I am considering purchasing VRay and had a question. Once VRay has calculated global illumination for an interior scene, does it have to re-calculate the solution for each different camera view or is this processed once for the entire scene? In other words, is the radiosity calculation view dependent as many Monte Carlo type renderers are? Or can the calculation be solved and saved and be applied to other views with accuracy?
It depends a lot on which GI method you use, and how divergent your cameras are.
If you use only photon mapping - no problem at all.
If you use irradience mapping - everything visible to the camera will be rendered, if you render from a different angle, you will need to do an incremental add in order to add in what is visible from the new angle that wasn’t before (this is the same for zooming in as well).
If you use light cacheing - you have to recalc the lighting at every view since it is view (and resolution) specific.