Support NeoGAF

[Larxia]

I would love VR headsets where you can dynamically adjust the focal lenght, like you would do with binoculars, so you can wear it without glasses no matter your eyesight, and without having to replace the lenses.

Actually, I'm so nearsighted that I would be able to see properly with no glasses in a VR headset if there were not already lenses pushing the image further away. The thing with the additionnal lenses is that it seems kind of dumb, from what I understand it's another pair of lenses inside that act like glasses, so... you have lenses in the headset that push the image further away (which is needed for people with a regular eyesight, it's to make that the image is about 2 meters away instead of a few centimeters from your face), and then you have lenses that bring it closer, acting as glasses, which must created quite a lot of distortion / chromatic aberration, and eyestrain.

I never tried it so I can't really say, but I just wish we could tweak the focal lenght easily on vr headsets so stuff like this wouldn't be an issue.

 

[coffinbirth]

Lots of sims enthusiasts would love this headset. Plenty of high end driving and flying VR games

That warrants a $700-$2K upgrade over headsets from the last four or five years? Are any of the games you speak of from more recently than that, even?

For people that have never had a PCVR headset, I can see the appeal, but otherwise...

 

[Danknugz]

pimax feels like a very fly by night company and i would not trust my money there

 

[bitbydeath]

Wake me up when there's new games to go along with any of these new headsets. I haven't seen hide nor hair of anything that warrants hardware upgrades going on YEARS now.

VR needs more people to buy-in to create those games; yet more games need to be created to get people to buy-in.

 

[Zug]

Asgards Wrath was a AAA VR game. I haven’t started the second game yet, but from what I hear, it’s an even bigger/better game.

Apologies, I should have precised : AAA VR games worthy of interest.
I really love VR but let's be honest, actual content has been lacking for years now, it didn't work out and the PSVR2 failure will be the nail in the coffin of this tech iteration, that we might see appear again in a decade ('member the Virtual boy and the VR headsets from the early 2000s).
The fact that Cyberpunk and Starfield didn't even get a VR mode, despite being the perfect genre for VR, with all the hype and budget they got, shows a lot.

 

[cash_longfellow]

Until someone can solve the problem of the lack of immersion for people with non-correctable amblyopia on VR, I’m out. Every game has legitimately looked like a blurry super low resolution game, with literally no feel of being in the game to me, on every headset I’ve tried (including a sneak peak I had of the Apple set).

 

[poppabk]

No, they just tried to save on cost by excluding it. I can't believe you're trying to treat dynamic foveated rendering as the one with "limitations". Fixed is the 'limited' tech. Having eye tracking means you can do whatever fixed foveated rendering does.

Except use upscaling tech like DLSS 2 and frame interpolation tech like DLSS3.

 

[Deerock71]

WHAT A BOUTIQUE PRODUCT!

 

[Reallink]

I wouldn't wish a Pimax product on my worst enemies, they're probably the shadiest company most people are ever likely to encounter in modern regulated markets. They've used early buyers as full price testers for pre-alpha hardware with every piece of hardware they've released, their required software is notoriously worse than terrible (routinely bricks headsets on firmware update, a great many games are completely non-functional on Pimax hardware, and it's basically a CCP rootkit), and they completely fleeced their Kickstarter backers failing to deliver every single stretch goal, and outright stole the $100 they offered people to "downgrade" models. You'd be better off lighting your money on fire than giving it to Pimax.

 

[poppabk]

Apologies, I should have precised : AAA VR games worthy of interest.
I really love VR but let's be honest, actual content has been lacking for years now, it didn't work out and the PSVR2 failure will be the nail in the coffin of this tech iteration, that we might see appear again in a decade ('member the Virtual boy and the VR headsets from the early 2000s).
The fact that Cyberpunk and Starfield didn't even get a VR mode, despite being the perfect genre for VR, with all the hype and budget they got, shows a lot.

Starfield is because MS hates VR. They don't even put PCVR games they own on gamepass and hide the VR functions of other games on gamepass. I wouldn't be surprised if they even drag their feet on VR for FS2024. VR isn't going anywhere in the PC space though, too many sim players for that to happen.

