Using CSM with real depth textures (D24X8) and HW PCF, how is the issue of oversampling in the farther cascades solved? This is solved with MIPs and anisotropic filtering for standard color maps. But with depth maps, you can't create MIPs (except manually) and you can't autogen MIPs. Is the only solution, short of VSM, to dump HW PCF and go with FP24 or FP32 textures, which allow autogen of MIP levels?
0
Cascade Shadow Mapping - Solving oversampling on distant cascades
Started by Scoob Droolins, Sep 12 2012 10:11 AM
1 reply to this topic
Ad:
#2 Moderators - Reputation: 5417
Posted 12 September 2012 - 01:38 PM
Even if you did generate mipmaps they wouldn't be useful, since in general filtering depth isn't useful when using it for for shadow comparison. What you really want is to compute the shadow occlusion value and filter that, but that's not really practical. VSM kind of works around this using a probabilistic model for determining shadow occlusion, where the value stored in the shadow maps (moments of a distribution of depth values) can be prefiltered and still have meaningful results.
