fact that subds are build of less points and will always be smooth. in a big scene this really matters for productions. haven't worked with cps yet in depth.
so one question : is it possible to use the proxy as a kind wrap deformer ? just like i can do this in subds.

[bigfatMELon]

i think we shouldn't loose the fact that subds are build of less points and will always be smooth. in a big scene this really matters for productions.

Well, sure... and no at the same time.:) In a big scene it might not matter at all because of heavy use of non-deformable stand-ins and other old school tricks that might be employed. There are advantages to both but it all comes down to individual needs. There is no "One Modeling Method Uber Alles."

As of 4.5, I think the real subs will finally have a more clear advantage in being able to convert to patches. And I still see myself in many cases getting work done with CPS or even a mixture of the two.

so one question : is it possible to use the proxy as a kind wrap deformer ? just like i can do this in subds.

If you mean binding/deforming the cage and such, then yes. The two technologies are practically the same thing.

-jl

[underdog]

Well technically CPS isn't subD, just a proxy smoothed object. Its a very nice set of tools, allowing you to see how your smoothed mesh will look in real time. Along with easily incresing decreasing the divisions, keeps the center vertices aligned, some other things I forget. Along with be an actual poly object the entire time, you have every poly tool available.

Maya's subD's are very powerful. You also get some nice features, creasing, hiearchy edits. And since you can go between the subD and proxy object you can still use all the poly tools, though it does get slow if your not careful. I prefer these to CPS, I like the workflow, and it just seems cleaner to work with. Hierachy subD is very powerful(just look at stahlburgs work)

[-wT-]

Originally posted by wgeddes
"they" is one of my teachers, why?

Just curious, you just sounded like meaning something big like ILM or Pixar or something :rolleyes:

[bigfatMELon]

Well technically CPS isn't subD, just a proxy smoothed object.
Not to quibble but...

Technically, that's exactly what it is: Catmull-Clark applied to polygons. It may not be an explicit output surface rather than an implied surface type. But it's still a kind of subD.

-jl

[underdog]

OK, true, point taken. You know what I meant.

[wrend]

wondering, in terms of transportability of models between applications.

will a poly cage smoothed w the same algo produce the same results in whatever package? as in ..if i model a cage in maya, and take that cage to xsi and Catmull-Clark it, will it not produce the same smoothed surface? same for PRMan too? This is with all modelling done on the cage.

what im scared the answer might be is that maya's subDs tend to propogate to higher levels in their hierachial subdivision scheme to capture curvature that might be implied by the cage? - common experience shows that certain cages will produce level 1 and 2 pnts on the subDed surface (levels you didnt call/refine for). thus it resolves surfaces which will be distinct from 'standard' Catmall-Clark subdivision. hey, do AW even use Catmull-Clark? or is it just elements of, i mean how do they sort this hierchial buisness in with it all?

I guess the question is: is maya's subdivision scheme so diviergent from Catmull-Clark that it is not safe to expect that base level modelling w maya's subDs (so all in the polycage) will produce the same surface if that cage is ported for smoothing elsewhere by Catmull-Clark algo.?

and just quickly, is standard poly subdivision (ie what CPS uses) done w catmull-clark? - would it be a safer bet for similar results?

cheers all.

[Waboflex]

Yes, good question, I've been looking into this recently. And to cut a long story short, there doesn't seem to be any significant difference using different smoothing algorithms. If you know where to look (typically areas where surface curvature changes a lot in a small space), you might be able to see a few (very) slight differences, but unless you're doing ID work it's unlikely to be of any concern.

I know what you mean about Maya adding extra level 2 points without you asking, but this just seems to be how they were reworked in Maya 4. In Maya 3 it was different, you only got level 2 points and beyond when you asked for them, but the end result is much the same however it's done.

So I reckon that, as long as the model's built well (all quads if possible), you should be able to take that cage into any package and convert to subd's / smoothed version, and you'll get pretty much the same resulting surface in any of them. However mapping them's probably a whole different kettle of fish...

BTW, I've used both CPS and subd's a fair bit, each has it's pros and cons, but in my experience I found the real subd's to be a lot faster as the model gets more complex. Just as long as vertex editing is done in hierarchy mode, and convert back and forth to poly's if topology edits are needed.

End of waffle.

Wabo.

[OrestesMantra]

You people use to much technical mumbo-jumble:eek:

lol, its a lost cause for me.I just accept that the tools are there and that they work.Nothing more,nothing less:beer:

[GrafOrlok]

I love both CPS and Maya SubD's:thumbsup:! To me CPS is a great poly modeling tool, and Maya SubD's are a great detailing (and rendering) tool. From what I know Maya treats SubD's a bit different than other packages do. Usually ordinary SubD's work much like CPS, only with a highly tesselated and smoothly rendered result.

In Maya, as pointed out earlier, you have a hierarchy which makes it perfect to go into detail at certain areas. You have 13 levels (I think) to refine one polygon, but you seldom need more than 4... I did a Dino for a commercial where I started out in CPS creating the basic shape and doing some minor detailing. After doing all UV's and stuff I then used the lo-poly to convert to SubD's and there I detailed the muscles, creases, folds and other skin details. Then we used the lo-poly to drive the SubD with a wrap deformer.

As to use CPS as a wrap deformer: It works fine to bind the proxy to the skeleton and use it to drive the resulting geometry. You might not get the precision you want (however same as with a wrap deformer), but it works.

[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.