 

[Three]

I think nemiroff point is that we can not "set lens distortion aside". Much of the benefit of rendering at higher resolutions, outside of the sweet spot, are lost in real world use cases. Obviously, future developments will mitigate this fact, and eye tracking will become more beneficial than it is with most currently available headsets. My limited use of PSVR2 drove this point home. The sweet spot is so small that eye tracking loses many of its benefits (outside of potential streaming optimization).

It's mostly gibberish. The edge clarity is not bad enough to mitigate the heavy edge pixelation seen when using FFR especially at higher level. Form factor is not an issue because its a simple camera and micro leds. It would be like saying your phone shouldn't have a front facing camera because it affects weight or form factor. The changes are minimal and the form factor is near identical to the non-budget Pimax version which has eyetracking DFR. The Pimax Light still using heavy and large glass lenses it's clear this isn't about form factor or lack of edge clarity at all. It's BOM cost driven and most in the VR community agree. That's why the Light was designed in the first place, to push affordability but maintain the clarity and sweet spot. What's become standard or 'worth it' has just been decided by lower cost headsets (which don't have eye tracking) becoming mainstream and so supporting it less beneficial for software. It's not because eye tracking is not currently beneficial or noticeable.

 

[Three]

Except use upscaling tech like DLSS 2 and frame interpolation tech like DLSS3.

Frame interpolation tech like DLSS3 isn't suited for VR anyway due to latency. It's much better to use something like ASW/ATW. As for upscaling are you suggesting FFR is better suited for upscaling than DFR? Because that's not really true. The focal point can be changed and it makes no difference to DLSS. Here's a free paper investigating DLSS with foveated rendering.

 

[cash_longfellow]

I would love VR headsets where you can dynamically adjust the focal lenght, like you would do with binoculars, so you can wear it without glasses no matter your eyesight, and without having to replace the lenses.

Actually, I'm so nearsighted that I would be able to see properly with no glasses in a VR headset if there were not already lenses pushing the image further away. The thing with the additionnal lenses is that it seems kind of dumb, from what I understand it's another pair of lenses inside that act like glasses, so... you have lenses in the headset that push the image further away (which is needed for people with a regular eyesight, it's to make that the image is about 2 meters away instead of a few centimeters from your face), and then you have lenses that bring it closer, acting as glasses, which must created quite a lot of distortion / chromatic aberration, and eyestrain.

I never tried it so I can't really say, but I just wish we could tweak the focal lenght easily on vr headsets so stuff like this wouldn't be an issue.

This, and non-compensation for other vision problems is exactly why VR isn’t close to becoming mainstream.

 

[Dacvak]

Pretty neat. Pimax makes some pretty decent hardware. I might check out their roadshow in DC

 

[Three]

I would love VR headsets where you can dynamically adjust the focal lenght, like you would do with binoculars, so you can wear it without glasses no matter your eyesight, and without having to replace the lenses.

The ultimate goal is varifocal that adjusts automatically.



It would also help with Vergence-accommodation conflict meaning less eye strain and headache.

 

[MarkMe2525]

It's mostly gibberish. The edge clarity is not bad enough to mitigate the heavy edge pixelation seen when using FFR especially at higher level. Form factor is not an issue because its a simple camera and micro leds. It would be like saying your phone shouldn't have a front facing camera because it affects weight or form factor. The changes are minimal and the form factor is near identical to the non-budget Pimax version which has eyetracking DFR. The Pimax Light still using heavy and large glass lenses it's clear this isn't about form factor or lack of edge clarity at all. It's BOM cost driven and most in the VR community agree. That's why the Light was designed in the first place, to push affordability but maintain the clarity and sweet spot. What's become standard or 'worth it' has just been decided by lower cost headsets (which don't have eye tracking) becoming mainstream and so supporting it less beneficial for software. It's not because eye tracking is not currently beneficial or noticeable.

"It's not because eye tracking is not currently beneficial or noticeable."

Why resort to a strawman, as neither I nor the user I was referring to claimed it wasn't beneficial nor noticeable. As I stated, much of the benefit is lost due to distortion, not all or even most. There is no need to exaggerate my claims.

I can't speak for everyone in here, but I certainly am not arguing that DFR was removed for any other reason than weight and cost reduction. Matter of fact, I wasn't arguing any particular stance. I just thought I could add clarity to another users stance, that many seemed to misinterpreted or worse, misrepresent.

