I agree with ingomc.
If there's 2 vertices there then you have a 5 sided poly which become pretty ugly when smoothed. review your mesh or cut the Ngon so it become 2 4 sided poly.
Hope it helps.:chinese:
Why are 5 sided polys so bad? Ive read that in countless tutorials. But Im a firm believer that to understand something, is to know it. Just "knowing" something only scratches the surface. So why are five polys so bad when smoothed out? Is the smooth engine based on pairs? Is there a deeper explenation that just "because thats how it is??"
To be honest, I have only few knowledge about what you asked. But what I know is that depending of the subdivision method you choose (NURMS, Quads, Classic), smoothing Ngons can give bad or in few cases acceptable results. Since I don't use a lot Quads & Classic I'll try to talk only about NURMS (the most used). NURMS are slightly like NURBS which means that they use a mathematical approximation :banghead: which results rely greatly on the type of Ngon to be smoothed. Briefly (I don't want to talk any further about the details) you have the best results with quads, & Tris are barely acceptable(in some cases, when workin in editable mesh mode, maybe I'm not sure, because a face is decomposed to Tri elements). It's a matter of mathematics :banghead: .
If you're interested you can search the web about this topics (NURBS & NURMS).
Hope I didn't say something wrong .
Ahh...I see. So its all just realy based on triangles. It seems therefore that most mesh errors occur when the computer is calculating the interaction of the verticies and trying to triangulate them into a solution. This makes perfect sense to me. This is largely the same error that occus when you boolean complex froms together and the results go nuts! Thanks! The concepts are slowly cementing in my brain.
@ Half - See first attachement, try to follow it and see if it helps you.
@ fouedjeljeli & STIGG- Let not get confused..... There are no "rules" on side numbers on polyfaces that's ONLY if you don't use smooth modifier on your geometries. Otherword, It is not illeagal using 5 sided or tri poly, but fact is quad polyface always gives the best result when smoothing, sometimes NGon or Tri polies could give best result too only if they are really small, not "eye-catching" thing. Look at second attachment for examples.
I've never said it's illeagal to use 5 or 3 sided polys. I only said that quads gives better results when smoothed using NURMS output:banghead: .
Being a mechanical engineer, I only know about NURBS but I think that NURMs doesn't differ too much. Quads gives better results because in NURBS approximation a surface or a poly is better defined when with 4 sides or to be honest with 4 vertices which will be considered as "director or control points". In other words a NURBS surface is better approximated and lately "handled" when having 4 vertices or control points.
Hope that I didn't say something wrong .
Gomennasa´ :chinese: (sorry in japanese) for intefering in your WIP with this pure technical discussion .
I absolutely agree with what Equinox showed you in his attachment. I think that the mesh will look better if you do like he said.
Keep up the good work.