Hello! I am watched a tutorial about Subdiv surfaces and the tutorial was not now, around 2008. I have the feeling that polygons took over subdiv completely for the past couple of years since you can smooth view your object by just pressing “3”. I am not 100% aware of 3d modeling so I have actually never used Subdiv surfaces before. Are they still used as general? Or are they forgotten because of the poly improvements?
Thank you!
Sub-D surfaces in Maya still used??
I have literally never used them once in my 5 years working…but I’m sure someone has a use for them…It’s not like you can’t use them, and I’m sure I could benefit from them in random situations, I just never took the time to check them out.
no one is using them anymore and the tools have been removed from the menus in maya 2014, though 2014 still supports them if your scene has them. People still might use them as a means to convert polygons into NURBS surfaces, but that’s about all you hear about them anymore.
Polygons and smooth approximation took over. SubD’s never had all the tools polys had anyway and they were a pain for multiple reasons.
Smooth Approx is the same as SubD surfaces. [subD is polygons]
Make a box and duplicate it… move it to the right.
On the first box do a “mesh” “smooth”,… and division levels at 2.
and on the other box hit “3” on the keyboard and hold down Shift and rclick hold down on it and move down to “Polygon Display”,… move to the right, and down to “Toggle Hidden Triangles”.
And… you can see, its the same mesh.
The Smooth Approximation does the same thing as hitting the 3 key. Or even setting the “3” key for rendertime.
And lastly… we at work are still modeling for subD.
… btw… don’t confuse normals for mesh smoothing… very very different things.
Maybe not(sure if) directly related…but this video made my wonder if NURBS will have an update (soon?).
Anyone any rumor or something?
I’m wondering how in your opinion NURBS should be updated ?
Maya’s NURBS work okay for some things, animation and some modeling aspects, but compared to every other NURBS application around they’re very basic and beginner, and lack many tools and functions we take for granted in Rhino, Solidworks, and others. As a result, I really never use them at all except for animation stuff. Rhino’s are so much more vastly powerful and potent, there’s really no comparison.
I’ve only used Sub-D a handful of time in the past, but prefer Maya’s polygonal toolset by far.
I’m wondering how in your opinion NURBS should be updated ?
Exactly as InfernalDarkness said.
Maya nurbs are not bad, a lot of stuff is possible to do as it is, but compared to other programs like Rhino, Alias and others a lot more stuff can be added also to Maya.
Someone might say…if you want more options then just use one of those apps…that is true but again i think it would be nice if Maya nurbs would get an update…not needed some super robust upgrade but more options would be most welcome in my opinion.
Or maybe something like T-Splines in Maya. There was apparently t-splines plugin for Maya, not sure what happened with that…as if suddenly disappeared into thin air.
Anyone know anything about that plugin?
My guess is that it would be easier to implement Rhino as a plugin for Maya than to overhaul Maya’s NURBS outright. Also, considering how many other applications specialize in this area, it seems silly (for Autodesk) to waste that much time and energy on something which won’t be a game-changer at all. Much like people using Zbrush/Mudbox/whatever for sculpting and Nuke/After Effects/whatever for compositing, Maya really can’t do everything for everyone. And I imagine most people who really need NURBS are already using their favorite tools. I model all my arch/viz stuff with NURBS in Rhino, then just export to .obj for Maya texturing/rendering. Maya is fun but they’d have to implement 90% of Rhino’s tools for me to even consider switching to modeling architecture in Maya - and I still don’t see Maya being as print-friendly as Rhino. Permit plans from Maya? I don’t see any benefit to such a thing! Again, the other NURBS programs are all as superior in their areas too, I’m just using Rhino as an example.
Sorry to derail the topic of Sub-D any further, though. Party on!
Hierarchical Subdivision Surfaces are not polygons, and isn’t the same thing as doing a “Smooth Mesh”, or hitting the 3 key on a poly. They are a completely different type of shape, but basically target the same kind of workflow.
