Coding without a laptop: Two weeks with AR glasses and Linux on Android

So this is probably a silly question but… can’t you fool your eyes into focusing at any distance you want if you’ve got a stereo screen at a fixed distance (ie a headset)?

Isn’t this just a function of the parallax when rendering both screens?

Focus is distinct from convergence - convergence is how much you have to cross your eyes to look at something, but focus is where muscles squish and stretch your eyeballs to change the distance of your retinas from your pupils. Just like a camera, if your eye is not focused at the real distance to subject, it will look blurry because your pupil is not a perfect pinhole, but has area (so a single eye is already seeing the same object from slightly different angles, on either side of the pupil).

Usually your brain learns a strong correspondence between focus and convergence, but this can be unlearned quite easily, and indeed must be in order to view e.g. VR, 3D films, Magic Eye pictures, etc... - all of which encode 3D information through convergence, while requiring your eyes to focus on a fixed plane.

No, and VR’s inability to match focal distance with parallax causes the vergence-accommodation conflict.

This is a real problem, but it's fine for most VR use cases as you're usually looking at content that's rendered at a distance greater than the focal plane. The problems start to occur when you look at nearby content - that is, content less than 2m away - as it ends up being extremely uncomfortable for your eyes.

There are solutions being developed for this, but they have not been successfully miniaturised and/or cost-reduced for productisation. It's unclear how far away it is at this time, but Reality Labs has several generations of solutions that physically change the distance between the lenses and the displays, and alternate solutions like lightfields capable of simultaneously displaying content at different focal planes are being investigated.