 

[Three]

"It's not because eye tracking is not currently beneficial or noticeable."

Why resort to a strawman, as neither I nor the user I was referring to claimed it wasn't beneficial nor noticeable. As I stated, much of the benefit is lost due to distortion, not all or even most. There is no need to exaggerate my claims.

I didn’t say you said this but when somebody else is saying that Eye tracking is for nothing but performance benefits (pure bullshit) they're suggesting that DFR has no noticeable visual benefit either. There is no strawman, that's exactly what they're saying:

"eye tracking is used so thet you dont have to move your head to get a constant clear picture,"

And now my head hurts..

No; that's not what we have eye-tracking for in VR headsets, I promise.. It's just not it.

Eye-tracking is purely for performance, period

 

[MarkMe2525]

I didn’t say you said this but when somebody else is saying that Eye tracking is for nothing but performance benefits (pure bullshit) they're suggesting that DFR has no noticeable visual benefit either. There is no strawman, that's exactly what they're saying:

I'll buy that, as I wasn't entirely sure if you were referring to my comment or the other guys. Tbh, I think y'all were talking past each other, as I don't take his statement to mean it isn't beneficial nor noticeable.

 

[alucard0712_rus]

"Fixed" Foveated.....yah dat aint it.

FR is unimportant on PC. It's mainly used to save on performance and the people who buying those things ain't using internal graphics.

 

[nemiroff]

I think nemiroff point is that we can not "set lens distortion aside". Much of the benefit of rendering at higher resolutions, outside of the sweet spot, are lost in real world use cases. Obviously, future developments will mitigate this fact, and eye tracking will become more beneficial than it is with most currently available headsets. My limited use of PSVR2 drove this point home. The sweet spot is so small that eye tracking loses many of its benefits (outside of potential streaming optimization).

That’s correct. The lens blur is what it is, and you can’t fix it with FFR or DFR. However, you can use it for easy performance gains. Due to the inherent lens blur, FFR is visually transparent to the user (if you follow the developer guidelines about foveation maps).

Now, the irony is that the Pimax Light has superior aspheric glass lenses (for optical performance) compared to, for example, the Bigscreen Beyond’s plastic pancake lenses (due to design and form factor). Therefore, eyetracking would be a good fit for the Pimax Light (as an option perhaps).

(And no, with the Pimax Light’s aspheric lenses, you do not have to move your head to keep at the center of clarity. The sweet spot in these headsets is much wider than, for example, in the PSVR2 (and even some PCVR headsets, for that matter))

Let’s take a step back and remember what this discussion started with: a knee-jerk dumb notion about how the lack of DFR would make the Pimax Light worthless.

I later tried to explain that historically FFR has had a larger benefit from zero than the relative performance difference between FFR and DFR. To illustrate, this example was inserted:

If we look at the numbers, we’ll see that with a level 1 foveation map and a render target of 1.5 (yes, I’m conveniently picking this scenario to drive the point home), the flat performance gain for FFR is 34%, while DFR yields 37%.

= 34% vs 4%

So, my original point was simply that there's more nuance and context to this. Especially for headsets where form factor and design are more important than a single-digit performance gain.

 

[alucard0712_rus]

That’s correct. The lens blur is what it is, and you can’t fix it with FFR or DFR. However, you can use it for easy performance gains. Due to the inherent lens blur, FFR is visually transparent to the user (if you follow the developer guidelines about foveation maps).

Now, the irony is that the Pimax Light has superior aspheric glass lenses (for optical performance) compared to, for example, the Bigscreen Beyond’s plastic pancake lenses (due to design and form factor). Therefore, eyetracking would be a good fit for the Pimax Light (as an option perhaps).

(And no, with the Pimax Light’s aspheric lenses, you do not have to move your head to keep at the center of clarity. The sweet spot in these headsets is much wider than, for example, in the PSVR2 (and even some PCVR headsets, for that matter))

Let’s take a step back and remember what this discussion started with: a knee-jerk dumb notion about how the lack of DFR would make the Pimax Light worthless.

I later tried to explain that historically FFR has had a larger benefit from zero than the relative performance difference between FFR and DFR. To illustrate, this example was inserted:

If we look at the numbers, we’ll see that with a level 1 foveation map and a render target of 1.5 (yes, I’m conveniently picking this scenario to drive the point home), the flat performance gain for FFR is 34%, while DFR yields 37%.

