We lose substantial time in the render loop from the UI and particle system features that use two additional cameras (CameraPlaceholder and "Text Camera"). The former is used as part of the docked UI system where the 3D view camera is embedded in the UI panel, and the latter "Text Camera" is used to render the hit and miss X and O markers as particles.
Some change to this where we are not re-rendering otherwise static cameras every frame is needed to not literally add 20-30% frame time overhead for 0 reason.
