Hello Keyan,

On 11/09/2011 01:48 AM, Keyan wrote:
> I tried to add a usleep, but the behavior didn't change, therefore, I 
> uploaded a screenshot to picpaste

hmm, not sure why sleeping would make a difference. The order in which 
things (shadow map and scene) are rendered for the first time are 
relevant, not so much the timing of these operations. The GPU can only 
do one thing at a time anyway which that serializes operations, luckily 
we don't have race conditions on the GPU in the way you get them on CPUs 
yet ;)

> http://www.picpaste.com/ace364df969ce68907ae34d443b55312.jpg
>
> The artifacts are stable, i.e. they don't change while the simulation is 
> running, or lets say, some movie is playing in quicktime (i use mac os x).
>
> I use the coordinate system of the physics engine, therefore, the ground 
> plate is z = 0, and upwards in the scene is z>0, downwards is z<0. The camera 
> is located higher than the walking machine, hence, i looks into negative z 
> direction. if i rotate the camera towards positive z-axis, the artifacts 
> disappear, and don't show up anymore.
>
> It happens with any shadow mode (I tried them all, with different smoothness 
> parameters).

hmm, those artefacts don't look like they would come from the shadow 
mapping [1], for that I would have expected "random" black splotches or 
something along those lines.
The only time I've seen artefacts like those in the image was when I had 
a graphics card that was about to die and the memory chips started failing.
Do you see artefacts with any other OpenGL programs? Are your drivers up 
to date?

        Cheers,
                Carsten

[1] It may well be that enabling shadows triggers this for you, because 
that causes the scene to be rendered into an FBO instead of the 
application frame buffer, so the memory touched is quite different with 
shadows than without.

------------------------------------------------------------------------------
RSA(R) Conference 2012
Save $700 by Nov 18
Register now
http://p.sf.net/sfu/rsa-sfdev2dev1
_______________________________________________
Opensg-users mailing list
Opensg-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/opensg-users

Reply via email to