= 34% vs 4%

So, my original point was simply that there's more nuance and context to this. Especially for headsets where form factor and design are more important than a single-digit performance gain.

While this is true (as Carmack said long time ago) DFR is harder to spot, it's better quality as I've seen on PSVR2.

 

[crackajack]

Until someone can solve the problem of the lack of immersion for people with non-correctable amblyopia on VR, I’m out. Every game has legitimately looked like a blurry super low resolution game, with literally no feel of being in the game to me, on every headset I’ve tried (including a sneak peak I had of the Apple set).

I don't know what amblyopia is, but I guess anyone, who really cannot see VR, is better off with something like TrackIR if those are still a thing.

 

[cash_longfellow]

I don't know what amblyopia is, but I guess anyone, who really cannot see VR, is better off with something like TrackIR if those are still a thing.

It’s essentially lazy eye where my vision is super blurry in my left eye and my depth perception is janky. All these devs are saying the headsets should appear just like I see in real life. Definitely not the case…it just looks like I’m watching a tiny tv screen with terrible resolution.

 

[poppabk]

Frame interpolation tech like DLSS3 isn't suited for VR anyway due to latency. It's much better to use something like ASW/ATW. As for upscaling are you suggesting FFR is better suited for upscaling than DFR? Because that's not really true. The focal point can be changed and it makes no difference to DLSS. Here's a free paper investigating DLSS with foveated rendering.

That paper indicates that they had to include a TAA pass to get DLSS to work with foveated rendering and there was still an image quality cost.
I'm suggesting that currently you may get the same performance boost with minimal or no foveated rendering and using upscaling techniques and maybe frame gen, I have AMD so haven't had a chance to try it out personally - but ASW often looks completely shit in sims.

 

[FunkMiller]

Asgards Wrath was a AAA VR game. I haven’t started the second game yet, but from what I hear, it’s an even bigger/better game.

I tried both Asgard's Wrath games, and neither has the fidelity, environmental detail, or the narrative of Alyx. Both feel very cartoony in execution - especially on the lesser Quest 2 hardware in the second game's case. They're better than the typical short, pointless experience for sure, but neither grab you or immerse you the way Alyx does. I actually found Lone Echo to be a more immersive AAA feeling game than Asgard's - but still not on the same level as Valve's game.

 

[Three]

That paper indicates that they had to include a TAA pass to get DLSS to work with foveated rendering and there was still an image quality cost.

Yeah because DLSS' temporal instability is compounded but i was referring more to section 2.2:

However, our focus point for the eyes is freely movable in real time without any issues. This means that we could integrate our implementation with eye tracking hardware to make it adapt to the user without any obvious problems.

Because I suspected you were saying eyetracked DFR was limited compared to FFR.

I'm suggesting that currently you may get the same performance boost with minimal or no foveated rendering and using upscaling techniques and maybe frame gen, I have AMD so haven't had a chance to try it out personally - but ASW often looks completely shit in sims.

Framegen for VR has the big hurdle of latency being more noticeable/problematic in VR at the moment but maybe eventually.

DLSS upscaling with the same performance boost would be nice but FR is built with VR specifically to have no visible loss in quality or artifacts. With eye tracking can you imagine the possibilities? Eye tracked DoF. Imagine UE5 with foveated rendering. Where mesh data is tied to resolution and could be reduced based on where you look.

 

[crackajack]

it just looks like I’m watching a tiny tv screen with terrible resolution.

My vision ain't perfect either, but that's probably the best for now for everyone. The fov isn't there yet, a big junk of the peripheral view is just black nothing, making everything tunnel vision devices, and the resolution takes a massive hit when looking at those pixels that close. Probably a gen down. VR is probably the only area where 8k and even beyond would be beneficial, incl reduced gaps between the subpixels, while it won't have a significant effect on any TV below, I don't know, 100inch at normal viewing distances.

 

[poppabk]

Because I suspected you were saying eyetracked DFR was limited compared to FFR.

