I have some issues with shrinkwrapping files which are going to be 3D printed.
In my grasshopper definition I’m using Isopod and then shrinkwrap to lower the mesh density and a bit of smoothing. The mesh output is closed but some of the faces are basically overlapping.
Here’s a detail and the file is below
2026-09-29_shrinkwrap-issue.3dm (2.8 MB)
I’ll try to create a simplified example.
Curious, have you tried sending them to a printer?
A lot of slicing software will simply ignore this type of issue as long as the part is closed.
Yes I had numerous files like this printed successfully.
Does the same file do this in Rhino 8?
Ok here’s a file which replicates the problem.
The problematic mesh faces can be seen best when the mesh is baked and then Concept mode shows sharp creases. I added points on four edges.
The example in the first post has worse overlaps than this:
2026-09-29_shrinkwrap-issue_02.gh (104.0 KB)
2026-09-29_shrinkwrap-issue_02.3dm (1.0 MB)
@Trav or @theoutside what’s the quickest way to find those faces?
Mesh has 1 face where the face normal differs substantially from the vertex normals.
To find them.. i dont think theres an existing command but a script can locate them pretty quickly.
However, the RebuildMeshNormals command should probably fix them without having to hassle with them.
import Rhino, scriptcontext as sc
from Rhino.Geometry import ComponentIndex, ComponentIndexType, Vector3f
rc, ref = Rhino.Input.RhinoGet.GetOneObject("Mesh", False, Rhino.DocObjects.ObjectType.Mesh)
if rc == Rhino.Commands.Result.Success:
obj = ref.Object()
m = ref.Mesh().DuplicateMesh()
m.FaceNormals.ComputeFaceNormals()
sc.doc.Objects.UnselectAll()
hits = []
for i in range(m.Faces.Count):
f = m.Faces[i]
vis = [f.A, f.B, f.C, f.D] if f.IsQuad else [f.A, f.B, f.C]
avg = Vector3f(0, 0, 0)
for vi in vis:
avg += m.Normals[vi]
avg.Unitize()
fn = m.FaceNormals[i]
fn.Unitize()
if avg * fn < 0:
hits.append(i)
obj.SelectSubObject(ComponentIndex(ComponentIndexType.MeshFace, i), True, True)
sc.doc.Views.Redraw()
print("Bad faces: {}".format(hits))Ok so one way to get rid of the narrow faces is by running CollapseMeshFacesByAspectRatio with a certain threshold until no faces hit the threshold.
Thanks @Trav that helps.
And I clauded something to eliminate the slivers in Grasshopper.
using System;
using System.Collections.Generic;
using System.Linq;
using Rhino.Geometry;
public class Script_Instance : GH_ScriptInstance
{
private void RunScript(Mesh M, double MaxAR, int MaxIter, ref object OutMesh, ref object Log)
{
if (M == null) return;
if (MaxAR <= 1.0) MaxAR = 25.0;
if (MaxIter <= 0) MaxIter = 100;
Mesh work = M.DuplicateMesh();
work.Faces.ConvertQuadsToTriangles();
work.Vertices.CombineIdentical(true, true);
work.Vertices.CullUnused();
int startFaces = work.Faces.Count;
int totalCollapsed = 0;
int pass = 0;
for (pass = 0; pass < MaxIter; pass++)
{
int n = work.CollapseFacesByByAspectRatio(MaxAR); // modifies 'work' in place
if (n == 0) break;
totalCollapsed += n;
// tidy up so the next pass sees a clean mesh
work.Faces.CullDegenerateFaces();
work.Vertices.CullUnused();
work.Compact();
}
work.Faces.CullDegenerateFaces();
work.Vertices.CullUnused();
work.Normals.ComputeNormals();
work.Compact();
// Report what is left, using Rhino's own aspect ratio measure
int remaining = 0;
double worst = 0;
for (int i = 0; i < work.Faces.Count; i++)
{
double ar = work.Faces.GetFaceAspectRatio(i);
if (ar > MaxAR) remaining++;
if (ar > worst) worst = ar;
}
OutMesh = work;
Log = string.Format(
"Passes: {0} | Collapsed: {1} | Faces: {2} -> {3} | Faces still > {4}: {5} | Worst AR: {6:0.##}",
pass + 1, totalCollapsed, startFaces, work.Faces.Count, MaxAR, remaining, worst);
}
}
@DanielPiker as far as I understand you developed shrinkwrap.
Could you improve the command so that narrow slivers are eliminated?
Dale and I did shrinkwrap. Are you finding these on all pieces or certain scenarios. We can certainly add a clean up pass if its needed.
Smoothing and optimization seem to add more of those slivers.
I was just looking at this and i can see where I did this clean up process early on in the 8 development and removed it, but I’m hazy on the why I pulled it (I’ve had a lot of Mt. Dew since then). My guess is it was causing other issues in enough cases. We’ll take another look. The problem shows up when using the polygon reduction most certainly.
