Why using half pixel distribution in biasing of variance shadow mapping I have a question of calculating bias in variance shadow mapping.