What I was trying to say was that quality upscaling that uses information from previous frames in the reconstruction is hard when the previous frames included low resolution in the part you now need to be high resolution and a small portion of high resolution in the area that is now low resolution so currently I don't think they are easily compatible. Is Nvidia still even supporting VRSS?

 

[Three]

What I was trying to say was that quality upscaling that uses information from previous frames in the reconstruction is hard when the previous frames included low resolution in the part you now need to be high resolution and a small portion of high resolution in the area that is now low resolution so currently I don't think they are easily compatible.

They're compatible according to the paper but with DLSS it does have its a slight hit in temporal quality. I think FR is a bigger plus for VR even if DLSS wasn't possible with it.

Is Nvidia still even supporting VRSS?

AFAIK they haven't dropped support but like with most things VR it isn't that widely used or covered.

 

[nemiroff]

It’s essentially lazy eye where my vision is super blurry in my left eye and my depth perception is janky. All these devs are saying the headsets should appear just like I see in real life. Definitely not the case…it just looks like I’m watching a tiny tv screen with terrible resolution.

I have no problem with your impression, but I'm just curious about what VR headsets in combination with what apps you've tried. For most people vision impairment shouldn't make much of a fundamental difference in VR. And that's very much my own experience as well, having four different real medical issues ruining some of my vision.

 

[Magic Carpet]

I tried out a Pimax at a friends house Once. It took so long to set up that I Just wrote it off as dead tech.
My friend had several different VR helmets and had to set each up individually. A disaster of wires.
He was very impressed with my quest 3 and zero wires. He’s keeping the pimax though.

 

[cash_longfellow]

I have no problem with your impression, but I'm just curious about what VR headsets in combination with what apps you've tried. For most people vision impairment shouldn't make much of a fundamental difference in VR. And that's very much my own experience as well, having four different real medical issues ruining some of my vision.

Valve Index, Meta Quest 3, HTC Vive, both PSVR, Apple Vision Pro. Plus some older cheaper headsets. None are any better than the other one’s either, and none come remotely close to how I see the real world at all.

 

[nemiroff]

Edit: Eh, double post for some reason...

 

[nemiroff]

I tried out a Pimax at a friends house Once. It took so long to set up that I Just wrote it off as dead tech.
My friend had several different VR helmets and had to set each up individually. A disaster of wires.
He was very impressed with my quest 3 and zero wires. He’s keeping the pimax though.

It's worth noting that the Pimax Crystal (not Light AFAIK) is designed to be ready for an optional wireless connection. It already has a battery and a Snapdragon chip and will function as a wireless PCVR headset with the airlink module.

Why Does The Pimax Crystal Have An XR2 Chip, Even For PCVR? | Pimax

Online several users wondered why the Pimax Crystal comes with a Qualcomm Snapdragon XR2 chip, because if you would just want to use it for PCVR, why use this

pimax.com

60G Airlink Demo & Current Status | Pimax

The Pimax Crystal from the start was designed and built to provide the highest clarity as well as wireless VR. This is also why it has a battery and a

pimax.com

 

[cyberheater]

Apparently the local dimming version is going to drop to $799

 

[THE DUCK]

In using the quest 3 earlier today, I was wishing for 3 things:

1. Lighter weight
2. Higher resolution panels
3. Wider fov

The binocular vision was definately bugging me today. Yet the 100" or so youtube screen I had floating in my living room was better than expected. I can just imagine just how good the apple headset with the higher res oled screens would be. (Just too expensive)

20 years from now more people will just put on glasses for a superior display than a tv, just not here yet. It will help usher in better vr.

 

[evanft]

Any headset over $500 that doesn't including real time eye-tracking with dynamic foveated rendering is DOA, IMO.

 

[nemiroff]

Any headset over $500 that doesn't including real time eye-tracking with dynamic foveated rendering is DOA, IMO.

I don't think so. This headset is most likely on its way to sell out its pre-orders according to VR flight sim guy (he buys and test pretty much all VR headsets out there). And also, he's saying that in f.ex. MSFS, eyetracked foveated rendering only yields 3fps more than FFR. I mean, you're entitled to your opinion, but the blanket dismissal is like throwing the baby out with the bathwater. I'd f.ex. take the Crystal Light's aspheric glass lenses (or f.ex. the Bigscreen Beyond's micro-OLEDs) any day over ET in an unbalanced headset.