 |
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
I'm having a problem with a smooth mesh2. For a face recognition project I
make smooth mesh2 objects from datasets. Each data set is a little
different, so is every mesh.
After solving the strange lighting problems with the double_illuminate, I
kept some strange bright triangles:
www.wattsup06.com/images/render.png
As I just said, every mesh2 is a little different; most parts of the renders
are smooth, except for 1 to 3 triangles which appear at different places for
slighty different meshes.
Help would be appreciated :)
Sid
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Mousehunter <svi### [at] click2c net> wrote:
> After solving the strange lighting problems with the double_illuminate, I
> kept some strange bright triangles:
This is most probably caused by the inverted normals problem described
here:
http://tag.povray.org/povQandT/SmoothTriangleArtifact/index.html
Also this is good to know about smooth triangles:
http://tag.povray.org/povQandT/TheShadowLineArtifact/index.html
Unfortunately there's no solution currently. If your objects cast
shadows then you may want to try making them closed so that they
will self-shadow. If they are shadowless, then the problem becomes
more difficult.
Increasing the mesh resolution (ie. more triangles) might help
a bit (at least reduce the sice of the artifacts).
--
- Warp
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
>
> This is most probably caused by the inverted normals problem described
> here:
>
> http://tag.povray.org/povQandT/SmoothTriangleArtifact/index.html
>
> Also this is good to know about smooth triangles:
>
> http://tag.povray.org/povQandT/TheShadowLineArtifact/index.html
>
I read those, but I found it strange that it all works for all the other
triangles. A small change in position or normal makes the artifact come or
disappear...
> Increasing the mesh resolution (ie. more triangles) might help
> a bit (at least reduce the sice of the artifacts).
>
Indeed, but the data I have is on a 40x40 grid. And when the normals are
determined, the outer bounds are lost, so only 38x38 remains. Not really
fancy for face recognition...
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
