I just confirmed that this is fixed by David's submission in r13067.
-Paul
On 4/12/2012 10:22 PM, Roland Hill wrote:
Hi,
I believe osg_NormalMatrix may be being calculated incorrectly in the trunk
version of OSG. At present, it is calculated as the inverse 3x3 of the
modelview matrix,
The weekend has been a total loss work-wise, so I didn't get too far with this,
sorry.
But I did back-date trunk to r13018, which was the last time oversimple3.cpp was
modified, and the lines of code in your proposed change haven't been modified
since then either. The example ran and lit
Hi Paul,
Thanks for looking into this. At least I know that it can be replicated outside
my environment!
Cheers,
Roland
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=47064#47064
___
osg-users
Hi Roland -- I dug in quite a bit further this evening and have narrowed down
the issue only a little.
The issue seems to be related to r13050, r13059, and the latest r13060, so I've
copied their authors on this post to solicit their feedback on this thread and
help track down what the root
Hi Paul,
Just a bit more information from me. I'm using Ubuntu 11.04 64 bit on two
machines (1 Nividia 1 ATI). I also compiled for Win64 and got the same
result. I'm building OSG for GL2 - the only deviation from a 'standard' build
is that I disable Display Lists on both machines otherwise
Something odd is going on here, as this works correctly in 3.0.1. I don't have
time to dig in to the math right now; maybe over the weekend...
-Paul
On 4/12/2012 10:22 PM, Roland Hill wrote:
Hi,
I believe osg_NormalMatrix may be being calculated incorrectly in the trunk
version of OSG.
Hi,
I believe osg_NormalMatrix may be being calculated incorrectly in the trunk
version of OSG. At present, it is calculated as the inverse 3x3 of the
modelview matrix, but should be the transpose inverse. The patch below appears
to correct this.
To test, try running osgsimplegl3 cow.osg. You
7 matches
Mail list logo
