Tech Talk #6 : The Key Elements of VR HMD Clarity

Hey,

We’ll talk about essential components that contribute to increase clarity in VR this week.

While reading any VR article, we frequently come across three terms: Resolution, PPI (pixels per inch), and PPD (pixels per degree).

Who is the key figure in all of this?

Aside from that, different panels have different screen techs and different sub-pixel layouts,what sort of display panel do you like for a better VR headset?
What’s better, the LCD , OLED or Micro OLED/LED?

Do you believe the lens design has anything to do with the clarity?

What hardware specifications would you advertise to the world if you were to build the ultimate VR headset?

**Feel free to leave a comment or suggest a topic for discussion.**we will select one lucky user to receive a SteamGift Card!

Regards.

5 Likes

Resolutions of next gen VR headsets are going to be very high with displays hitting the market right now with uOLED displays hitting PPI numbers 3500 and LCD displays hitting around 1764 PPI and OLED hitting around 800PPI.

uOLED: But these Micro OLED displays are limited to size right now with a size limit of 1.1inchs a display size not suitable for a Pimax HMD. Currently there is only one company that is a exception to this size rule. They have manged to merge two uOLED displays to form a 2.1inch display.

OLED: Unlike uOLED regular OLED displays have very low PPI with current OLED technology we are hitting about 800PPI but this is considered the latest and greatest with most likely high cost associated with it. But a 3840x2160 5.5inch OLED could be made using this tech which could slot right into Current Pimax Vision HMDs.

LCD/CLPL: LCD displays are not limited by size but other features such as brightness, contrast, heat, local dimming zones, persistence/pixel response times. But currently is not favourited among the community due to some of these downsides. I don’t personally see an issue with the technology but currently don’t want it in my VR headsets I want to go back to a form of OLED. I would 100% prefer a 4K OLED display at 100% coverage of the DCPI3 in my Pimax HMD then a 6K LCD display even if it covers the 90% of the BT2020 colour space with local dimming. there would be faster pixel response times on a OLED display with per pixel dimming, these characteristics would be more preferable in my opinion. It gives more immersion to the content.

All three technology’s have to be paired with a lens for HMDs. Some requiring certain lenses to get the most out of the display. A uOLED needs to cover the users FOV which needs heavy magnification that’s where Pancake optics are used. OLED doesn’t have a high PPI and brightness so the displays are going to remain fairly large meaning aspheric lenses are the best option. LCD is highly versatile meaning operation with both lens types are possible but the more suited lens would be aspheric due to contrast and brightness issues with really high brightness displays needed for pancake optics. But I am super curious about Liquid Crystal lenses which should make variable focus possible for VR. It will be interesting to see what Pimax is going to use the Crystal name for but I guess we will see in a few days.

OLED, but this is user preference

100% affects the clarity. The lens changes how the light hits the eye. Artifacts from lenses change how people like their headset such as godrays and chromatic aberration

And to answer this question you would first need to know what the pros and cons are to the ultimate VR headset to set it apart in marketing it.

1 Like

Heh, personally I only associate the word “clarity” specifically with just how see-through the lens is – i.e. no turbidity and no otherwise diffusing properties. :stuck_out_tongue:

2 Likes

I believe the varjo aero has debunked the resolution is all debate. yes you need a good resolution for clarity but there is a limit. I believe clarity, right now, is a mixture of everything but I think both contrast and lens design are more important than most everything else. with good lenses you get better clarity and a larger clarity window. also, contrast is important for making things look 3d and sharper. oled will always be king here.

so if i could make the clearest display in the world id go micro oled for the immediate center of the picture, with regular oled around the outside closer to the density of the 8k we have already. preferably you would want both to have quantum dot technology for color. make sure the lens is resistant to scratches, and can be adjustable to the individual’s ipd. this also means the lenses need to have the ability for depth adjustments from the person’s face because some people have weird faces like me.

2 Likes

I agree, the contrast is vital for the clarity. The cons of OLED is the limit pixel density.

1 Like

Resolution, PPI, PPD…

All of these measurements can be misleading in terms of actual clarity. What matters at the end of the day is actual resolving power in a real virtual scene.

There’s a world in VRChat which lets you take an eye exam. It’s the kind where you’re trying to read letters accurately, and each line of letters is smaller and smaller. I think this kind of approach is the right way to measure clarity because it measures all factors end to end in the pipeline that will affect your vision. Doing it the same way it’s done in RL.

This test doesn’t measure aspects like color accuracy or contrast range. But I would argue that while these are important factors, too, I don’t think they fall under “clarity”. Or if they do, then clarity is being used as a catch-all term for anything which affects the quality of the video image. Whereas I would see clarity more specifically as the appearance of sharpness and being able to truly discern fine details in the image.

