After implementing Cascaded shadow maps, render method for directional shadows become a hidden iterative method, and iteration
is based on the layers of frame buffer. When testing EVSM, I find
that custom handling is very confusing.
Current logic:
Render method is called iteratively, attaching new depth layers for each call
Expected logic:
Maybe: All frame buffer attachments are attached layer by layer
Better one:
Interface for showing what is expected behaviour
Best one:
Interface to choose iteration, and iterate which frame buffer objects
After implementing Cascaded shadow maps, render method for directional shadows become a hidden iterative method, and iteration
is based on the layers of frame buffer. When testing EVSM, I find
that custom handling is very confusing.
Current logic:
Render method is called iteratively, attaching new depth layers for each call
Expected logic:
Maybe: All frame buffer attachments are attached layer by layer
Better one:
Interface for showing what is expected behaviour
Best one:
Interface to choose iteration, and iterate which frame buffer objects