I really wonder if the next-gen engines and Oculus are a perfect match, at least in the near term. UE4 really has awesome graphics, but it obviously comes at a cost. And this seems to be even more significant for the Oculus, where you have to reach consistent fps of 75 @1080 instead of 30/60, and despite the fact that a lot has to be rendered twice. The solution can’t be that you need a GTX 780 to run an Oculus game - how many end users have one of those? So maybe right now it might be better to use an engine for Oculus games that is optimized to run on lower-end hardware, and sacrifice some of the graphical bells and whistles?
Right now there is no entry on the roadmap for VR development. I haven’t looked at the Oculus plugin code yet - is it really optimizing the render loop for the Oculus? I mean, you can e.g. take a brute-force approach and render everything twice, or you could share some render information between both eyes, but probably you need special shaders etc. for that? I am no expert on this, just wondering how much optimization potential there still is. And if so, will this make its way on the roadmap?