SubD for Hollows

I saw @theoutside 's latest vid where he does some SubD shelling. This is a time-consuming thing in jewelry.

But I’m failing to see the benefit of SubD. Like any complex surface that would be a PITA to clean up with OffsetSrf is a PITA to clean up with OffsetSubD. Here’s 1/4 of a ring head that was difficult to hollow:

Forgetting my other design constraints (side wall min thickness is 1mm, top wall min = 1.25mm), I tried creating a minimally complex SubD obj to get a 1mm offset.

At least with OffsetSrf, I had a base to use for projecting curves to create the inner srf by hand, but I can’t see how this has any use.

OffsetSrf vs SubD.3dm (436.9 KB)

SubD has benefit as a tool only if one finds its tools more effective towards a specific application, over other tools at one’s disposal.

Ok, that’s an oversimplification of course, but the point is that one needs to have experience with a tool (any tool), leading to knowledge of where best to employ it - or alternates. In your work, such may be: never, occasionally, often - based on both your specific discipline and experience.

In my specific experience, SubD is used a minority of the time, but when it is used, it is a VERY useful toolset.

1 Like

Curious - have you tried creating the ring as a solid SubD mass, convert to NURBS, and then Shell it to create the hollow form? Success with that route may hinge on how well the Shell command deals with the resultant polysurface…

A key topic in Kyle’s video was about how much better shelling is with SubD than srfs. If that’s true, I would like to know how because this is a major pain point in jewelry.

I am a long-time TSplines user. Creating a ring as a single body is an incredible waste of time. Having to artificially adjust the subdivision levels between 100’s of components (shank, head, prongs, seats, galleries, etc) is time-consuming and leaves you with something wholly uneditable.

It’s the same with jewelry. Very limited usefulness in which it is very useful.

I’m not a jeweler, but can imagine the precision required, and that NURBS modeling is well suited. It is my hope that, as things develop, McNeel finds novel ways to bring some elements of precision, typical with NURBS, to their flavor of SubD. Such would move-the-needle.

As you know, SubD is just a tool amongst others. Perhaps you’ll find utility with the Rhino flavor in time.

1 Like

Has anyone tried meshmixer? If you import an an stl file it will give the ability to create an interior wall and delete the solid portion inside of it. It’s a free download although you may need to watch some YouTube videos to get a good feel for it.
Download at:

here are my thoughts on it using 100% rhino

  • from left to right-

original in blue,

original meshed- red
offset mesh at .25mm- Green

Green quadmeshed and mesh offset again at .25mm, then quadremeshed again. -orange

orange converted to subd and cleaned up a bit (delete edges that were causing issues and creased corners, snapped edges flat to match ring quarters-

convert to nurbs and cap- blue

However, with stuff like this the absolute best way I’ve found to to do it is with a voxel program like freeform or 3dcoat. Because voxels do not self intersect you get perfect offsets and can use thier smoothing tools to knock off any spikes or problems.

Once you get your voxel offset, you can quadremesh it, convert to nurbs and then join it in (or boolean out) for a nurbs wall thickness offset.

Not a 100% rhino work flow per se, but it is bulletproof, I’ve used it hundred of times for very very complicated parts.

I agree, I like this GH add-on which works in Rhino for Windows.

All those steps feel more complicated than using _OffsetSrf or _CageEdit to get a base for doing surface modeling.

I also have 3DCoat, but rarely am able to take advantage of voxel offsets. It produces a clean hollow, but I need a _Shell with an open side:

A clean SubD hollow would be easy to edit into a shell, but a multistep workflow involving voxels and meshing conversions sounds entirely unappealing.

I’ll keep watching the SubD space for things that might be useful to my industry, but so far it’s irrelevant.
And it’s frustrating to see 99% of dev going into these features, while decades-old bugs that would give a 10x boost to efficiency are thrown on the maybe someday pile.

1 Like