Is VR dead?

  • Thread starter Deleted member 197115
  • Start date
  • Deleted member 197115

Did the G2 put you off VR ? Was the G9 with head tracking just that good?
Or was it a combination of both?
I don't think it's G2, it was an eye opening step up from Odyssey Plus in image quality. It's just even at best settings I can squeeze out of it with OCed 3080, it's a pale, degraded representation of what I can see on monitor. Especially in ACC with HDR on.
My field of view is completely covered, TrackIR allows me to an look at apexes and mirrors, I am getting 100+ fps with all settings on Epic and 110% resolution scale, after short period of adaptation getting same and even better times and racing feels just as immersive as in VR.
I am quite a pixel picker, so for me improved overall image quality makes it up completely for missing 3D depth.
In AC I can run same settings in VR and monitor with recommended 100% (200% native SS), but it still doesn't look as good as screen image. Some things just do not translate 1:1 and I am not sure if it's just lower resolution, certain effects are just not rendered the same way or are completely missing in VR.
I don't know what should happen in VR space to overcome current limitations, something very revolutionary for sure as jut jacking up resolutions does not work that well for this gen HW.

VR future driven by FB does not look too bright at the moment either.
And there is no one else having much incentive to burn in more money into this. Seems like big tech, and I mean tech, not social platforms or digital game store, just pretends that VR does not exist anymore. And why we have steep decline in new VR titles, because VR content unless it's racing or flying is plain stupid and 99.99% of games work much better on the screen.
Afraid our sim market is really small niche to make a noticeable dent. I was under impression that enterprise situation is much better but got my hands on some stats recently and to my surprise it's just a small fraction of game market.
 
Concerning TrackIR: Are there any such trackers that just use facial recognition instead of a dorky hat attachment? A solution specifically for drive/fly sim applications? That should be easy these days with a Raspberry Pi, etc.

Also, how is the experience? I presume you move your head to move the displayed image, but focus your eyes elsewhere. That quickly becomes transparently natural?
 
Last edited:
Concerning TrackIR: Are there any such trackers that just use facial recognition instead of a dorky hat attachment? A solution specifically for drive/fly sim applications? That should be easy these days with a Raspberry Pi, etc.

Also, how is the experience? I presume you move your head to move the displayed image, but focus your eyes elsewhere. That quickly becomes transparently natural?
There is something like that, using a webcam to track your face. I cant remember what it was called or if it is still popular but at the time it was never as fluid as tracking a light source/point as track ir.

It does become very natural very fast, its not requring large head movements so your eyes stay on the monitor. It seems your brain is easily reprogrammed to perform smaller head movements to get larger turns.

I used to love it for flying before VR and was good in racing when I had single screen. For some reason TIR when combined with triples for me didn't work well.
 
  • Deleted member 197115

Concerning TrackIR: Are there any such trackers that just use facial recognition instead of a dorky hat attachment? A solution specifically for drive/fly sim applications? That should be easy these days with a Raspberry Pi, etc.

Also, how is the experience? I presume you move your head to move the displayed image, but focus your eyes elsewhere. That quickly becomes transparently natural?
You can check this thread for options. https://www.racedepartment.com/threads/head-tracking-options.198356/
Tobii eye tracker is probably the closest to what you want, BUT it only works attached to the bottom of the monitor (area usually covered by steering wheel) and limited with monitor sizes it supports, 49" is a no no, I think it stops somewhere at 27" or something.

With ultra wide curved screen like G9 you already turn your head to see the sides of monitor, so your eyes will be pointed in that direction, you just need very small degree of extra movement to see a bit more beyond screen, effect is very subtle and natural.
And I know what you mean as I've used head tracking with other monitors before, and it was quite disorienting, not with G9 though.
 
Thanks for the answers.

I thought about it some more, and while it absolutely could be done, I think the frame rate and latency could be poor, at least on a Pi. It's a fun idea, though.

Since I don't use my G2 much at all, I was thinking about upgrading my display, especially to one with a faster frame rate, and ideas like this are interesting to think about.
 
I have the Samsung CRG90 49" ultra wide and actually prefer it to VR and triples , the ultra wide for me was a game changer in that the immersion went up 10 fold , I believe it's because of no bezels ( they really really mess up the immersion ) , having one big screen in front of you with no bezels makes all the difference , it's the perfect sim rig racing monitor.
I have a G2 from when they first came out and have used it probably about 10 times and haven't touched it for months , actually I should sell it as I don't use it anymore.

 
This is really not a Fresnel lens with zero god rays, etc..

The price tag doesn't bother me because this is the first of it's kind. I'm more excited that this level of technology exists now and that this is STEAM compatible. It just means if we wait, it will get here. My "guess" is that by the time this tech is more affordable, there will be a GPU that can push it comfortably, so... 2-3 years ?

 
Last edited:
I've read that DCS will soon release a new version completely ported to the Vulcan game engine that supports serious multithreading of the rendering and other sections.

DCS is a very detailed flight sim and this is the direction sims should be going.
 
Last edited:
given the ongoing disparity between increased resolution f VR versus power of graphics cards that never seem to catch up to each other, its time for a "VR Card", a dedicated card next to your graphics card, that takes the load from the graphics card and hardware accelerates it for the VR headset, all the AA fucntions etc, taking all that load off the system, and meaning if you can get 150fps in 4K 2d, you can also get 150fps in VR.

