Hacker News Reader

Background: 100+ hours in Vive, hobbyist VR dev, and 20 demos or so

I see two kinds of VR "sickness".

1. There is sickness caused by locomotion implementations. Testing seems to indicate there are some people who will just never adjust to these. I really don't think FOV and eye tracking will fix it. I've experienced locomotion sickness in some games, and in all of them I've been able to completely adjust in 1 to 2 hours. That adjustment seems to be relatively permanent as well. I don't think it is a rare trait, but it seems that not everyone can do that.

2. There is this sort of initial "VR Fatigue" where the eye tracking issues might fit in. I've noticed a trend the first or second time people use VR for a real session (1 hour +), it is somewhat common (50%?) they get a little nauseous. I've really just been chalking that up to "new stuff is weird" and the heat the headset puts out. This sort of sickness seems to universally fade quite quickly for people, and could have to do with the head on a pivot thing. From what I've seen, most people find the sweet spot/head pivoting natural almost immediately.

TL;DR I spend a lot of time in VR and with people in VR. Eye tracking or larger sweet spots is near the bottom of the list of problems/discomforts.

Nausea inducing incongruous locomotion is of course an absolute fail. But that is not a problem unique to VR, and for the most part you know it going in if you're likely to get motion sickness.

The actual issues with VR of eye fatigue are more subtle, and often hard for users to identify. They simply get an increasingly negative attitude towards the experience over time, and over repeated use. This is why an entire industry can spin up and have what seems like a great product, but have an odd lack of audience.

The exact same thing happened with stereo in movies. The technology existed for a long time, but lost favor because it was a subtly unpleasant experience. Avatar and the modern resurgence of stereo has been more successful because of the decision to always converge the subject. Even now many supposedly expert "stereographers" try to emphasis the "stereo" effect by pushing the subject into a different depth from the screen by not converging on the subject.

Watch a stereo version of Avatar without 3d glasses. You'll find that 95% of the film is completely watchable without glasses because the subject is always converged, or the shot isn't even stereo.

I get motion sick in some games, notably the Half-Life series, unless mouse input smoothing is enabled. I also get motion sick in cars sometimes. Would this be the former type of motion sickness?

I have a little trouble understanding what you're trying to say here but I don't think it's really accurate. Positionally tracked headsets like the Vive and Rift do not model vision as if our eyes are permanently stuck forward. They are very much intended to be used with normal eye movement within the field of view of the headset while the head is stationary.

Both are low persistence displays with a 90Hz refresh rate. Low persistence was an innovation from Valve explicitly designed to reduce blur and ghosting resulting from eye motion relative to the display as you get when fixated on a stationary virtual object while turning or moving your head or when trying to track a moving virtual object with your eyes while keeping your head still.

In an ideal world we would have even lower latency head tracking and higher refresh rates but current headsets already allow for legible text when moving your head thanks to low persistence.

I'm not familiar with the Minecraft hack you mention but approaches to snapping view changes are generally designed to address issues with using a controller rather than actual 1:1 head motion to change view orientation. View rotation independent of head movement tends to produce nausea due to the disconnect between visual cues and cues from the inner ear but really has nothing to do with any supposed assumption of eyes being stuck forward relative to head orientation.

Issues with convergence, accommodation and depth of field raised in other replies do still exist with current display technologies but they are again not particularly tied to any assumptions of fixed eye orientation.

Try wearing progressive bifocals. I think you will find that the experience is very similar. You have this little window where you can see stuff, but it has very small angle of visibility. In the top half you can focus out to infinity. In the bottom half you can focus up close. If you look in the wrong place, there is absolutely no way to bring what you are looking at into focus. For the vast majority of things, that means you have a single focal length (infinity) and you bring the thing into the visible part of your gaze by moving your head.

The thing that is really strange is that within a few weeks of wearing progressive bifocals, everything just looks normal. Your brain compensates. Or to put it in a different way, your brain is always compensating so that while you think you have great vision, the reality is that your brain is paving over all of the badness. It really does an amazing job even with something as unbelievably bad as progressive bifocals.

VR does not need to provide the same experience as reality (without glasses). It just needs to provide something good enough that the brain can be trained to pave over all the badness. I've only ever played with VR headsets at a store, but I can tell you that they are already better than my bifocals ;-)

This, and the fact that you cannot focus into the distance. In real life most of your view is blurred except the part you are focussing on. In VR it still just looks flat, because if you focus on something in the distance, nothing happens.

I wonder what happens if one uses VR goggles for too long. Will one lose the ability to focus?

If you wear prisms that invert the world, your brain will eventually see the world inverted when you remove the prisms, and this effect will auto-correct after some duration.

It also gives you one hell of a migraine.

Things might work differently for small children who wear VR goggles for extended durations. Their brains might get wired the wrong way, and there might be no "auto-correct" in this case.

I have a PHD in small children vision, I concur.

why would an adult who grew up with normal vision be able to adjust to something else but not the opposite?

