At what point will VR tech "look like" 1080p?

[Cavemanhead]

I have a 24", 1920 x 1080 monitor which I suspect the majority of people have. Sure, if I had a higher resolution, I could eliminate some AA, etc... but for a flight sim, this resolution looks fine and dandy.

 

How about for VR? What would be the equivalent pixel density required to push an image that "looks like" what I see now on a 24" monitor? Is 2k x 2 enough or will it take 4k x 2 (8k total)? If the latter, I suppose it will be at least another 2 years before the tech can drive that sort of pixel count.

 

I've tried the Samsung VR (using an S7) and it has a higher resolution than the Rift, yet I've heard the Rift's image is slightly better, so there's more going on there than meets the eye (ha ha).

 

So... When will we see VR get the kind of IQ we have using a 24" 1920 x 1080 monitors?

[Aluminum Donkey]

I have a 24", 1920 x 1080 monitor which I suspect the majority of people have. Sure, if I had a higher resolution, I could eliminate some AA, etc... but for a flight sim, this resolution looks fine and dandy.

 

How about for VR? What would be the equivalent pixel density required to push an image that "looks like" what I see now on a 24" monitor? Is 2k x 2 enough or will it take 4k x 2 (8k total)? If the latter, I suppose it will be at least another 2 years before the tech can drive that sort of pixel count.

 

I've tried the Samsung VR (using an S7) and it has a higher resolution than the Rift, yet I've heard the Rift's image is slightly better, so there's more going on there than meets the eye (ha ha).

 

So... When will we see VR get the kind of IQ we have using a 24" 1920 x 1080 monitors?

 

It probably already does--if you're willing to spend more money than it costs to learn to fly a real aircraft :)

 

Consumer grade? It'll be a while. Enjoy your 1080p single screen :)

[Cavemanhead]

It probably already does--if you're willing to spend more money than it costs to learn to fly a real aircraft :)

 

Consumer grade? It'll be a while. Enjoy your 1080p single screen :)

 

Thanks for responding but I anticipating a more technical/quantifiable discussion in terms of pixel densities, "Rift image tricks", and current video card technology, OLED displays, etc...

 

Anyone?

 

As a baseline, I believe as low as a 970 is able to push 40? fps at the required resolution of the Rift? And 1080 is able to go a full 90 FPS, yes?

 

2k x 2 eyes = 4k which can be done now with a 1080 comfortable on a single screen, but wondering how much the Rift would drop the FPS.

 

And... at what point does a Rift look like today's 1080or 1440 displays?

 

Do we need 4k x 2 = 8k to get there? Seems like overkill but I'm just ball parking. Certainly the VR manufacturer's are asking/answering these same questions as fast as possible. Just wondering what the real quantifiable data and roadmaps from GPU designers says "we'll get there"...

 

Anyone?

[Haukka81]

At what point will VR tech "look like" 1080p?

 

My 970 can push 90fps even in dcs (when not near big city) , but cities hit hard to framrate, 1.5 is worse than 2.0

 

Other VR games run quite fine 90fps about 99% of time.

 

Dcs is getting better and better , more optimization will come in 2.5 i think.

 

 

Sent from my iPhone using Tapatalk

Edited December 11, 2016 by Haukka81

[hansangb]

I have a 24", 1920 x 1080 monitor which I suspect the majority of people have. Sure, if I had a higher resolution, I could eliminate some AA, etc... but for a flight sim, this resolution looks fine and dandy.

 

How about for VR? What would be the equivalent pixel density required to push an image that "looks like" what I see now on a 24" monitor? Is 2k x 2 enough or will it take 4k x 2 (8k total)? If the latter, I suppose it will be at least another 2 years before the tech can drive that sort of pixel count.

 

I've tried the Samsung VR (using an S7) and it has a higher resolution than the Rift, yet I've heard the Rift's image is slightly better, so there's more going on there than meets the eye (ha ha).

 

So... When will we see VR get the kind of IQ we have using a 24" 1920 x 1080 monitors?

