Horde3D

I found this thread on gamedev interesting and thought I'd share it here:
http://www.gamedev.net/community/forums ... _id=525805

These guys are trying to come up with a bias function based on the distance to the light, instead of using a constant bias value.

That is very useful! I detest shadowmap-acne, and tend to avoid shadows where possible to side-step the issue entirely. This definitely merits some thought and testing.

_________________
Tristam MacDonald - [swiftcoding]

Thanks for the interesting link. This subforum is definitely the right place for such things (actually there should be more link sharing )

I would say in the end the bias problem is closely related to the general EPSILON problem of floating point comparisons. Here is an interesting read:
http://realtimecollisiondetection.net/pubs/GDC07_Ericson_Physics_Tutorial_Numerical_Robustness.ppt