Two bugs
8167 10 0-
- regget
- Member
- 73 posts
- Joined: July 2005
- Offline
I've found two bugs in build 8.1661
1. When rendering with Mantra a (closed) transparent surface (a sphere with supermaterial, tuned to transparency, with an alpha map on) with deep shadow map generation the rendered image shows up small dots out of color…, so very ugly. With normal raytracing (no shadow maps, area shadows) this problem disapperas, but this is not the way I wished to render. (Not present in build 8.1655)
2. If I make a new desktop, which is associated with a file (when opens it uses that) than the viewer pane is totally out of place, when switching to build desktop than switching back to the desktop which it wants to use when opens than everything turns ok.
1. When rendering with Mantra a (closed) transparent surface (a sphere with supermaterial, tuned to transparency, with an alpha map on) with deep shadow map generation the rendered image shows up small dots out of color…, so very ugly. With normal raytracing (no shadow maps, area shadows) this problem disapperas, but this is not the way I wished to render. (Not present in build 8.1655)
2. If I make a new desktop, which is associated with a file (when opens it uses that) than the viewer pane is totally out of place, when switching to build desktop than switching back to the desktop which it wants to use when opens than everything turns ok.
-
- regget
- Member
- 73 posts
- Joined: July 2005
- Offline
-
- mark
- Staff
- 2675 posts
- Joined: July 2005
- Offline
regget
I've found two bugs in build 8.1661
1. When rendering with Mantra a (closed) transparent surface (a sphere with supermaterial, tuned to transparency, with an alpha map on) with deep shadow map generation the rendered image shows up small dots out of color…, so very ugly. With normal raytracing (no shadow maps, area shadows) this problem disapperas, but this is not the way I wished to render. (Not present in build 8.1655)
There was a small “fix” made to transparent surface rendering in 8.1.659. I expect this has broken the case you describe.
However, I can't seem to reproduce it. I've attached my test case (which generates the enclosed image).
I've only tested on Linux, so it's possible that there's a compiler difference between Windows and Linux. Could you please post a sample?
Thanks.
-
- djorzgul
- Member
- 235 posts
- Joined:
- Offline
-
- regget
- Member
- 73 posts
- Joined: July 2005
- Offline
-
- regget
- Member
- 73 posts
- Joined: July 2005
- Offline
in some areas the render is even worse, and the same result when u look through and finding first a different surface, not the closed sphere (actually it is an opened globe at the top and bottom, underneath a floor catching shadows, when u look through the nearer part towards the floor, the same result, and even worse)
-
- regget
- Member
- 73 posts
- Joined: July 2005
- Offline
-
- mark
- Staff
- 2675 posts
- Joined: July 2005
- Offline
regget
Other example, the floor one
I've tried sticking a map on, with no luck. Would it be possible to bundle up a simple example and either send it to me by mail (mark at sidefx.com) or post it to the list?
My test case seems to render ok on Windows (even with a map), so there must be something different. Please try to bundle up an example file.
Thanks,
-
- regget
- Member
- 73 posts
- Joined: July 2005
- Offline
-
- mark
- Staff
- 2675 posts
- Joined: July 2005
- Offline
regget
Ok, sent to mark at sidefx.com
The rendering bug should be fixed in 8.1.666. If you're superstitious, you might want to wait till 667.
-
- regget
- Member
- 73 posts
- Joined: July 2005
- Offline
-
- Quick Links


