Would pinhole lenses work with a VR headset?

I'm no optics expert, but I would say no. A pinhole only allows light through that is already in focus, which you probably don't want to do. Also, a pinhole is a very small aperture, so it doesn't let much light through, which is bad for obvious reasons.

The Rift's lens needs to be a collimating lens that takes converging light and bends it into parallel paths.

collimating.png

Your eye's lens then focuses the collimated light onto your retina, which is why you see a clear image.

oculus.png

These images aren't representative of the actual shape of the lenses. This is just a quick draw up of my amateur understanding of the optical path.

Your eye image is also largely inrepresentative. Collimation lens ensures that light emitted from a certain point, would exit the lens at the same angle.

Thanks, that certainly makes more sense. I'll admit, thinking about optics doesn't come intuitively for me. Trying to figure out how NVidia's lens array works makes my brain melt.

It is actually fairly simple. Given condition of collimation, rays that are exit the lens at a specific angle, are all focused at one single point beneath it. The eye takes all the parallel light as a single point also. Basically, we're seeing objects at different places not because the light comes from different locations, but because it comes from different angles. Since collimated light is perceived similarry from any distance and is easy to focus onto, using the microlens array could open the road for right-next-to-the-eye displays. The lens could be finely adjusted to carefully give out only the right pixels at any given angle. However, unlike large lens, exact eye location creates great angle bias for the image, making it impossible to finely adjust microlens array in the first place. Or at least very challenging. Also, whereas large lens make headset precise location (and IPD for that matter) largely irrelevant, perceiving of microlensed display will change a lot if it moves about, because basically the whole picture will move. With large lens, as long as you keep the angle constant, the picture will also remain constant. Long story short, this approach has a lot of challenges to resolve.

The reason why you can't just put an individual microlens on top of every individual pixel is because the light would still exit the lens at all the angles as with unlensed pixel, save having that tiny area of an individual pixel collimated. But that area is much much smaller than iris area, and that essentially makes no difference as if there was no microlens in the first place. Imagine if on the image, all the rays were the same color - the picture would be blurry. Effectively, it's the same as watching bare display being that far away from you. Even if it's possible to only make "red" rays contain the picture, it would require the eyes to be in precisely defined locations to even see anything, which is just impossible.

"raidho36" wrote:
Your eye image is also largely inrepresentative. Collimation lens ensures that light emitted from a certain point, would exit the lens at the same angle.

Looks like you created a convex wide-angle lens. ;)