 

My guess with be by 2018. It's actually pretty good today.

 

But you're really missing the point about VR. Resolution is a factor. But with today's VR - even with lower resolution - it's simply stunning. Not because of what you see on the HMD, but because how it makes you *feel*. The immersion, the ability to track targets, wingman, and SAMs with 100% SA. Something TrackIR simply *cannot* do because your brain knows it's not real.

 

The best comparison I can make is this. "When will Ferrari's get 25MPG? I'll wait until then" in the meantime, you're missing out on the growl of the engine, turning on rails, being slammed back into your seat etc! :)

[SimHog]

There's 3 things you need to factor in ...

 

- Pixel density

- Screen size

- Eye distance

 

Using Apple's Retina paradigm ... "It turns out there’s a magic number right around 300 pixels per inch, that when you hold something around to 10 to 12 inches away from your eyes"

 

But in VR things aren't that simple, changing one variable, affects the other two. Take the 300 ppi display and move it closer, you loose image sharpness and you introduce eye strain.

 

Oculus Rift uses a 5" display at approx 2" from your eyes. Not taking into account the lens affects like "screen door, you would need a density ~1700 ppi or higher.

 

The best mobile phone displays we have on the market now are 4k (3840x2160) that's a density of 881ppi and "Retina" quality can be maintained up to 4" from your eyes. But 4" from your eyes is too far away for VR as you loose FOV.

 

So to get "Retina" quality in VR you'd be looking at 8k displays (7680×4320 ~1760 ppi) @ 2" away from your eyes.

 

That's the holly grail of VR imo.

 

Samsung was reported to be researching 11k mobile phone displays (2250 ppi) last year.

Edited December 12, 2016 by CoderX71

[DoctorStrop]

So to get "Retina" quality in VR you'd be looking at 8k displays (7680×4320 ~1760 ppi) @ 2" away from your eyes.

 

You're probably right. Also a GPU capable of running those resolutions at 90 FPS is probably years off.

[SimHog]

You're probably right. Also a GPU capable of running those resolutions at 90 FPS is probably years off.

 

That's even a bigger hurdle. There would have to be a major in shift in GPU thinking / architecture I think to achieve 8K resolution at 90Hz. :idea:

 

With current architecture the only way to achieve more is to go smaller die, but we're already reaching the limit on that front.

Edited December 12, 2016 by CoderX71

[DoctorStrop]

That's even a bigger hurdle. There would have to be a major in shift in GPU thinking / architecture I think to achieve 8K resolution at 90Hz. :idea:

 

With current architecture the only way to achieve more is to go smaller die, but we're already reaching the limit on that front.

 

..and would the compute power be enough at the moment?

[BlackLion213]

It's actually a pretty hard question to answer because the perceptive experience is really different with these binocular HMDs compared to a normal monitor.

 

During a normal day, our eyes and brain spend a lot of time "signal averaging" and dismissing visual bits of info that our brain feels are irrelevant. This is probably why the screendoor effect disappears for most rift users - the pixel lines are not aligned between the 2 images so you have to actively focus on the screendoor to see it.

 

I am not yet a rift user (Christmas gift :D), but I tried it out at the USAF Museum in Dayton OH. They had a spacewalk experience using the Rift CV1 and it was really insightful. If you can try a Rift even for a few minutes it will probably answer a lot of your questions. The actual Rift demo wasn't the best use IMHO (almost no near objects, no "body" for the user, just floating in space, etc), but the sense of depth was really impressive.

 

When I first put the HMD on I saw the screendoor, but the moment I focused on anything it disappeared. Actual objects where clearer than I expected. If I focused on something with one eye, screendoor was more noticeable and some jagged features (the stuff that AA handles) was more obvious. But the binocular image smoothed out some things and really hid the screendoor in my experience.

 

I think the more telling bit of info is the lack of people on this forum who tried the Rift (or Vive) and either returned it or still use their monitor most of the time - I can't think of anyone! (maybe I didn't look hard enough...).

