Per Reflection Capture Resolution

In order to blend between multiple reflection captures efficiently we have to put them into a cubemap array, which requires that all cubemaps are the same size. I don’t expect this requirement to change.