3
$\begingroup$

I have closed poly curve triangle. Curve lines as instances on its points. Endpoints of these lines are offset along triangle normals.

first case - I use curve normal node - you see the normals do not intersect in the triangle incenter

second case - custom constructed normals from point positions intersect well

what is wrong with the curve normal node - why is this happening?

normals

$\endgroup$
5
  • $\begingroup$ Why is what happening? I don't understand the question. Specifically, I don't understand what the first two cases represent, the third and fourth have seemingly the same calculation from normals, though there's also a horizontal line for some reason... $\endgroup$ Commented Sep 1 at 21:17
  • $\begingroup$ there are just two cases, if you mean first two image figures they just describe the adjustement of the triangle before 3rd (1st case) and 4th (2nd case) figure (its zoomed in detail to the two most left points) OR just open the file and use the yellow switch $\endgroup$ Commented Sep 1 at 21:30
  • $\begingroup$ That's really interesting... I can confirm that this behavior is strange, and from my point of view, it actually looks like a bug. In your calculation, the normals are generated exactly as you would expect from the "Normal" node, and I think you've encountered some rounding errors here. I would report that. $\endgroup$ Commented Sep 2 at 8:28
  • 1
    $\begingroup$ ok reported.... $\endgroup$ Commented Sep 2 at 9:01
  • $\begingroup$ @quellenform, reply from the dev and if anyone else is more interested in this LINK $\endgroup$ Commented Sep 4 at 9:07

1 Answer 1

2
$\begingroup$

After experimenting for a while, I suspect you might have different Z's for these points.

If I have the triangle completely flat, the normals behave as expected even in extreme ratios:

Normals on flat triangle

But if I have different Z's for the points, I introduce twists that change the normals off the plane (we don't notice the reason if we're on an orthogonal view)

Normals on z-flat triangle

$\endgroup$
2
  • $\begingroup$ good point, but the curve should be 2D flat i.imgur.com/7bQuFPr.jpeg I am in blender 4.5. Try open the original file and let me know when you flatten all to Z=0 are normals ok? $\endgroup$ Commented Sep 2 at 6:55
  • $\begingroup$ 01: curve normal node i.imgur.com/GVEi1bS.gif 02: comp. normals i.imgur.com/Oe7WiAA.gif $\endgroup$ Commented Sep 2 at 7:22

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.