Not sure what architecture that would require or how it would work, but, occurs to me a LOT more VR headsets would be sold if it didnt ALSO mean you needed to buy the latest 3080ti card just to scrape by on FPS.
 
Last edited:
Not sure what that solves. You'd still need a 3080Ti-like card, plus the new add in card. This problem is already solved, yet killed by the very industry we're asking for solutions from.

SLI is perfect for VR. Two cards, one rendering each eye. I have no idea why SLI gets so much hate. I had both crossfire and SLI setups in the past and although there were some headaches, when I had my dual GTX 970's, nothing came close to touching their performance from a single card. It was faster than a 980Ti, easily. With the release of DX12 it was supposed to make things even easier, by using the total pool of VRAM as a resource rather than requiring a mirroring of the memory buffer that DX11 demanded. It was also supposed to make all cards agnostic, treating them as a complete resource of rendering power, instead of requiring two of the same GPU.

Say what you will about the issues surrounding SLI, but having less options is far inferior than having access to choose to use something that may not be for everyone. If we had the theoretical ability to combine 2 6800XT's, 3080's or their bigger brothers/sisters, imagine the result in terms of sheer performance.
 
SLI is perfect for VR. Two cards, one rendering each eye. I have no idea why SLI gets so much hate. I had both crossfire and SLI setups in the past and although there were some headaches, when I had my dual GTX 970's, nothing came close to touching their performance from a single card. It was faster than a 980Ti, easily. With the release of DX12 it was supposed to make things even easier, by using the total pool of VRAM as a resource rather than requiring a mirroring of the memory buffer that DX11 demanded. It was also supposed to make all cards agnostic, treating them as a complete resource of rendering power, instead of requiring two of the same GPU.
I'm not qualified to comment on SLI personally but it was something I did wonder about for VR when I first got it but searching about it brought up lots of comments about latency options. Here's a post I found from a Natural Locomotion dev.

The main trick for SLI was to add a full frame of latency and have each card render alternate frames. With VR you have to avoid all kinds of latency. And even if you render each eye with each card (which requires explicit game support), you still have to send the whole image from one card to the other, and shadows must be calculated on both.

In other words, no dice unless game engines add transparent support for multi GPU VR. That will probably happen eventually, with Vulkan or DX12.

Like I say, this is not my opinion/experience whatever, but it is along the lines of other posts I've seen.
 
Also might not be great for manufacturers if two older GPUs are out preforming a new GPU?
That said it is likely the software that has to catch up.

Well, there are two things to think about there. One, they sell you two GPU's as opposed to a single one. But two, probably the more important to us, is that maybe we can start to see actual significant improvements in the upgrades that we purchase. I'm getting really really bored of the 30% increase in performance for slightly/significantly more money at every segment of the GPU market. In fact, it goes for all sectors of PC hardware. There are definitely stockpiles being secured in terms of performance and we are being drip fed just enough to keep us barely interested.

The great thing is there is major competition in both graphics and computing from AMD. Perhaps the "next-gen" might be a little more interesting. We'll see what happens, especially from Nvidia who now have something to think about. AMD are just going to go full bore with what they have and I can't wait to see it all pan out.
 
they sell you two GPU's as opposed to a single one.
That's almost so crazy that it could work :)

Considering these are parallel processing cores and that they could bridge their Quadro together with full memory access on both sides, why couldn't they just build a twin or even quad GPU card?

Commercial work require lots of expensive memory which would require a second card just to fit all the supporting logic, hence the Quadro connection.

That means that this would be entirely for VR and ultra high resolution triple screen setups or people who wanted to run 8K with high refresh and that is likely not a large enough market.

So I retract my initial optimism. They likely won't build this.
 
Last edited:
  • Deleted member 197115

 
Sale of 4 million Quest 2's at $299 each = $1.2B

Cost of lawsuits and plastic surgery for 4 million people priceless ;)
 
NVIDIA ‘Ada Lovelace’ & AMD ‘RDNA 3’ GPUs To Be Power Chugging Beasts; Over 400W Rumored
Rumors from Kopite7kimi, Greymon55, and Bondrewd suggest that the NVIDIA GeForce RTX 40 series flagship will consume around 400-500 Watts of power, a 20-40% increase.

According to Bondrewd at Beyond3D Forums,
Navi 31 GPUs will feature a 600-650mm2 die, a 15-25% size increase,
consuming 420-450 Watts, a 25-35% increase over RX 6900 XTX.

2.2x to 2.5x expected performance gains.
(The article also tabulates rumored specifications)
 
Last edited:
I'm OK with that power draw if they really are twice as powerful.

And knowing video cards like that are coming will mean VR headsets will be able to do what we want them to do. Give us more resolution, more field of view and a solid framerate.

I'm looking forward to seeing what the Index 2 looks like :)
 

Latest News

What's needed for simracing in 2024?

  • More games, period

  • Better graphics/visuals

  • Advanced physics and handling

  • More cars and tracks

  • AI improvements

  • AI engineering

  • Cross-platform play

  • New game Modes

  • Other, post your idea


Results are only viewable after voting.
Back
Top