PDA

View Full Version : MR Warning:"warn 542009:"...


Andrei2k
06-17-2007, 09:57 PM
finalgather ray returned NaN for the finalgather point at pixel 101.245 42.5437

This happens a lot throughout the frame at different x and y values.

Anyone know what it means and how to "fix" it?

gga
06-18-2007, 11:32 PM
It means at least one of your shaders used in your scene is badly coded.
It is returning some NAN (not a number) values. These are floating point errors, caused, for example, from trying to divide by zero.
Check the objects at that pixel value, and, since the problem is with finalgather, with the objects surrounding it.
The only way to really fix the problem is to use a good shader. Changing a parameter in one of the buggy shaders may also fix the issue, but unless you know where the bug is, it is a hit-and-miss situation.

Puppet|
06-19-2007, 09:53 AM
As additional to Gonzalo said...
Sometimes problem with NaN may be because of (maya's bump + reflection + objects without UV). In other words object with maya's bump with reflection may reflect other object that have no UV.
Check you objects, may some of them have no UV mapping.
I suppose it's maya's bump shader problem.

Also you may try some shader that will replace NaN value, for example like this BlacksportsRemover:
http://dl3d.free.fr/download_shaders.php
But better find the problem and fix it.

Ironhalo
06-19-2007, 05:27 PM
As additional to Gonzalo said...
Sometimes problem with NaN may be because of (maya's bump + reflection + objects without UV). In other words object with maya's bump with reflection may reflect other object that have no UV.
Check you objects, may some of them have no UV mapping.
I suppose it's maya's bump shader problem.

Also you may try some shader that will fix NaN value, for example like this BlacksportsRemover:
http://dl3d.free.fr/download_shaders.php
But better find the problem and fix it.

good ideas. ive had nan problems in the past, but they seemed to be related to models sometimes as well. i'd hide a specific mesh and the nan would go away. normally it came from particularly unclean mesh topology.

Andrei2k
06-20-2007, 08:04 AM
Thanks for the suggestions I will definitely use them to single out the problem(s).

Would nurbs meshes ever do this as well? I'm dealing with nurbs that were brought in from StudioTools ...

MaNdRaK18
06-20-2007, 04:02 PM
Would nurbs meshes ever do this as well? ...

Doubt it. NURBS meshes ALWAYS have UV's hehehhhhh

techmage
06-20-2007, 10:57 PM
Don't know if this will be of any help. But I found that using the mia_material + roundcorners creates a high probability of NaN pixels if the surfaces start overlapping anywhere.

CGTalk Moderation
06-20-2007, 10:57 PM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.