Obviously. But some of the best VR experiences are a lot less demanding graphically than E. Also the resolution is the biggest weak-point (imo) in the VR right now; I'm sure that most of the people would be very happy with 4K @ Low Settings instead of 1440p @ Medium or High settings on the more demanding titles.
I'm betting most would not. I'm fully aware that you can have compelling experiences at lower settings, but I don't think those will be what is popular or pushes adoption. People will want demanding VR.
I also understand the benefits resolution provides. I think 1440p will provide an adequate resolution for what current VR wants to do, to the point where people will choose better graphics at 1440p over poorer graphics at 4k.
I'm sure high-end gaming in 4K would be problematic, but what about everything else? Especially pre-rendered video?
And of course you can still run something at a lower resolution while benefitting from the reduced SDE of the display's resolution. Apparently the scaling should introduce a really negligible amount of latency.
I'm talking specifically about a diffuser over the screen, which doesn't really care what the rendering resolution is - it acts upon the absolute resolution of the screen. Having tried diffusers in the past, I'd go with a sharp display any day. Diffusers reminds me quite a bit of looking at the world without glasses on (assuming you need them). It's kind of aggravating.
Well, as I said earlier, that's not going to go away any time soon. They think the "retina" resolution for VR is 8k x 8k per eye. It's going to be a looooong time before we get that sort of display technology.
You have to think about it like this: VR today is what the NES was back in 1985. What you're asking for is what, say, an Xbox 360 can do with 2D. It'll just take time for this stuff to get to that level.
By turning stuff down, yes. I can absolutely bring a PC to it's knees, even with dual 980s, with a game like Elite.
And no systems really run elite well, period. They all are subject to microstutter and judder.
My bad, you're talking exclusively about the DK2. I haven't yet gotten a chance to play too much of Elite (but was playing it at 1440p with my 780Ti when it was in Beta). Here's hoping it improves...certainly the judder in games on DK2 makes me beyond nauseous. That's the main thing I hope they personally improve on consumer release.
You have to think about it like this: VR today is what the NES was back in 1985. What you're asking for is what, say, an Xbox 360 can do with 2D. It'll just take time for this stuff to get to that level.
I'm talking specifically about a diffuser over the screen, which doesn't really care what the rendering resolution is - it acts upon the absolute resolution of the screen. Having tried diffusers in the past, I'd go with a sharp display any day. Diffusers reminds me quite a bit of looking at the world without glasses on (assuming you need them). It's kind of aggravating.
I'm not trying to say we absolutely need 8k per eye.
I'm just saying 4k is probably the minimum before most people won't notice pixels when actively using a VR device. 1440p is really close but in some situations (namely videos, some games, menus, and text) it is noticeable. Text is still pretty bad unless it's "close" to your face and blown up more.
I get you, but I'm saying what research indicates is that 4k really isn't the minimum when people stop noticing pixels. Research indicates we won't reach that level until 8k^2 per eye.
Which is all a long way of saying - when we finally do reach 4k, you might be disappointed that you still see the screen door.
I'm betting most would not. I'm fully aware that you can have compelling experiences at lower settings, but I don't think those will be what is popular or pushes adoption. People will want demanding VR.
I also understand the benefits resolution provides. I think 1440p will provide an adequate resolution for what current VR wants to do, to the point where people will choose better graphics at 1440p over poorer graphics at 4k.
I can run much more graphically demanding games on my DK2, sure, but the experience is much worse due to resolution difference. Keep in mind the jump from 1440p to 4k is bigger than 1080p to 1440p. Except for people that demand the best visuals possible so many people would take 4k with less demanding visuals. I wouldn't dream of watching video on a 1440p VR device if a 4k device was available.
You have to remember the people who really honest to God care about extreme visuals in PC games is quite small compared to the rest of PC gamers. Also mass market VR cares not for serious gaming anyway.
I get you, but I'm saying what research indicates is that 4k really isn't the minimum when people stop noticing pixels. Research indicates we won't reach that level until 8k^2 per eye.
Which is all a long way of saying - when we finally do reach 4k, you might be disappointed that you still see the screen door.