Uniformity of Clarity

Uniformity is very important to clarity, but I don’t think I’ve seen it discussed. In terms of VR immersion, what we really want is to get clarity off the table as something that the user is paying attention to. Rather like you tend to only notice glasses you’re wearing when there’s a smudge on them. Otherwise, as long as they’re working correctly, eye glasses tend to vanish from your perception while you’re paying attention to the world around you instead.

It draws the user’s attention to lack of clarity when parts of their view are sharper than others. Most typically, the center of the view has much higher clarity than the rest. I think this is to be avoided. The user will get used to whatever level of clarity the VR headset is able to provide as long as its consistent, and it will come to feel normal and ignorable. But when the central part of the display is so noticeably sharper than the rest, the user is constantly reminded about how other parts of their vision are less clear by comparison.

In terms of VR immersion, I think you’d be better off with a lower clarity which is uniform across the entire display than you would be with a higher clarity only in the center. This is for much the same reason that a lower but steady frame rate is generally better than a higher but erratic frame rate.

So, in my opinion, the ideal VR headset for consumers would have a wide FOV that covers their whole vision and does so with uniform clarity where ever they look (which does allow for eye tracking and rendering foviation). Under these conditions the specific level of clarity its able to achieve becomes less critical. Obviously sharper is better. But I think the first goal is just to reach wide and uniform, and then after that you go for incremental improvements in clarity in successive generations.

1 Like

I would also love to see improved motion clearity - fast moving “pixels” e.g. when doing turning motions are still to blurry today - OLED, like the QDOLED might be a way out off that. And larger sweet spots so that a view mm dont cause a totaly different experience.

1 Like

Lense design is really important. Fresnel lenses do not have as large a sweet spot, and aspheric lenses can have much distortion.

A hybrid fresnel aspheric lend is the best solution, or aspheric lens with software to correct distortion for a distortion free experience in wide fov like StarVR and XTAL

PPD is most useful for VR, PPI is how many pixels in relation to the screen size and widrh, however PPD is the resolution divided by the horizontal or vertical FOV assuming there is 3 subpixels per pixel ( a red, a green and a blue one) and no white subpixels, and is much more useful of a measurement for VR compared to PPI

QLED and QD-OLED (quantum dot oled) are superior to MicroOLED, OLED and LCD but Micro OLED is superior to OLED which is superior to LCD in terms of colour, however OLED often has pentile displays (2 subpixels per pixels rather than 3)

  1. It would have display port 2.0 cable for 14k.

  2. 14k X 6k per eye (custom display) with 60-90 PPD

  3. Full RGB subpixel displays (no pentile) with 3 subpixels per pixel, a red subpixek, a green subpixel and a blue subpixel, with no white subpixels.

  4. HDR color range (qled or quantum dot)

  5. 240 horizontal fov

  6. 135 vertical fov

  7. 160hz refresh rate

  8. Leap motion Handtracking

  9. Tobii eye tracking

  10. Dynamic Foveated rendering/foveated transport so only 5% of performance is required to render

  11. Eye relief (bring lens closer or further away from headset)

  12. IPD adjustment physical dial

  13. Virtual IPD offset

  14. Proximity sensor to allow headset to stay plugged in but in sleep model

  15. Lighthouse tracking

  16. 200g-500g in weight maximum

  17. Super compact and not hanging far off the face like Arpara

  18. PSVR Halo headstrap and comfort kit

  19. No distortion at all over whole picture

  20. Very large sweet spot with no blur when looking at the centre of the lenses

  21. Aspheric or hybrid aspheric fresnel lenses.

  22. 6-10m long singular input cable

(not needed but also excellent additions)

  1. Wireless capability

  2. Mouth tracking like decagear

14000 horizontal pixels per eye × 6000 vertical pixels per eye × 2 eyes×160hz is 26,880,000,000 (27 billion) pixels per second. Assuming foveated rendering allows for the image to be rendered with only 10% of the performance requirements for the entire displays, that is only double the 8KX performance requirements of 1.3 billion pixels per second for the 8KX and is feasible to do on next generation graphics cards if the display can be manufactured. Samsung has manufactured 10k PPI displays so it is possible
Samsung Has Created the World's First 10,000 PPI Screen - Despatch.

For these three terms I would say:

  • Resolution is the general term how much of a detail we can see. For an easy definition I would define Resolution by PPD.
  • PPD is probably the most convenient VR equivalent to what we call resolution on regular flat panels.
  • PPI (as in Pixels Per Inch resolution of the display) is useless for VR application, because it does not say anything about what the user actually see - as it depends also on the optical properties of the lens.

Lens design is actually the only factor which defines clarity. If the lens cannot properly focus the image, the user will not be able to benefit from the PPD.
There is however unfortunately no accepted term to describe the clarity of the headset. It is somehow understood/accepted that the area in focus is usually not covering the whole FOV, but how much of the FOV it is is usually a surprise coming with each new headset design.

