> which seems like it would be uncomfortable in real gaming scenarios.
I routinely want to grab things of tables behind me in VR without turning my head off of the enemy in front. It's one of the things that makes VR feel immersive, the world keeps existing even if its outside your field of view.
Sure but for a 1-2 second grab extrapolation + accelerometers should do the trick.
Maybe but I'm skeptical it would be good enough.
The industry seems to be hurrying to cut costs with lower end solutions and yet the high end isn't even close to perfect yet IMO.
Its too early to commoditize VR in my opinion. Google tried with daydream and its crap. When Vice launched it was remarkable and since then the quality of the VR experience has only degraded. Lots of "we can get almost as good results doing this cheaper option" steps have been taken which cement VR as a gimmick rather than a truly compelling experience.
John Carmack has also talked a decent bit about their inside-out tracking extrapolation tracking code to keep latency low. Their tracking predicts where your hand will be at the time a frame is rendered, not at the time the frame begins. This not only means the controller tracking will appear low latency but it also means it can figure out from the controller's sensors where it is when it can't get visual range for a second or two. So I imagine you'd have to hold your hand directly behind your back for more than a few seconds to really see the effects, and by the time your controller comes in range (way before you'd see your hand position) it should correct itself.
That being said, some people with hands on did say the inside out tracking didn't do so well when the controller got too close to the front of the headset, so it's up in the air how bow and arrow type of movements will work.