I'm not worried about the screen door as much as you think. I'm just saying in terms of average enjoyment once VR is more mainstream is it will be less noticeable and more easily dismissed at 4k. It can already be hard to spot in some VR instances with Gear VR. A slight push will make it less of an issue and not detrimental to an enjoyable VR experience. That's all.
I can run much more graphically demanding games on my DK2, sure, but the experience is much worse due to resolution difference. Keep in mind the jump from 1440p to 4k is bigger than 1080p to 1440p. Except for people that demand the best visuals possible so many people would take 4k with less demanding visuals. I wouldn't dream of watching video on a 1440p VR device if a 4k device was available.
You have to remember the people who really honest to God care about extreme visuals in PC games is quite small compared to the rest of PC gamers. Also mass market VR cares not for serious gaming anyway.
Having experienced just about every demo currently available on both dks and gear vr, I'll say that you quickly get tired of clearly low rent art at good resolutions. Alien isolation, elite dangerous, project cars - a good part of why these are so exhilarating is the AAA effort behind them. Nobody is going gaga over ashnar wars after trying elite.
Having experienced just about every demo currently available on both dks and gear vr, I'll say that you quickly get tired of clearly low rent art at good resolutions. Alien isolation, elite dangerous, project cars - a good part of why these are so exhilarating is the AAA effort behind them. Nobody is going gaga over ashnar wars after trying elite.
At the moment sure. In the future it won't be. VR isn't going to end up as only a gamer hobby. That's thinking too small.
Oculus Cinema is (outside of the Mars 4k pictures) the best use of VR I've currently experienced. A movie theater on your head is very cool and awesome.
Also 360 video with real clarity is probably going to be a HUGE selling point of VR in the future. Its got so much potential it is insane. Its much trickier to get right because it needs an insane amount of bandwidth and 360 video is still in the growing stages but it will be amazing in the future. Mass market will eat up VR for virtual tours of places around the world that can be experienced in VR
Well first up, I don't think the nes is bad to begin with. But you can't compare output resolution like that - vr screens and resolutions behave differently than a television screen half a room away.
I'm not talking about diffusers, but having a 4k display in the Oculus Rift. Shouldn't be that far off, yea? 2017 at the absolute latest, I imagine.
Just to clarify, 4k isn't the same as 4kx4k per eye. 4k is 3840x2160, which is 8294400/2 pixels per eye, which is 4147200 pixels per eye. 4k x 4k per eye is 16000000 pixels per eye, or 32000000 pixels total. Which, according to research, likely still isn't enough to reach retina resolution.
We are a loooong ways off from that type of display technology.
Vr will be the diving force behind display technology for decades. It will be bleeding edge technology the same way video cards in general are bleeding edge technology.
Vr today is what 3d accelerators were in the early 90s.
Vr will be the diving force behind display technology for decades. It will be bleeding edge technology the same way video cards in general are bleeding edge technology.
Vr today is what 3d accelerators were in the early 90s.
That's my entire point, at these absurd resolutions, even "simple" content requires insane hardware. We don't have this technology yet, nor will we for decades.
Let's not forget that the DK2 is not the most current incarnation of the rift, Crescent Bay was being raved about back in September, it's feasible that they could release the updated version of that prototype in the spring.
I get you, but I'm saying what research indicates is that 4k really isn't the minimum when people stop noticing pixels. Research indicates we won't reach that level until 8k^2 per eye.
There is a difference between Retina-like resolution and being able to distinguish pixels.
I have a 1080p monitor and cannot identify individual pixels. But I do know that pixel grid is still creating an obfuscating effect, reducing clarity. I don't think we necessarily need Retina-like fidelity for people to be quite happy with the clarity.
There is a difference between Retina-like resolution and being able to distinguish pixels.
I have a 1080p monitor and cannot identify individual pixels. But I do know that pixel grid is still creating an obfuscating effect, reducing clarity. I don't think we necessarily need Retina-like fidelity for people to be quite happy with the clarity.
Yep there is a point at which the average consumer just can't tell the difference or just doesn't care. For me, as long as the pixilation is not preventing me from let's say reading a scan of a magazine I'm happy.