The terminology can be really confusing though since most people call standard Catmull-Clark subdivision “Subd modeling”…but in Maya, Hierarchical Subd modeling is very different, and I think that’s what the original poster was asking about.
http://facultypages.scad.edu/~mkesson/AppNotes/hierarchsubdiv.html
So these aren’t polygons?
from the pages…
"
Compatibility with Maya Subdivision Surfaces
The path specification for Maya subdivision surfaces differs slightly from PRMan's specification. In particular, Maya does not differentiate between extraordinary and regular faces when ordering vertices after the first level of subdivision. The following code fragment using the Maya subdivision surface API illustrates how to compensate for this when converting a hierarchical subdivision path from Maya to PRMan."
So what is it?.. the rendering interpolation? And/or is it just prman?
Saw this thread also,… http://forums.cgsociety.org/showthread.php?t=118938
http://en.wikipedia.org/wiki/Catmull–Clark_subdivision_surface
Pixar - people call it a true curved surface, just like NURBS.
So you can think of it as the third way of representing your geometry along with poly’s and NURBS.
The fact that SubD uses same kind of proxy cage as you would use with polygon proxy modeling usually confuses people but the ability to crease / partially crease edges and vertexes makes it a totally different thing.
Why Pixar and Disney uses it?
Because it’s the most efficient way (atleast from polycount perspective) of presenting arbitrary geometry and their renderer supports it .
The magic comes in the way which renderman compliant renderers handle the geometry.
It’ s a thing called micro-tessellation, so the smooth surface is “done” on rendertime.
That’s what they are doing with OpenSubDiv, the same thing but calculated with GPU.
And if you take a good look at Ptex you’ll see that SubD’s was the actual (and only) reason why it was ever conceived.
Hope this helps.
/risto
Ill try to explain a few things without getting everyone even more confused than we already are
I think everyone knows what a polygonal surface is
it has vertices, two connected verts make up edges, connected set of edges make up faces
Various people (e.g. Ed Catmull and Jim Clark) have come up with sets of rules to modify a polygonal surface in such a way that it gets smoother and smoother, converging towards a perfectly smooth surface. This is what Catmull-Clark subdivision is. A set of rules that you apple to a poly (insert a new vertex here
average out the vertices there
), and you can apply it over and over again and you will approach the smooth surface.
In Maya, selecting a poly and hitting 3, or doing a Mesh -> Smooth, or a Proxy -> Subdiv Proxy, all use the Catmull-Clark set of rules. An important part of this kind of algorithm is that it is iterative
you apply it once and you get a more dense, and smoother, poly
you apply it a second time and you get smoother and more dense
In any of those operations in Maya you can provide the number of subdivision levels, which is how many times you want to apply the set of rules to the poly.
There are other sets of rules that do similar things, e.g. Loop subdivision was created by Charles Loop, but Maya uses Catmull-Clark.
So what are Mayas Hierarchal Subdivision Surfaces (and this will be a pretty high level hand-wavy kind of description)
The end result is basically the same thing: its a way to view/edit a smooth surface using some higher level control cage or representation. The key difference between polys and these things though is that it can decide what the smooth surface is without having to iteratively apply a set of rules until you are close enough to the smooth surface that you are happy. You can give it a control cage (which is just a poly) and it can jump from that to any point on the smooth surface in a single step. Thats the key difference in the approach that Jos Stam took and what I think he has a patent on (although I cant really remember if he actually has a patent on it or not).
Of course graphics cards and software renderers dont understand the Subd method that Maya uses on these things. For us to draw them in the viewport we have to triangulate it into something we can send to the graphics card, and software renderers like Mental Ray need to come up with their own way of approximating what it is you see in the viewport inside their renderer.
Things like the Catmull-Clark method have been around a long time, and people have started to extend them in new ways. E.g. maybe you want to come up with a method of allowing a user to make a set of edits to the poly beginning at a certain level? Then the user could keep the original low res poly and only start adding certain details at a more dense level. Thats not described in the original paper, but its the kind of thing that people have done over the years, and the term hierarchal has started getting mixed in there to describe various approaches.
But
in Maya
when we say Hierarchal Subdivision Surfaces we mean something that really only exists in Maya. Renderers like Mental Ray can provide approximations of them, and users can always model with them and then convert it to a poly, but you just wont ever see anything natively supporting it. E.g. you wont ever see a game engine using the exact same math as in Maya to draw a smooth version of a low res control cage.
Hopefully that clears something up
or maybe just generates a new list of questions
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.