-
This patch is exactly right and is now committed, thanks!
2009-10-05 13:13:21 UTC in Aqsis Renderer
-
c42f committed revision 2896 to the Aqsis Renderer SVN repository, changing 3 files.
2009-10-05 13:13:10 UTC in Aqsis Renderer
-
Ok, this is now fixed courtesy of a patch from Ricardo Mayor on the mailing list. Thanks Ricardo!
2009-10-05 12:54:36 UTC in Aqsis Renderer
-
c42f committed revision 2895 to the Aqsis Renderer SVN repository, changing 4 files.
2009-10-04 00:48:30 UTC in Aqsis Renderer
-
This one is fixed as of 809e878bf2c58bd2d76b2cc6bfcc3671029f8b79.
The artifacts were due to incorrect filtering. There was a bad interaction between the bucket overlap caching and "rendering" of buckets which were fully outside of the crop region: The sample region of a bucket could be reduced in side due to application of a cache from a non-rendered bucket, and then reduced in size again...
2009-10-03 09:33:15 UTC in Aqsis Renderer
-
c42f committed revision 2893 to the Aqsis Renderer SVN repository, changing 3 files.
2009-10-03 02:09:29 UTC in Aqsis Renderer
-
This bugs seems to be something to do with the filtering code: it's only present when the filter width is exactly an even integer. So if I have a crop window in conjunction with
PixelFilter "gaussian" 4 4
I see the bug, but with widths of 3.9 3.9 or 4.00001 4.00001 the bug goes away!
2009-09-29 05:41:00 UTC in Aqsis Renderer
-
c42f committed revision 2892 to the Aqsis Renderer SVN repository, changing 1 files.
2009-09-28 07:38:48 UTC in Aqsis Renderer
-
I can confirm this bug with the latest repository version; the artefacts look oddly like the geometry has been mirrored in the first row of pixels or something...
2009-09-28 07:25:17 UTC in Aqsis Renderer
-
This bug is fixed since the frustrum-based culling volume has been removed prior to aqsis-1.6. It was the frustrum culling which was causing geometry to be discarded incorrectly.
2009-09-28 07:19:11 UTC in Aqsis Renderer