I don't think it is the guess (unless some new information has come to light). The shape of the CB housing appears to be concealing a 21:9 panel (and a screengrab of some prototype circuitry in one of the Oculus videos might show connections for two panels). The lenses appear to be a different shape too (not round), yet supposedly the FOV hasn't improved very much. This could point to some kind of anamorphic solution that is corrected in the optics mainly to reduce the screen door effect rather than to increase FOV, for example 2560x1440 stretched across a 21:9 grid, with the optics bringing it back to look like 16:9, giving more densely-grouped pixels (horizontally).
Carmack also seemed to suggest in his Gear VR talk that the Note 4 panel couldn't do 90Hz (although this wasn't explicitly stated).
Gear VR's resolution seems fine to me. Yes, you can see it. Barely... it doesn't get in the way or ruin the experience. Nerds and tech enthusiasts might get upset, but I don't think a single casual would complain. 1440p seems like a perfectly good starting point, and if it helps keep the cost of the final product low even better. But I guess a lot can happen between now and the end of the year.
I don't think it is the guess (unless some new information has come to light). The shape of the CB housing appears to be concealing a 21:9 panel (and a screengrab of some prototype circuitry in one of the Oculus videos might show connections for two panels). The lenses appear to be a different shape too (not round), yet supposedly the FOV hasn't improved very much. This could point to some kind of anamorphic solution that is corrected in the optics mainly to reduce the screen door effect rather than to increase FOV, for example 2560x1440 stretched across a 21:9 grid, with the optics bringing it back to look like 16:9, giving more densely-grouped pixels (horizontally).
Carmack also seemed to suggest in his Gear VR talk that the Note 4 panel couldn't do 90Hz (although this wasn't explicitly stated).
Why would it lower the resolution? If it's properly matched to the display, wouldn't the resolution stay the same, while only increasing the apparent pixel fill ratio? Can they not diffuse per pixel?
That's my entire point, at these absurd resolutions, even "simple" content requires insane hardware. We don't have this technology yet, nor will we for decades.
Actually, no, there isn't. Retina display is already a bit of marketing speak because it's not a defined resolution, but rather the point where the ability to discern individual pixels vanishes. Obviously that's a function of distance, resolution, and pixel density. But given a relatively fixed distance that vr affords, it becomes easier to conceptualize a "retina" display for vr.
Why would it lower the resolution? If it's properly matched to the display, wouldn't the resolution stay the same, while only increasing the apparent pixel fill ratio? Can they not diffuse per pixel?
per Michael abrash, John Carmack, and Palmer lucky yes we are decades away from viable foveated rendering.
And, because every time I say this people bring up the same articles - no, that's not close to viable yet, from both a cost and performance standpoint.
That's my entire point, at these absurd resolutions, even "simple" content requires insane hardware. We don't have this technology yet, nor will we for decades.
Ok that's an overstatement if I ever saw one. For simple but useful content like the virtual cinema (or, perhaps even more useful, a virtual spherical desktop on PC) a single modern GPu is more than sufficient to run 4k² per eye. If we had the display hardware and interconnect to make it happen, there's lots of possible content where such resolutions would already be advantageous now, never mind a decade from now.
Ok that's an overstatement if I ever saw one. For simple but useful content like the virtual cinema (or, perhaps even more useful, a virtual spherical desktop on PC) a single modern GPu is more than sufficient to run 4k² per eye.
To begin with, the resolution quoted is 8k^2 per eye, and we don't even have a display standard that will support that bandwidth, nor will we for decades. Particularly at the refresh rates carmack dropped alongside his theatrical ideal resolution: 240 hz.
If we had the display hardware and interconnect to make it happen
To begin with, the resolution quoted is 8k^2 per eye, and we don't even have a display standard that will support that bandwidth, nor will we for decades. Particularly at the refresh rates carmack dropped alongside his theatrical ideal resolution: 240 hz.
Its being realistic, those who make said displays need to actually have a market for these technologies for them to exist. VR isnt a market driving revenue for display hareware yet, nor will it be for.a while. It's not unrealistic to expect 20 years of market growth before it becomes viable for said theoretical displays to become produced.
If you're going to argue that such a display is not inherently impossible today, so be it. But it's not even close to being ready for the market, nor will it be for a very long time.
