What exactly is CurvatureAnalysis telling me

Today I tried to improve my Nurbs modelling skills and e.g. made a Y branch that looked okay to me. Now CurvatureAnalysis Gaussian style looks :slight_smile: while Mean is more :cry:.

I read the manual, btw, but it is, well - can someone perhaps explain in other words whether my Y branch is more ‘good to go’ or ‘that doesn’t work when printed using metal’.

What are you basing your analysis of Guassian and mean curvature on? Are you assuming green is good and red is bad?[quote=“MarcusStrube, post:1, topic:39824”]
can someone perhaps explain in other words whether my Y branch is more ‘good to go’ or ‘that doesn’t work when printed using metal’.
[/quote]
Surface curvature has little to do with whether a part is suitable for printing using metal or any other material.

I understand that the assumption green is good, red is bad is not working.

Maybe I should have asked the question in a different way: How can I use the analysis tools to figure out whether a Polysurface is really a properly modeled Polysurface? So, everything really as expected/wanted at the edges of the surfaces that make it.

Hi Marcus,

Mean and Gaussian analysis tools are rather good for exploring flat, simple and double curved areas of a surface, as well as violent changes in curvature or whether the curvature is saddle like or bowl like.

If you want to know about the continuity of a surface (eg: how smoothly objects would reflect on it if it were built in a reflective material) than use Zebra or Emap.

If you want to know if a polysurface is closed (eg: a solid) run the _SelClosedPolySrf command and see if it gets selected or the _ShowEdges command and make sure you don’t have any naked edges.

Okay, that’s nice. I at least know all commands you are suggesting.

The point is that _SelClosedPolySrf and _ShowEdges give a real information: My surface is either closed or it has naked edges. It’s true or false.

Zebra and Emap however are helping me to make that decision on my own. And I often find it difficult (using Zebra and Emap) to figure out whether a Polysurface is bad, merely acceptable, quite okay, good, or even very good.

But, okay, I get it, I just thought there is perhaps also a true/false tool for the ‘would objects really reflect very smoothly on this polysurface if it were built in a reflective material’ question…

Thank you, Vanessa.

What do you mean by bad or merely acceptable? Is it to 3D print (eg: is it printable)? Is it to see if it’s smooth and continuous? Or if it’s a bad object and needs geometry repair?

Cheers!

Hi Marcus - one problem is that unlike two curves that meet or don’t meet at a particular continuity, (GCon), surfaces meet, or not, over some distance in space -along the length of the edges- so there is no yes-or-no about their continuity - it may be perfectly G2 here and show a crease there. Zebra allows you to see how that is all the way along - it is not perfect - there is no numerical info, for example, and it is highly dependent on the quality of the analysis mesh. But if you set it up with a very fine mesh and stripe width and direction as appropriate, it should give you not only good information about the continuity but also give you an idea of how the surface flows - generally you want to see very consistent and non-wobbly stripes on the surface - that can be as important as the local continuity at the seam.

EMap, with Fluorescent Tube as the map, and again a good super fine anlaysis mesh, is a good one for getting a sense of how reflections will flow.

-Pascal

Yes, that looks like something where I trust my own eyes… Thank you, Pascal.

But generally it would be possible to extract the necessary Isocurves of the left surface so that I then have a bunch of ‘curves that meet’ on both sides along the edge and Rhino doesn’t do that, because that approach is taking to long for complex polysurfaces? I mean, more or less?

Hi Marcus - that might be workable in cases where the isocurve direction is exactly relevant to the curvature of the surfaces along a seam, but that is generally not the case.

-Pascal

Ah, okay, now I get it. Thank you.