Head position-informed dynamic fixed foveated rendering? Probably an obvious idea, but here goes.