This isn't actually true. At least the part about redesigns. Oculus still hasn't shown a protype from this year. The last prototype from them was shown off last Sept. So they will have plenty of time to design something new and better by fall. Look at the jump sony accomplished.
Sony VR Headset/Project Morpheus/PlayStation VR
This isn't actually true. At least the part about redesigns. Oculus still hasn't shown a protype from this year. The last prototype from them was shown off last Sept. So they will have plenty of time to design something new and better by fall. Look at the jump sony accomplished.
It was planned tweaks. Last year, when sony have shown a prototype, they said they were planning to get an oled screen, that they might not have the frame interpolation in hardware, and they said the 40ms lag was something they were planning to reduce... A year later they made every single one of these tweaks. It's not a complete redesign. They also tweaked the tracking leds, and might have a few surprises left... everybody is tied to the same limitations of parts procurement.
It sounds like Oculus may have somthing big to reveal yet with regards to interface as well. Not particularly relevant to headset quality but it does show that they are still a ways off from an end consumer product.
With regards to the comparisons about image quality at this stage, I really don't think it makes a lot of sense to be making these comparisons on the basis of the light weight software currently available. Is the experience still going to be comparable in say Crysis 4? How much visual fidelity would the PS4 have to drop in that game for example merely to achieve 1080p/60fps while Oculus on a sufficiently powerful rig will likely be maxing out the games visuals at 1440p/90fps. I don't think expecting near parity between PS4 and high end PC's in those types of games is particularly realistic.
With regards to the comparisons about image quality at this stage, I really don't think it makes a lot of sense to be making these comparisons on the basis of the light weight software currently available. Is the experience still going to be comparable in say Crysis 4? How much visual fidelity would the PS4 have to drop in that game for example merely to achieve 1080p/60fps while Oculus on a sufficiently powerful rig will likely be maxing out the games visuals at 1440p/90fps. I don't think expecting near parity between PS4 and high end PC's in those types of games is particularly realistic.
Crysis 4 at 60 fps on PS4? VR on consoles will be a very different experience than on PC I think most of the game will be dedicated VR experience on PS4... GT 7 or Project Car will probably be there. Adventure action games with slow movement, spatial games Like Eve Valkyrie or No Man Sky...
Last edited:
D
Deleted member 11852
Guest
This is the question I'd really like answered and which I expect there are many answers. On whatever arbitrary scale, how much 'visual fidelity' do you have to surrender to produce a stereoscopic output of a scene compared to a single frame 2D output intended for a TV?
Whatever the answer, it's clear that Sony have fully control of their API to make this easier on developers than Oculus or Valve do in their respective platforms. Assuming the APIs a pivotal part of the pipeline.
I doubt API control will help any. The renderer will have to scale, and that's down to the devs to create scaled assets, shaders, and other resources. The API provides the interface to get the hardware to draw stuff, but isn't high-level enough to handle management of the resources with rescalability in mind. You'd look to the engine for that. Something created on UE4, say, could be built in two flavours, for 2D and VR. Something built on a proprietary engine like No Man's Sky is going to need a whole custom build for a VR mode, I expect.
D
Deleted member 11852
Guest
Very true! Same question from an engine perspective or do we just need new engines designed to abstract and minimise the hardware impact of producing a render of two frames suitable for 3D rather than one. Maybe we need hardware designed with this in mind?
Logically the two frames will be more alike than that will different and I'm curious how much you can fudge to take advantage of this rather than do two completely separate independent renders, one for each eye.
Sebbbi et al have spoken to some length about 3D reprojection techniques. It shouldn't need a whole new engine as the core principles of geometry and pixel values are will remain the same when producing any image. It just requires a suitably effective extrapolation method to produce two images from the one rendered output (or whatever interim buffer format). I can see a central perspective 'render' of geometry being used which can be perspective corrected for each eye, and then that virtually textured. I guess the demise of 3D put a dampers on this research, but VR should start it back up with a vengeance. At least on console. PC may just rely on gamers to upgrade their hardware!
Of course, with foveated rendering this wouldn't be an issue. I wonder what the limitation on its adoption is? Not fast enough?
(going off-topic)
Well yeah, as I said I'm hoping/expecting the consumer version will have a 1440p RGB, and an improved optical filtering.
Wildcard is if they delay to 2016, then they have a potential 2160p pentile from samsung. But it's not such a big improvement over 1440p full rgb, as can be seen when comparing Sony's 1080p rgb versus the competitor's 1440p diamond pentile.
Does the foveated rendering come with eye-tracking?
Somewhat unseriously, it would compromise the artistic vision of cinematic AAA games that blur out whatever they think we shouldn't care to twitch our eyes towards. At least until they implement an eye-controlling VR headset.
Press X to moisten corneas.
edit: spelling
Somewhat unseriously, it would compromise the artistic vision of cinematic AAA games that blur out whatever they think we shouldn't care to twitch our eyes towards. At least until they implement an eye-controlling VR headset.
Press X to moisten corneas.
edit: spelling
Last edited:
Two 1440x2560 panels would be nice. Esp if they are smaller than the s6 screen. PPI on that is almost 600. So if they can shrink it and keep the resolution you could be looking at a nice vr experience. It would trump the 1200x1080 screens in the vive or the 960 per eye of Morpheus.
Remember they went dual screens with cresent bay. So its all up to the resolution they can get on the screens.
Isn't reprojection what Nvidia uses for it's 3D Vision compatibility mode? And what Crytek use in Crysis 2? If so then it's certainly not comparable to rendering two completely separate scenes and I'm not sure it would be sufficient for VR.
isnt reprojection is the method that result in "cardboard cutout" 3D? then it will looks weird in VR
Two small displays are the same total surface as a big one. There are many advantages but it's not allowing more resolution, the limitation of oled is the sub-pixel density, so splitting the screen in two doesn't change that maximum.
The smaller displays per eye must be square or just a bit taller. That would be 1280x1440p pentile for currently available samsung lines. What I'm saying is that a 2016 launch could possibly bring either 1920x2160p pentile or 1280x1440p RGB per eye. This also applies to Sony but they are targeting a lower price for mass adoption, and I doubt they'll change the resolution before launch. The external processor would have to scale 1080p120 up to 4K/120 which is quite an extreme upgrade.
Yes that was my thought. Both Crysis 2 and Nvidia's 3dvision compatibility mode have the "cardboard cutout" effect.
but some games maybe can take advantage of that characteristic.
VR Paper Mario. VR Visual Novel. etc.
VR Paper Mario. VR Visual Novel. etc.
There are multiple ways to tackle the problem. For turning 2D games into 3D games with a forced driver feature, you're limited in what you can produce as you only have the final front-buffer to work from. Inside the game engine with full access to the buffers, you should be able to reconstruct the two 3D views with near complete accuracy at a fraction of the conventional 2x cost of simple brute-force rendering. With a virtually textured geometry buffer, you'll have a transformation of the buffer for left and right eyes, and then the texture draw of 2x the screen plus whatever overhead. Lighting may be a bit more costly, requiring a full two passes, but then you may be able to get away with faking that with a centralised viewpoint and duplicated lighting. Only in extremes do the eyes see things substantially differently.
.
I see what your saying but of course we need to see what the note 5 will be. That will be mass produced for an October time frame which would allow the same for a consumer rift
They did have a 4k res for 5 inch screen on their road map for 2015 http://www.ibtimes.co.uk/samsung-20...-displays-foldable-screens-smartphones-520643
Similar threads