 

If you are wondering when the Rift will be good enough that the vast majority of players prefer it to a monitor despite a few pros and cons - the answer is now.

 

Wondering when the Rift will have little to no drawbacks compared to a HD monitor - probably soon, 2018ish as Hansangb said seems reasonable.

 

When will VR be good enough that there will be no suspension of disbelief....probably a long time.

 

The good news is it's probably not a big deal, because what is available now is already better than a monitor for nearly everyone who tries it. Still I am excited to see how things change over the coming years, it looks really promising!

 

-Nick

[Torso]

We will see something new in 2017, 4th quarter I am betting. And it sounds like significantly better than current. Here is an article on the new displays. They have already shipped samples to the manufacturers for review. They are readying for production pretty much.

 

http://www.roadtovr.com/display-conglomerate-jdi-developing-ultra-high-resolution-panels-vr-headsets/

[SimHog]

..and would the compute power be enough at the moment?

 

I'm no GPU expert but I suspect not.

 

I'm guessing if nothing changes then GPU architecture would need to grow like CPU's have, i.e. multiple "complete" cores on 1 die sharing the work load. Then there's the bandwidth and VRAM requirements to think of - More power, more heat, more cost ... That's why I say there needs to be a complete re-think on how pixels are pushed to the screen. Who knows maybe quantum computing holds the key.

 

https://singularityhub.com/2016/10/10/massive-disruption-quantum-computing/

 

 

Edit: Here's a good explanation of the limits we're approaching.

Edited December 12, 2016 by CoderX71

[hannibal]

even if 4K HMDs were available.. theres no computer cost effective enough to push it at that resolution at 90fps..

[Precog]

Eye tracking and foveated rendering technology is the key to powering these high res displays. The future looks bright...I mean sharp!

[Haukka81]

Here is my experience so far:

 

I have 2560x1440 monitor and Oculus cv1 (and before i have triple screen setup)

 

-rift cv1 has very good vertical fov now, horizontal is still far from perfect. More fov would need even more resolution so i think that current compromise is good.

 

Ground target Spotting is harder than my monitor , but Sense of depth and 1:1 trakking makes it less pain than you would think. Once you see target its easy to "feel" where they are -> so Its harder but same time its easier :P

 

Spotting would be better if dcs would draw far away targets more dark (black dots etc..)

 

I dont use labels btw , i used custom labels bit but it wont feel good no mater how much i edit those. Somekind smart scaling is better (imho)

 

When i get back to my rig i write more , hard to write with phone now.

 

Sorry my crappy English

 

 

Sent from my iPhone using Tapatalk

[CallsignFrosty]

My guess with be by 2018. It's actually pretty good today.

 

But you're really missing the point about VR. Resolution is a factor. But with today's VR - even with lower resolution - it's simply stunning. Not because of what you see on the HMD, but because how it makes you *feel*. The immersion, the ability to track targets, wingman, and SAMs with 100% SA. Something TrackIR simply *cannot* do because your brain knows it's not real.

 

The best comparison I can make is this. "When will Ferrari's get 25MPG? I'll wait until then" in the meantime, you're missing out on the growl of the engine, turning on rails, being slammed back into your seat etc! :)

 

 

 

Well I can't afford a Ferrari either! Lol

 

 

Sent from my iPhone using Tapatalk

[CallsignFrosty]

I'm no GPU expert but I suspect not.

 

I'm guessing if nothing changes then GPU architecture would need to grow like CPU's have, i.e. multiple "complete" cores on 1 die sharing the work load. Then there's the bandwidth and VRAM requirements to think of - More power, more heat, more cost ... That's why I say there needs to be a complete re-think on how pixels are pushed to the screen. Who knows maybe quantum computing holds the key.

 

https://singularityhub.com/2016/10/10/massive-disruption-quantum-computing/

 

 

Edit: Here's a good explanation of the limits we're approaching.

 

 

 

Quantum computers will never replace conventional cpus for these kind of tasks. Plus I don't really want to have to have vats of liquid nitrogen in my house to fly on DCS in 2030!

