Cascaded shadow maps

In the previous two chapters when we covered lighting calculations, the directional light was always the first recipe. The directional light has no attenuation to worry about and it spans infinitely in every direction. These two qualities which make directional lighting easy to calculate, turn into an obstacle when calculating shadows.

In order to add shadow's support to the directional light, we need a way to map the entire scene area into a fixed size shadow map. For a very small scene, it may be possible to allocate a large shadow map and have it cover the entire scene. For a decent-sized scene, however, this will result in very bad scene to shadow map pixel ratio. If you recall from the introduction, this pixel ratio is a key factor in the amount of shadow-related artefacts we are going to see in the lighting/shadowing result image. Getting good quality shadows while lighting a decent-sized scene requires a more sophisticated solution.

Since we can't cover the whole...