What's going on with the weird shading on my torch?

[syclonix]

Does anyone know how I can fix the weird shading on my torch brackets? Admittedly I probably made them the wrong way by placing a circle, extruding inward, then extruding upward to give it volume; the normals were inverted so I had to invert the normals, but now my shading is really strange.

Any suggestions appreciated!

[NP5]

Check the normals direction.
Overlays dropdown, tick ‘Face Orientation’.
Blue is good outward facing, red is bad inside facing out.
Select all, Shift N recalculates outside.
Alt N provides a menu of options Useful to flip individual selected faces.

[syclonix]

Thank you for the tips NP5. With face orientation, the brackets were originally red, then I pressed Shift-N and inverted the red normals to blue. As you can see in the screenshot above, the circular brackets still have the irregular shading issue.

Do you know what might still be causing it?

[BH67]

@syclonix

Admittedly I probably made them the wrong way

No right or wrong way as long as the end result is good! I actually never thought to do it like that =)

I followed your repro steps: placed circle, extruded the edges inwards, A to select the resulting faces, extruded them up, and then flipped the entire object’s normals. No issues.

Since you’ve already checked the normals and there aren’t many modelling steps along the way, you most likely have either duplicate topology (which would usually cause obvious z-fighting) or excess topology (such as edges connecting vertices that should not be directly connected. Your second screenshot looks clean of this, though it might be hidden under duplicate topology). Some additional standard stuff to try:

• A, M–>By Distance
• Select one random vertex (or edge/face) at a time and test-grab it around to see if you have duplicates that the merge didn’t catch (this can happen if you accidentally apply a transform instead of cancelling it, or duplicate in Edit Mode without realizing it).
• Turn on X-Ray Mode (or Wireframe) and the Statistics Overlay, then carefully box-select some topology on one part of the bracket and see if your selection makes sense in the statistics. Same idea as the previous point, but a bit faster and broader in scope.
• Since it looks like you’ve got some sculpted stuff here ( ), you might also want to make sure you don’t have any modifiers unexpectedly affecting the area. I doubt that’s it, but might as well mention it.

If there’s one good thing, it’s that it looks like you have the same issue the entire way around both brackets, which should make it easier to detect. Good hunting!

[syclonix]

Thank you BH67. I followed the steps, but couldn’t fix the issue. I’m almost certain the issue cropped up because of the unsanitary way I made my previous brackets. The good news though is that by remaking the hollow cylinders brackets, I found a better way to make them

This time I created a cylinder, deleted the top and bottom faces, then used the "solidify modifier to give it thickness. It’s all good now!

[NP5]

I would next have suggested checking for custom split normals.
In the mesh data (green triangle icon) Geometry Data dropdown.
If, Clear Custom Split Normals Data was an option, click it.

However you have solved it another way, all good.