Carbon nanotube microprocessors seem like the next step and I assume they would be able to use a similar processor design but the manufacturing material would allow for higher speeds and thermal performance.

I have a friend working on those ATM for his PHD, I'll ask him over Christmas what his work may mean for DCS!

 

The only problem is convincing intel to release new tech when there is no serious competition in the marketplace :(.

(Cue some AMD users getting triggered haha)

 

 

Sent from my iPhone using Tapatalk

[hansangb]

;2981526']Well I can't afford a Ferrari either! Lol

 

 

Sent from my iPhone using Tapatalk

 

 

:D Good point! :)

[mx22]

I think the more telling bit of info is the lack of people on this forum who tried the Rift (or Vive) and either returned it or still use their monitor most of the time - I can't think of anyone! (maybe I didn't look hard enough...).

 

Price and how it correlates to current DCS experience would probably be high on the list of why VR sets are being returned. Immersion of flight is there, but playability is lagging behind regular monitor experience (lower resolution, harder target spotting, high hardware demands to begin with coupled with the fact you would need to lower graphics settings to make it playable).

[hansangb]

High return rate? Where did that come from?

[Rangi]

High return rate? Where did that come from?

 

You just made it up. :book:

[Mr_sukebe]

Commenting as a CV1 and 2480*1440 monitor user. The monitor is sharper and can run higher visual settings in DCS1.5. For all that, I never use my monitor when actually flying in DCS. If anything, DCS is probably the best VR demonstrator that I have for when friends ask about VR. The sheer feeling of for example being in a Huey is simply awesome.

[Sneetch]

I was a vive owner for 2 weeks and returned it due to the screen door effect, small 'sweet spot' for eyes to have clarity of view, lower resolution. small FOV like viewing world thru binoculars, light blooming effect and issue with shutting out the world (family) when playing. I returned to my triple screen setup for the time being. I am waiting for the next gen headsets to come out to see how they can overcome those challenges, well save the last one, that is my personal issue. :) It was a hard decision to return as the immersion that was mentioned is VERY true. The sense of scale and feeling of presence was incredible. Yet for me, it wasnt enough to keep the vive in its current state. The pixel density of display will help with visuals but also need to contend with improving the lenses since they fake out the wearers eyes to believe they have a 15-20 foot viewing distance. I believe it is the lenses that are creating the light bloom effect and it can be annoying. I am also curious to see if the nvidia api's to VR will be implemented, they have potential to take a huge strain of the GPU when rendering, yet too many times vendor specific solutions are poorly supported. (thinking about 3d Nvision and gsync support) All that aside, VR is where it is at for sims like DCS. :)

[Neldrion]

Prediction for VR in 5 years by Micheal Abrash:

 

[ATTACH]152666[/ATTACH]

 

My estimation:

 

I think the important number is pixels per degree per eye. Because if you raise the FOV you need way more resolution.

 

If you raise the resolution from 2,6 million pixels (Oculus CV1) to 30 million pixels (Oculus CV3) you get only twice the perceived resolution because FOV will propably increase also.

 

In VR Gen 2 in 2018 we will propably have a resolution increase of 133% which is not that much. It will still not be enough to feel like a 1080p normal desktop screen.

 

In VR Gen 3 the jump will be big, but that is far off in the future, 2020 earliest.

 

VR Gen 1

Year: 2016

FOV: 100°

Resolution for both eyes: 2.160 x 1.200 = 2,6 million pixels

Pixels per degree per eye horizontally: 12 pixels per degree (= 100%)

 

VR Gen 2

Year: 2018

FOV: 120°

Resolution for both eyes: 3.840 x 2.160 = 8,3 million pixels

Pixels per degree per eye horizontally: 16 pixels per degree (= 133%)

 

VR Gen 3

Year: 2020

FOV: 140°

Resolution for both eyes: 7.680 x 3.840 = 29,5 million pixels

Pixels per degree per eye horizontally: 27 pixels per degree (= 225%)