Shelling meshes wishes

Here’s an example of shelling : Rhino VS Meshmixer.

Heck ! Rhino even added a goatie to Grendizer (Goldorak).
Sacrilege !

Re-meshing done right…
This is what Daniel Piker’s “MeshMachine” should be if it was made robust :
A way to convert an ugly mesh into a mesh with “almost” equilateral faces with adaptive edge lengths depending on curvature.

Rhino provides a ton of ways to collapse edges based on a quintillion criteria… what good does that do ?
I just want a reliable tool that turns a crappy mesh into a good one.
Sadly, the only one I found is from AutoDesk and is free, mind you : MeshMixer

MeshMixer “Adaptive remesh” tool :

2 Likes

Here’s for Shrinkwrap.

The initial mesh :

And here’s the shrinkwrapped one :

Still a bad mesh, and in particular, with self-intersecting faces which are a real pain to deal with.

1 Like

Regarding “ReduceMesh”, it’s the following issue : https://mcneel.myjetbrains.com/youtrack/issue/RH-62170

How can this not be corrected yet ?

And here’s the best part. Check the comment about “Largish meshes” :

OF COURSE we are going to use this tool for large meshes which take time to process, that’s the whole point of “ReduceMesh” ##### !!!

3 Likes

Can you please share the initial mesh of this statue you tried to shrinkwrap?

Hi Martin, I’m afraid not.
The issue is really obvious with “OffsetMesh”, when asking for a distance that makes the offset mesh self-intersect.

But Scott Davidson pointed out that ShrinkWrap can be used for shelling, and it seems to do the trick.

I don’t like the “XYZ cubes” patterns that the voxel-based approach of Shrinkwrap create, but at least it can now be said that Mesh shelling in Rhino is a thing.

Shrinkwrap to create a smoothed shell works for me. You can also shrinkwrap and then TriRemesh…

I know Martin, but :
-TriRemesh often fails
-It is not adaptive

I don’t know if you ever used Daniel’s “MeshMachine”.
The concept was excellent, but sadly, he never made a robust tool out of it.
Currently, I need MeshMixer to produce good meshes that are both leaning towards equilateral triangles AND denser in curved regions.

Yes I’ve used mesh machine.

Have you tried remesh by color?

have you watched these videos?

2 Likes

my suggestion for workflow-

mesh you exterior however you want to achieve the result you want.

then shrinkwrap the model at a lower poly count (use a higher optimization setting for more decimation)

then offset inwards with this shrinkwrapped part.

do a mesh boolean with this new part to create your shell.

if you need a closed part, you may need to make a small post off the inside part in order to create the shell, you can then repair the holes as needed.

If the goal is printing, you may want to leave a larger hole in the part somewhere to drain uncured resin, or powder if using sla or sls.

if you are printing PLA you can make the hole very small and just fill it with a dab of CA glue.

1 Like

Very nive videos Kyle !

No. It sounds fun though :slight_smile:
Any reference, documentation ?

Here’s an example of the GH “Triremesh” tool giving an un-reliable result :

What is un-nerving is that, if I feed any of the original surfaces individually, it will produce a proper mesh.
This kind of erratic behavior makes it discouraging to build more complex tools with such components.

Definition sent to the devs, I can’t share this geometry.

I hadn’t received this file until just this moment.
It looks like there’s maybe an issue with a hardcoded tolerance value - I’ll look into it.
It works fine with the input geometry (and edge length to keep the same proportion) scaled down by 0.01.


image

Also, the behaviour of the component for individual surfaces is the same as when feeding them as a list - it solves each of them separately.

Will we get adaptive Tri-remesh at some point ?