I don’t know the answer to your question, but AFAIK VR Lab made by Valve is mostly Unity and Source 2, forward rendered with MSAA and supersampled, so it can be super clean. Both, Valve and Oculus, seem to be advocating against screen space and post processing (like FXAA and Temporal AA) methods in VR.
Oculus just released their VR renderer for UE4
More about it here: https://youtu.be/Xk3WUk5T2TQ?t=2525
It’s also worth watching Valve’s presentations, like the Advanced VR rendering (GDC 2015 and 2016). Here is an article that summed up some of the informations from the last one.
Thanks for the info, that makes a lot more sense now, I think I will stick to stock UE at moment however once my CV1 turns up I will give both deffered and forward rendering a test.