Adjusting to the inability to focus is different from learning to focus. Like the jellyfish who can't orient themselves properly if they were born in zero gravity [1].

http://www.businessinsider.com/space-born-animals-adjust-to-...

I don't think it would be catastrophic, unless these children spend 90% of their time playing VR games. I would expect it to be similar to exposing a child to a second language. If you replace their exposure to the previous language completely with the new one, their proficiency at the first one will suffer but if you provide enough time for both, the brain will improve in both.

I'm watching developments with light field displays for this with interest.

That said, you don't really notice this effect, or much else, once you're immersed in something. I've hundreds of hours in the vive and I still punch walls, the ceiling, lean on objects that don't exist.

The point is that regardless of the various things that would make it better (4k light field displays, bigger fov, higher refresh, blacker blacks, etc etc) the tech is good enough today - the fact that one is able to surrogate away from "I am stood in my living room wearing some silly goggles" to "I am stood in the snow atop a peak watching the sun set, I should get inside or I'll get cold" is the litmus test. Literally everyone I've jammed in it, young and old, has lost themselves in it.

I'm really curious if this is based on your actual experience with a Vive or if this is an assumption based on what you know about the technology.

I read an article at some point that claimed this was an issue so I consciously tested it in the unit - focusing on something near my face and then farther away without moving my head and it actually feels (to me) pretty correct.

I don't really know the science behind this versus light field displays or whatever, but if you have access to a unit I encourage you to try it - it's very difficult to detect issues with focus and it absolutely does not look flat (again, to me).

I would be very interested of all the opinions in this thread which of them are from people who own a unit and have tried these things and which are assumptions.

This technology is so incredibly convincing I'm embarrassed to say that I fell over trying to lean on a virtual pool table at one point. I simply forgot that it wasn't a real thing. :)

There's no focus in any VR headset currently. So whatever you experienced was not that. You might be thinking of the stereo convergence, where the left and right images come together where you look at them. But they don't have any capability to adjust to the focus of your eyes right now.

Yep, this is the problem. And we've known about it for more than a decade.

I tested using VR when we were building the Virtual Production system for Avatar back in the summer 2005. We were seeing if it would help the actor's eye lines to actually see the virtual world they were performing in. The problems then are still the problems now. Nothing meaningful has changed with VR yet.

- Lack of convergence. (Movies, done well, are converged on the subject, VR is not)

- Lack of depth of field.

Latency, fov, and resolution were also issues but to a much lesser degree.

No one will be able to use VR for long periods of time until those two issues are resolved. In fact, I think it is actually irresponsible to let anyone under the age of 12 or so use VR without correct convergence and focus due to how it may effect ocular muscle development.

> I personally think it's what makes people sick.

It is possibly 'one' of the factors that make people sick but that doesn't preclude the other very common issues that make people sick like artificial locomotion. After giving ~100 VR demos I haven't found anybody that gets sick with non-artificial locomotion games (including my mum who can't even watch a video game on a TV without getting sick).

That's why I look forward to StarVR [1] from Starbreeze Studios (or what the competition will do if it hit's the market. I'm a bit skeptical if Starbreeze has enough resources), since it has a field of view of 210 degrees, which allows the user to look to the corner of the eye and see stuff like in real life.

It seems in the next prototype they want to also integrate eye tracking, which hopefully works great together. The downside of the current prototype seems to be the weight, only 60 hz refresh and hard to make optics which have a tight sweet spot, and if you are not in the sweet spot it looks bad.

[1] http://www.starvr.com/

FOVE is the first VR start up working on eye tracking for this very reason. http://www.getfove.com

Imo the biggest problem with VR is lack of controls. All of the games I've seen (that aren't driving/flying simulators) use same "teleportation" movement gimmick with some slight leaning/ducking or they just have you stand in place while stuff happens around you.

There's also less to render, if you only render where you're looking. You can render lod proportional to retina receptor density.

Can be done with eye-tracking, or contact lenses.

Big the big problem is latency: it has to be even quicker than what current VR headsets can get away with, especially for saccades.

We may have direct visual neural implants before we have that.

> You can read a sentence while shaking your and twisting your head. Try that with an oculus (or any other headset).

Actually, with the issues with small fonts (and subpixel gaps in the DK2) in Elite: Dangerous - one suggested fix is shaking your head from side to side for easier reading of text, and it works fine?

It's more like you're wearing goggles than "fixed eyes", you can still look around, your field of view is just restricted because of hardware limitation at the moment.

This also is not a common reason for motion sickness in VR, that's just incorrect.

> You can read a sentence while shaking your and twisting your head.

I tried to read the rest of your post while doing that, and it gave me motion sickness. I'm not usually susceptible to it so I guess that says something. Maybe reality is actually as bad as VR?

> The minecraft hack that snapped view changes to 15 degree increments

Link? Haven't heard of this, and I sometimes have nausea when playing.