Its being realistic, those who make said displays need to actually have a market for these technologies for them to exist. VR isnt a market driving revenue for display hareware yet, nor will it be for.a while. It's not unrealistic to expect 20 years of market growth before it becomes viable for said theoretical displays to become produced.
If you're going to argue that such a display is not inherently impossible today, so be it. But it's not even close to being ready for the market, nor will it be for a very long time.
I think you missed my point. Again, as you were talking about content - and that only matters for the rendering part of the equation - I was talking about rendering hardware, not displays.
Physical dot pitch of the display isn't relevant in VR. What's relevant is how much resolution you have per arcminute of the FoV you want to cover. So you can either decrease the FoV (a big nono) or increase the resolution. Or, if it's just about screendoor, blur (or diffuse to put it more in more easily digestible language ).
They actually want to do this with a custom display - higher pixel densities towards the center where each eye focuses. The problem is that, currently, oculus has to justify such displays to samsung for non-vr applications because samsung isnt producing displays solely for vr headsets. Unlike carmack's research into low latency, which will manifest in phones as "sharper" screens when scrolling, the type of display oculus proposes for vr doesnt really benefit non-vr applications.
I think you missed my point. Again, as you were talking about content - and that only matters for the rendering part of the equation - I was talking about rendering hardware, not displays.
I think you missed my point. Again, as you were talking about content - and that only matters for the rendering part of the equation - I was talking about rendering hardware, not displays.
Physical dot pitch of the display isn't relevant in VR. What's relevant is how much resolution you have per arcminute of the FoV you want to cover. So you can either decrease the FoV (a big nono) or increase the resolution. Or, if it's just about screendoor, blur (or diffuse to put it more in more easily digestible language ).
I don't follow you. If the screen door effect being talked about here isn't about seeing the non-illuminated spaces between pixels, then what is it you are seeing?
Oh, then I agree. I mean, it's not really an open engineering challenge (the interconnect that is), we already have production HW you can buy that can transport 300 GBit/s (EDR Infiniband), but there's very little consumer market incentive for something like that even in the longer term.
I don't follow you. If the screen door effect being talked about here isn't about seeing the non-illuminated spaces between pixels, then what is it you are seeing?
That is what is being talked about. However, the important point is how big those spaces are in your field of view, not how big they are on the physical display. E.g. if you build a HMD with a 4" or an 8" screen with the same resolution and fill factor, obviously the 4" screen has a far superior PPI. However, if the optics for both HMDs are designed to cover the same FoV, you'll get the same amount of screendoor effect.
I don't follow you. If the screen door effect being talked about here isn't about seeing the non-illuminated spaces between pixels, then what is it you are seeing?
So, when talking about physical dot pitch, you can talk about two separat phenomenons that seem to get lumped together under the umbrella term "screen door effect" by the common person. Dot pitch would do more to mitigate the jailbar effect - bands of subpixels turned off for a specific color that result in bars running from top to bottom (or, technically, left to right). The classic rgb subpixel matrix exhibits this problem. A pentile subpixel structure, as currently used, reduces jailbars by not letting an entire band of sold color run uninterrupted.
Which is what i assume durante is talking about. That is technically not related to the screen door effect at all, even if they seem to get lumped into the same term erroneously. Pixel density itself is just one of the factors that plays into the classic screen door effect as you describe.
E.g. if you build a HMD with a 4" or an 8" screen with the same resolution and fill factor, obviously the 4" screen has a far superior PPI. However, if the optics for both HMDs are designed to cover the same FoV, you'll get the same amount of screendoor effect.
Right, this is what I meant with my last sentence. Pixel density alone doesn't define the screen door effect. But, given equal everything else, a higher pixel density would result in a reduced screen door effect.
Problem is each new display demands completely reworked optics.
Fixed, and yes. I made a couple of illustrations when I was trying to explain fill ratio a few months back. First, we have a typical LCD, with a 25% PFR, and also a typical OLED, with an 80% PFR.
(Fig. 1)
(Fig. 2)
Same display size. Same resolution. Same dot pitch. Same pixel density. From a software perspective, the displays are identical, but the improved PFR greatly reduces SDE.
