nope, from what i understand from his explainations, is that he uses a cubemap of some type for each light and does a depth test on each pixel if a pixel returns a closer value than the location of that point he darkens them.