Using Perception-Based Filtering to Hide Shadow Artifacts (DAAV)

Combining filtering techniques with shadow mapping is a common tool to simulate soft shadows in real-time applications.  A positive side-effect of such approaches is that the filtering also blurs aliasing artifacts caused by low resolution shadow maps, thereby improving the visual quality of the shadow.

In this work we investigate the correlation between filter radius and shadow map resolution to optimize computational performance while mostly preserving the visual quality of the soft shadow.

We present the results of a user study and offer a ready-to-use function to compute for shadow map aliasing artifacts a respective filter size that makes it unrecognizable.