I lately obtained an iPhone 7 and tried out Google Cardboard. Hidden Temple is a really good demonstration of its talents.
This video will assist me check calibration for what I contemplate to be probably the most snug and efficient most depth for rendered pixels. This would range for various individuals due to the variations in eye positions on the face, so I’ll experiment asking a couple of individuals to let me know which worth works greatest for them.
If you might have an iPhone7 and Google Cardboard harness, you possibly can assist by letting me know right here which worth provides you the deepest depth seen comfortably. Thanks! The values vary from zero to 100 then return to zero once more. For my half, the most effective worth is round 57 or 58 whereas the video is enjoying. However, if I pause the video, I can nonetheless resolve the depth all the best way right down to 50 or so (and with additional relaxed experimentation, all the best way right down to 37, however I do not know if my eyes are pointing like a chameleon at that worth – for now, I am continuing with 40, which can not work for individuals whose eyes are nearer collectively than mine). It could also be higher if I render an precise scene with far distant points quite than a dot. In the morning, whereas my eyes have been nonetheless drained, I discovered the dot separating at round 60, although I can take a look at distant timber out the window simply high quality.
Once I can set up a great calibration for deepest depth, I can attempt rendering some scenes and see how they end up, and add them to share. I haven’t got calculations for the lens curvatures but, so the sides of the scenes will probably be distorted within the lenses till I can publish course of the scenes with the right edge curvatures.
I do not at present have a dev surroundings for mobile units, so I’ll fudge some makes an attempt at making movies that look good in Google Cardboard on a four” mobile system.
This video is formatted for four” long 16:9 mobile screens ONLY. I don’t know how well devices other than an iPhone 7 will work with this, but I’m assuming any screen that is 4″ lengthy with a skinny edge that matches into the Google Cardboard harness will work.
If you want to make use of this your self to calibrate the place you need to put the ahead vanishing level in your eye scenes, simply look ahead to the dot to separate on the primary cross and are available collectively on the second move. The place of the left dot alongside the middle axis throughout the screen is (2.eight + the quantity * zero.005) / eight.zero. zero.5 can be lifeless middle within the scene if it was a potential end result (highest potential is zero.475 (100)). I often get a quantity round zero.3875 (60), which is to the left of middle. Different individuals will get totally different numbers because of the totally different width of their eyes. The white vertical division within the center is four pixels large out of 1334 (wider screens would use extra pixels in my technique), and the width of the 2 eye scenes does NOT embrace that quantity, so you will need to make your eye scene widths barely narrower than merely dividing the screen in half. The two scenes are flush with the left and proper sides of the screen – the remaining is taken up by the center division. I exploit a vertical subject of view of 77.eight levels which comes near actual world angles represented close to the middle of the view (the lens curvatures trigger the sting extremes to be off a bit) . Again, this check was made for four” screens.
I’ve had a couple of different individuals do this app. My worth is 60, one other was 52, and one other was 70. People that measure at decrease numbers can resolve the upper numbers and nonetheless get a way of depth, however the perceived quantity of depth won’t be as pronounced for them. So I’m wondering actually how any VR headset or apps that do not permit bodily calibration can present a constant expertise for everybody. The ones with their eyes closest collectively and that may nonetheless comfortably resolve the distant pixels will get probably the most pronounced 3d expertise.