2
$\begingroup$

I think there is something wrong with this mesh, but I cannot find the problem. Am I missing something?

enter image description here

Download blend file here:

https://blend-exchange.giantcowfilms.com/b/2889/

$\endgroup$
1
  • 1
    $\begingroup$ you have smooth shading... that is the effect in 3d viewport. try with the default cube, it looks weird with smooth shading. Your model is ok and it doesn't mean it wil not render good later $\endgroup$
    – m.ardito
    Mar 7, 2017 at 12:14

1 Answer 1

2
$\begingroup$

Like m.ardito said, it it an issue with smooth shading. Contrary to what he said though, it won't render correctly in many situations. That is how Blender calculates the normals...

To fix the problem, you could add a bevel modifier and potentially turn on auto-smooth in the mesh settings.

Here is what it looks like with this bevel modifier:

enter image description here

enter image description here


To understand what's going on, let's look at the default cube with smooth shading.

enter image description here

It too has rather hideous shading "problems" shown by the sharp gradient lines between the closest and furthest vertex on the top. As I understand it, smooth shading averages the normals from the vertices so that there isn't a sharp change at the face's edges. But unless you use a normal map, you must use some sort of math to set what the normals will be across the face. For flat shading, the normals are the same across the face, but have a hard cut at the edges. An easy way to fix this is to just add more geometry. This allows the smoothing to look much more like what you'd think. Here are a few examples of ways to introduce this geometry.

Bevel Modifier

enter image description here

Notice how the normals get better and better with the higher amounts of geometry.

Subsurf Modifier

enter image description here

Notice that the subsurf modifier is just subdividing the faces. It is not doing any smoothing itself because it is set to simple instead of Catmull-Clark.

$\endgroup$
7
  • $\begingroup$ I think you are right, this solves the problem. Could you perhaps explain a bit more what is the problem with the normals and why you chose this bevel modifier for solving the problem. Why these settings of the modifier. I just want to understand. $\endgroup$
    – user13877
    Mar 7, 2017 at 12:33
  • 1
    $\begingroup$ by the way, do you have a suggestion for a better title for my question (so that other people can find the specific problem/solution) ? $\endgroup$
    – user13877
    Mar 7, 2017 at 12:34
  • 1
    $\begingroup$ FYI (and I'm surprised about that), it minimize the effect, but it is still there. $\endgroup$
    – lemon
    Mar 7, 2017 at 14:56
  • $\begingroup$ @lemon Are you surprised that it minimizes the effect or that it is still there? $\endgroup$
    – JakeD
    Mar 7, 2017 at 21:31
  • 2
    $\begingroup$ @lemon With the bevel technique, it is still averaging the vertex normals across the face, and the top-most 4 vertices are still not straight up and down. So the angle improves dramatically (because it was originally 45º off), but it still isn't 0º. That's the best explanation I have for it. $\endgroup$
    – JakeD
    Mar 8, 2017 at 12:47

You must log in to answer this question.