Hi Gary,
It's a cool idea. But I have to ask, what about the >problems with Nurbs like 4 sidedness, continuity, seams, >and all that nonsense that Sub doesn't have to deal with?
Well, the way I've thought about it is this: - You need to plan ahead similar to the way you would with a low-poly object to a subd surface, make the original low-poly object as low as possible, then run the 'Conversion to NURBS' script, it takes each edge seperately and creates a curve from each edge, that's a lot of curves, but at the same time useful, then the script joins each curve at exactly the point where the vertice would be in polygon mode, then the script would fill in the area inbetween with a NURBS surface, don't know which way yet, whether it's lofted, or a simple NURBS plane is put in with whatever values YOU say you want (U and V iso's), then the 'script' would stitch or attach or Global Stitch every Nurbs Surface together, thus creating the resultant NURBS object from the low-polygon version.
I'm not saying Sub is the answer to everything or that it >can't be done with Nurbs, just that I remember all those >Nurbs issues being a major pain.
Yeah, the beauty is you don't really need to think about them as your working the polygon model to shape what you want, the thing is though, the resultant NURBS surface is a hell of lot nicer to work with for animation than subds are, and there would be no difference in look to a low-poly model converted to subds and 'this' scripted output, all in all you end up with a seamless mesh, but NURBS instead.
Also, I know depending on the situation, SubD can be slow >for some things. Not going into why and so on, just... >well isn't Nurbs slow too sometimes. I mean trims and >continuity calculations, fillets, etc. Is it really that >much faster to have the same model in Nurbs rather then >SubD?
Yes, because the initial model is low-polygon seamless model, the script is simply converting 'that', so fillets, blends and trims don't even figure into it, all the result should be, should be a SEAMLESS Nurbs mesh. Remember you can up the U and V iso's at 'script' time to get what you want, or do it thru tesselation at render time. The only downside I can see so far is 'edits', subd's has hiearchy mode for that, but is useless once you bind it to a skeleton, whereas my way of doing it would require edits to the mesh at the final stage, I think the best bet would be to not edit it unless it's really needed, and do all the modelling at the low-poly stage. Still u can still edit after the script has been run, there's nothing stopping that, it's just inserting an isoparm across the belly of a character would go right round it, whereas rather you could have done some localised detail at the low-poly stage instead. U could also retain the low-poly model and copy to new layer if u wish thru the script to use for animation purposes too.
But like I said, cool ideas. I've done a little MEL and >I'd say what you are talking about is possible, though not >something to attempt without really knowing MEL. Just at a >design level, unless the whole model is a 4 sided patch ?>that is made of 4 sided patches, aren't you gonna run into >some big problems?
Probably in some areas, which is why I would recommend people who would use it to model using quads wherever possible, and possibly Tri's may be implemented.
Also, about SUBNURBS.... sounds kinda funny. Isn't SubD >already a combo of poly and Nurbs? I have a pretty good >understanding of how SubD works and I don't think we want >it anymore Nurbs like, though faster is always good. But
well, yes it is, but from what I meant with SUBURBS is an actual NURBS type model (so speed is an issue here), but one in which Polygon type tools could be used, like grabbing an iso and extruding from that like extrude edge, or grabbing a single patch and being able to bevel. All just ideas running around in my head atm.
then you have to expect a surface that can do all subd can >and works how it does to be a little slow, especially in >it's first release in Maya.
Yes, but it shouldn't have been that way, as the subd technology has been around for about 20 or so years now, you'd think they'd have gotten it right with so much 20 years research material on the subject!
Anywayz, I 'really' need to learn MEL now to see if I can bring this 'thing' to life! 8)
BTW, for experienced MELlers, try it out for me, let me know what sort of stuff u come up with. Hmm, this could be very interesting, especially when there was a TABOO that you 'couldn't' do this with polys. 8)