To be clear, dot pitch is the distance from the center of one pixel to the center of the next, measured in millimeters or whatever. That's important for overall image quality, but screen door comes from the empty spaces between the lights, and the difference between lit and not-lit is expressed as the fill ratio. Increasing the dot pitch/pixel density will improve resolution at the cost of increased computational expense, but that doesn't really change the screen door effect that much; even though the mesh is finer now, it's still blocking out 20% of your view overall (assuming 80% PFR). Yes, you're decreasing the gaps between the holes, but the holes themselves become proportionally smaller in the process, reducing your net gains considerably. No matter how many pixels we jam in to Figure 2, there will always be gaps between them, so we'll always get the screen door effect. If we're strictly discussing SDE, it doesn't matter if your display consists of 100 lights or 100,000,000 lights; the problem is that 20% of it isn't lit up at all. That's where diffusers come in. You can spread that light a tiny bit before it hits the eye, to help fill in those unlit gaps.
Sorry, but I still don't follow. If you start with an 80% PFR display as shown in Figure 2 above, then apply a 25% diffuse filter, wouldn't you end up with 100% PFR, and 0 SDE? If your pixels are blending, then you just need to dial back the diffusion a bit. I would think overlap would cause all sorts of weird color artifacts as well, when red pixels bleed over in to blue pixels, and vice versa. Am I missing something? Why would you use "too much" diffuse? I would think "not quite enough" would actually be preferable. =/
My math may not be accurate, but hopefully you get the idea. <3
per Michael abash, John Carmack, and Palmer lucky yes we are decades away from viable foveate rendering.
And, because every time I say this people bring up the same articles - no, that's not close to viable yet, from both a cost and performance standpoint.
I'm not sure what articles you're referring to, but what about SMI's Mobile Eye Tracking? It's only 60 Hz, which I suspect isn't fast enough for foveated rendering, but surely we're not "decades away" from something which is fast enough, right? 20+ years to get to what, 240 Hz? You mentioned cost as well; how much does this MET rig run?
Incidentally, these guys are actively working with Magic Lab. They've only publicly demonstrated eye tracking using the Camera, but I can't imagine that's the only thing they're working on.
Yep there is a point at which the average consumer just can't tell the difference or just doesn't care. For me, as long as the pixilation is not preventing me from let's say reading a scan of a magazine I'm happy.
I remember somebody at Oculus talking after Crescent Bay came out and noting that there are ways to achieve better 'clarity' through clever optics rather than just relying solely on resolution increases. People who have tried both Gear VR and Crescent Bay said that Crescent Bay looks better(some comments even suggesting the resolution must have been higher than 1440p), although none of the same games/experiences shown were directly comparable, either. But I do think that was strongly hinting that it will look better than what you'd expect if you just swapped a 1440p display into the DK2.
It's going to be big at basically every consumer event coming in the next year. I don't think people realize we're about to get a ton of awesome VR news in 2 days. Fox seems like they're going to bring their A game.
Actually, no, there isn't. Retina display is already a bit of marketing speak because it's not a defined resolution, but rather the point where the ability to discern individual pixels vanishes. Obviously that's a function of distance, resolution, and pixel density. But given a relatively fixed distance that vr affords, it becomes easier to conceptualize a "retina" display for vr.
Retina-like resolution is meant to be when there is a glass like quality to the display in terms of clarity. There is a gap between the point where you can stop being able to identify individual pixels and having a Retina display. Typically, we're talking at least 200+ PPI for Retina in a desktop situation, while my 1080p, 82ppi monitor is already enough for me to be unable to see individual pixels from a normal viewing distance.
Now if Oculus meant something else by this, its possible that you're right, but you nor I would have any idea about that. I'm merely going by the typical way that Apple classifies their Retina displays.
Anyways, if a decent resolution and foveated rendering are indeed 20+ years away, I'm going to have to do a flip in my feelings towards VR and say that I don't think it'll be successful due to the technology not being there yet. I was always under the impression that the tech would advance rapidly and it wouldn't be 20+ years before doing ultra high res VR would be feasible.
. Also the resolution is the biggest weak-point (imo) in the VR right now; I'm sure that most of the people would be very happy with 4K @ Low Settings instead of 1440p @ Medium or High settings on the more demanding titles.
).