3 Likes

Another post discussing the next-gen clarity
https://community.openmr.ai/t/apple-8k-8k-per-eye-small-fov-vs-pimax-12k-6k-per-eye-wide-fov-discussion/38654

1 Like

More clarity in VR, especially past the ~20/20 readability of the Pimax Vision 8kX, is about improving both comfort for long-term users, and learning curve for newer users briefly using the headset.

From the Pimax Vision 8kX, for what the image quality looks like, we are just at the point of diminishing returns. Much higher resolution and much less blur will not look so much different, because we close to average human vision.

Yet, being only as good as human vision does still cause some slight eyestrain. Better lenses will reduce blur (especially towards the edges), better resolution will improve the speed of readability. Both better lenses and higher resolution are still an important end goal for VR.

For long-term users, this may be a difference of 10hours/day instead of 12-18hours/day, and some minor aches from more head movement. From new users, I sometimes hear complaints, sometimes due to expectations of VR optical clarity already being exactly equal to ‘IRL’, sometimes due to impatience with the time taken for a person’s own eyes to adjust.

I think the Pimax 12k, especially if the center area high resolution projector becomes available at any price, will help improve the amount of hours VR developers can work solely in the headset, and the number of users who can get into VR. That in turn, will improve the usability of software in VR. So this is very important.

On a side note, everything I can do that I seem to need to at this point - whether hardware/software design, or exploration and competitive gaming - is already ‘virtual’. So optical quality - both from resolution and lenses - is vastly more important to me than the sort of transparent displays used for ‘see-through’ AR. For getting to my chair and keyboard, camera passthrough is very appropriate. So I would also say that higher display resolution, better lenses, and lower distortion, that VR headsets should continue to have over transparent AR glasses, makes those VR headsets far more important to me.

1 Like

At some level the thing really needed for high resolutions is accomodating lenses.

Near sighted people actually can focus their eyes more closely to objects than normal sighted people can. It takes away from focussing in the distance but I used to be able to see the pixels on my phone if I held it near my face. Now I have my eyes lasered and discovered I lost this ability unfortunately.

2 Likes

Good lord, I want the 8KX you have. Mine is nowhere close to that metric.

2 Likes

Keep in mind that there is also the brain’s ability to fill in the gaps. So some will be able to score higher on an eye examine.

To compare one would also want to ensure settings are the same in terms of render and SS etc…

It would be interesting to revisit the Snellen(?) Test that the three and others used on the 8k and 5k+.


That being said I wouldn’t say the 8kX is 20/20 clarity as at that point in a headset would we really need a new model? Save for maybe some more miniaturization and new features that the older model lacked.

Even though the Varjos are very restricted FoV hmds it would be interesting to perform the Snellen test on it and the G2.

Exactly.

Even on a VR-3 with its 70ppd(?) focus screens, we are still nowhere clear to human eye 20/20 vision.

I love the 8KX for its FOV but don’t understand the hyperbole.

1 Like

In case everyone has apparently forgotten (rolls eyes)… I actually tested the 8kX early on with a Snellen chart. Pimax Vision 8kX does indeed have 20/20 readability to the point of being able to, admittedly just barely, correctly read the 20/20 line on a Snellen chart without already knowing.

Keep in mind ‘just barely’ is still a long way from actually normal comfortable 20/20 vision, and even then only for someone whose eyes perfectly fit the optics of the headset, with thousands of hours of adapting across a variety of such headsets.

Importantly though, that roughly 20/20 vision is enough that even some newbies can comfortably read everything on most simulated flight decks (without leaning in), and usually the HUD looks perfectly sharp. Pimax Vision 8kX is still unprecedented for giving at least some users adequate optical clarity, resolution, and FOV.

I do hope Pimax gets headset VR to the best possible result sooner rather than later though. Looking forward to the Pimax 12K .

No unfortunately it really doesn’t(unrolling eyes). But yes your adaptability may make it seem to have. That being said people with some levels of Myopia may have better Snellen results.


Otherwise G2 and Varjo would also have 20/20 due to higher res and ppd in viewport compared to 8kX.

Seem to have? Either the 20/20 line is readable or not, being calibrated at the same distance and size as IRL. I went through a lot of trouble on that, actually, and apparently, no one remembers.

It’s unambiguous. It’s readable. Pimax Vision 8kX is 20/20. This was announced and settled a while ago. I don’t know why there would be any new debate about it now.

Well your the only one it seems to believe so.

You could have saved yourself time and used existing tests like @knob2001 's RoV test suite.

I did those too. At the time, they were rendering at well below native resolution, instead of >1.4x supersampled.