0
$\begingroup$

Has anyone seen this effect before? I made polar UV coordinates for the ring and applied the standard brick texture. The polar coordinates basically take a square in UV space and squash it onto a trapezoidal face. I would have expected linear interpolation throughout the face, but something odd happens near the middle. In the top left quadrant, I used ctrl-r to split the faces and line up the split edges with the texture lines. That straightens them out, so I'm sure the linear calculation is correct along all the face edges.

After taking the picture, I tried splitting a distorted quad along the diagonal, switching between fixed and fixed-alternate. It turns out, one of the split directions matches the distortion, and the other direction distorts the other way. So behind the scenes, quads are made into triangles for the UV mapping.

Unless there is a different UV calculation mode for blender, I think the only fix is to split the trapezoids radially.

enter image description here

Improve this question
$\endgroup$
2
  • $\begingroup$ Hello and welcome to BSE. Can you share a simple example blend blend-exchange.com $\endgroup$
    – Harry McKenzie ♦
    Commented Jun 2 at 2:11
  • $\begingroup$ Attached a blend file with a single quad showing the behavior. Try splitting the quad vertically with ctrl-r. then undo that and use ctrl-t to split on diagonal and swtich the diagonals back and forth. $\endgroup$
    – D Hughes
    Commented Jun 2 at 3:03

1 Answer 1

Reset to default
0
$\begingroup$

Your node setup is wrong.

It's easy to fix :

img

* because for a generated texture (Mapping) Vector also should be set to Generated

Improve this answer
$\endgroup$
3
  • $\begingroup$ No, the generated texture coordinates would end up with no distortion. The point here was to introduce distortion by moving the quad vertices. But each half of the quad is apparently calculated separately, so you see the texture being sheared in each triangle, but they don't match at the diagonal line. $\endgroup$
    – D Hughes
    Commented Jun 3 at 21:27
  • $\begingroup$ @DHughes — Be informed: I do not make guesses. I've figured out the solution investigating your test file. — So, please, just try to do the thing I've suggested in my answer. — After that tell me: weather it worked as good as it did with a test file, or did it not. — Ok? $\endgroup$
    – Yaroslav
    Commented Jun 4 at 0:00
  • $\begingroup$ Nope, did not fix it. If you load my demo blend and try your suggestion, you will see the bricks at the bottom are the same size as the bricks at the top. They need to be smaller (same count at bottom as at top). $\endgroup$
    – D Hughes
    Commented Jun 4 at 0:36

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .