Rendering: Do we leave the mono camera enabled? about osvr-unity

[DuFF14]

No, in stereo mode we disable the mono camera. Looks like they leave the mono camera enabled for part of the frame (but not rendering) so that it can be referenced easily by other game logic. Unity developers commonly use Camera.main to get the scene's main camera, but this usually assumes there is one camera in the scene.

I really like the EndOfFrame coroutine as it provides a system for frame timing in Unity which will be crucial for rendering optimizations.

[russell-taylor]

One approach is to have a vector of cameras that has one in it for a mono viewpoint and two for stereo. This could be generalized for a multi-zone stereo display.

[rpavlik]

I think there's probably value in adopting the approach used by the Cardboard plugin then, right?

[DuFF14]

Yes.

[DuFF14]

Working on this here: https://github.com/OSVR/OSVR-Unity/tree/enableCamera
Also added calls in the EndOfFrame coroutine to update the state of VRHead's tracker interface to see if it reduces perceived latency. To me, it feels the same or better in the demo scene. Will test in more complex scenes, and might make some more changes before opening a pull request. (I also am updating interfaces from LateUpdate in this branch exactly as in #64)
From the docs, WaitForEndOfFrame "Waits until the end of the frame after all cameras and GUI is rendered, just before displaying the frame on screen."
Also see Execution Order of Event Functions.

[DuFF14]

This was included with the DisplayConfig API update.