At the moment we only rely on the results of the main camera's culling. This is not accurate for shadow views: objects that are not visible in the main view are skipped, however, their shadows might still be visible.
One particular challenge in fixing this in EveryRay is the fact that ER_Camera has methods for culling but shadows are not using that class and instead create their own views with ER_Projector, ER_Frustum. We either need to move everything in the shadow system to ER_Camera or find some workaround.
At the moment we only rely on the results of the main camera's culling. This is not accurate for shadow views: objects that are not visible in the main view are skipped, however, their shadows might still be visible.
One particular challenge in fixing this in EveryRay is the fact that ER_Camera has methods for culling but shadows are not using that class and instead create their own views with ER_Projector, ER_Frustum. We either need to move everything in the shadow system to ER_Camera or find some workaround.