Is SubD Inset working correctly? Unexpected result

Hi,
I have a question regarding working with SubD. Quite often, when I use a command, it doesn’t behave the way I expect :frowning: – and I always have to find some workaround. Let me start with the Inset command.

I’ve prepared a model that clearly shows the issue. When I run Inset on an object that is convex, it still creates new vertices outside the object in some areas – and I really don’t understand why.
Yes, I can solve it by extracting the “control polygon”, offsetting it as a curve, and then using SubD loft. But I use this workflow quite a lot, and this “offset workaround” isn’t ideal for me.

Could you please advise if I’m doing something wrong, or if there’s a way to make Inset behave correctly? Or is this more of a “developer issue”? **
nom_inset_v01.3dm (88.3 KB)
**

It doesn’t look right in concave corners

Hi, great observation! I didn’t realize that the command calculates based on “concave corners” rather than the “area center.” To me, that seems more like a bug than a feature.

likely

i think you expect to much from a complex single-Face-non-quad SubD.
To Nurbs shows the underlaying triangulation:

the “error” (which is more limitation for n-gons) comes from this star-like underlaying surface.

my workflow to have a nice result:

gumballextrude the surface with a random value
_extrudeSubD → Direction N
to get the desired faceloop that corresponds to the expected Inset
now manually build the fill with multiple faces / quad only
i use:
_bridge _3dFace
delete the initial extruded Faceloop.

a nicer set up for the initial polygon / shape would give a more structured / nicer result.
for example use 8 Faces for 360 degree, 4 Faces for 180 Degree, 2 Faces for 90 Degree…
in your case 4 * 180 Degree = 16 Faces

kind regads - tom

nom_inset_v02_tp.3dm (3.0 MB)

2 Likes

Hi, thanks for taking the time to look into it. Sorry, I probably didn’t explain it clearly enough – it’s not about solving that specific surface, it’s just a model case. What I really mean is that I’d like the command to behave the way it should from a user’s perspective, without having to come up with some kind of “alternative” solution – whether it’s yours, mine, or anyone else’s. It just slows things down during modeling. And honestly, I don’t think I’m asking for anything “extra” from the command when it doesn’t even work properly in a simple case – as Martin pointed out.

This is definitely a bug. Thanks for reporting. I opened RH-87715 to get this fixed.

1 Like

looks like concave subd faces are not handled correct.
i would expect that they are handled correct or at least a failure is reported.
"inset failed - can not handle concave corners.

it has nothing to do with the n-gon as i thought initially.

also
inset does not work for a single subD face !

subD faces with concave corners are not meshed correctly in box mode

thanks @Joshua_Kennedy for having a look at this.
kind regards. -tom

1 Like