I am experimenting with split screen. Obviously, I have to make script agnostic from any "Camera.main" calls, because that would be ambiguous.
But in the case of things like procedural ribbon trails, it's important to orient faces toward the current rendering camera.
Normally, I am constructing geometry to face the camera inside of Update. I would assume that this is called just once, regardless of the number of cameras rendering concurrently.
Have I completely missed that there's already a "Render" function, called before each camera renders? If there is, obviously, the geometry generation would be more appropriate here!
When calling Camera.main from within such a call, I assume it would point to the current rendering camera?