I can now confirm, the Pimax Vision 8kX, has passed a reasonably credible Snellen Test at 20/20 readability. This is a very significant milestone for VR.
Basically 20/20 is ‘typical’ human vision. If you can read the 20/20 line of text on a correctly sized Snellen chart, standing at the correct distance, then you have normal human vision.
However, there are a few pitfalls.
Dim, standardized illumination may be used for eye exams, which is beyond the scope of this experiment.
At 20/20, discerning a P from an F, was marginal, which as I understand it, is a sub-clinical finding.
Blurring of the text was qualitatively very different from the kind of blurring expected from human vision defects. Small text tended to become subjectively more grey.
Headset IPD, sweet spot alignment, etc, were a little more carefully managed than usually practical.
Total SR was pushed rather high, >2x. Going beyond 1.5x was marginally different at most.
VirtualDesktop was conservatively used with distance/FOV estimations. Due to problems with dedicated VR headset test apps.
I welcome the opportunity to help others replicate these findings as more Pimax Vision 8kX units become available.
That’s really exciting. Gonna be perfect for flight simming and just about everything else. Gonna be a great fit for the 3080ti. I just hope the distortion is low or not noticable. Otherwise, seems perfecto.
It’s fantastic to see results given along with all known caveats called out as they should be. Good stuff.
@mirage335 this isn’t exactly 8K X specific, but might you be able to design a standardized FOV test for comparing HMD’s against one another that would be decent enough for relative comparison? I’m thinking in terms of a rigid mechanical rig to mount the HMD etc.
Edit: Just to be clear for everyone reading this, I’m talking about being able to compare HMD’s against one another in a non-subjective way. I’m sure trying to give an “absolute” figure of the FOV for the “average” human head would be even more difficult.
I realize the idea is fraught with perils, but I really feel like any serious effort made to come up with something would be a whole lot better than the current situation. Basically misleading numbers are all over the place.
I imagine attempting, with very limited resources, to make the rigid mount approximate a typical Human head might be tricky. Also there’s the question of how tight the strap needs to be, and the resulting compression of the facial interface. Using a head mannequin of some sort might be a way to start, but I have no idea what the camera rig would have to be like to do the job.
I just pointed out in another thread that I recently tested the Index and Reverb side by side. Many websites list both of those HMD’s as having 115 degree FOV which is ludicrous.
I know you’ve got plenty on your plate as it is, it’s just a thought.
An interesting validation check could be: Can a person with 20/20 vision still read the 20/20 test line in VR? And can a person with let’s say 20/25 vision not read it? If the latter still can then something is odd with the setup.
Wow, I was not sure this would get much attention.
By the way, there is one more pitfall I forgot to mention, which I will edit into the original post. AFAIK, eye exams are best done under a standard amount of illumination. IIRC, these tests are actually fairly dim. Brighter things can be easier to see.
@x15adam I am still expecting the 3080Ti to show ~20-30% improvement, and we can already achieve decent results with the 2080Ti. Not expecting it to be a game changer. What would be a game changer would be to get those flight sims to support SLi and even a tiny bit more CPU multithreadding.
@TripRodriguez
There exist FOV tests. Basically, you stand in a certain place, look forward, and without moving your eyes (covert shift of attention), determine which lines are still visible, step back, and read the numbers. Horizontal FOV in my experiments has been more or less consistent with Pimax’s marketing claims. I don’t consider the exact FOV numbers all that remarkable. So what if I can see 180deg FOV instead of 175deg? All I really care about from FOV is that it is not the ~100deg that makes me feel even more like a horse wearing blinders.
This is the link to one of the FOV test apps. Keep in mind the Snellen chart does not seem to have absolute calibration information, and the rendered resolution may be low. But just for FOV testing, it should work.
@NoamLoop
Since my test procedure required the Snellen chart to be visible on my desktop monitor as well as the headset, at the same angular scale, I can attest that it was consistent with my slightly better than 20/20 vision at that brightness.
@mirage335 What I’m looking for is not the FOV of visible “game world”,but how many degrees of the user’s ‘real life’ vision is encompassed by the display area vs. darkness.
I wouldn’t expect this to be a 1:1 ratio, but perhaps I’m incorrect. Assuming it is not, I’d like to compare the “hardware FOV” of the various headsets against one another.
I think that the real life/hardware FOV is more critical to immersion, and based on FOV settings in flat screen games it also seems to me you should be able to have the “rendered FOV” at least 10% less than the “real life” FOV which would create a more immersive experience with less rendering power. I think 10% is probably a very conservative number.
I also realize this might be a horrible idea, causing fish bowl effects and inducing motion sickness. Perhaps the margin is much smaller than I expect.
@Cdaked
That is exactly what I was referring to when I said ‘due to problems with dedicated VR headset test apps’. I have tried a number of workarounds, but I cannot get the Snellen chart in that app to render at full Pimax Vision 8kX 1.5x Total SR resolution. Consequently, the chart 20/20 line in that app becomes unreadable at rather incredibly short distances.
