Sony VR Headset/Project Morpheus/PlayStation VR

The only AAA games that will be made during this generation will probably be those that work well with or without VR. I guess FPS and Racing Games are a given, but I fear it will be worse than 3D where gamers were immediately aware of the sacrifice in quality they are giving up when switching, hopefully the advantages will be more significant.

The more gimmick and fun experiences might be really cool, and I hope indies will have fun with this. They need fun stuff for launch to demonstrate the possibilities, not just a popular AAA franchise.
 
I really want to see a game that looks like TLoU, and runs at a stable frame rate, on a 256mb ram PC with a gimped 128bits nvidia 7800.

That's an unfair comparison since the 7800 hasn't been actively supported in Nvidia drivers for years. You'd be better looking at how a modern architecture with similar performance plays an equivalent (ish) game. The closest comparison I could find is a 720m (somewhat weaker than RSX with a crazy low 14.4GB/s bandwidth) running Tombraider 2013:


So I'd conclude that yeah, if the 7800 series was still fully supported by Nivida today and still targeted by developers like modern architectures are then it should be capable of giving a reasonably comparable performance to the PS3 in modern games.

My interest here is that journalists didn't see a significant difference between vive on a recent $1500+ PC, compared to a $400 PS4 from 2013 with morpheus. 2 to 3 times more power is quite a step.

But they're comparing different applications so we haven't actually seen a direct, head to head comparison yet in the same application. Also Crescent Bay sounds like it's a step above Vive in image quality (likely a result of the higher resolution). Clearly though Sony are doing something right to be potentially competitive with Vive in image quality despite Vive running at a slightly higher resolution.
 
There seems to be many factors. Console fixed hardware, unified memory, timewarping from 60 to 120 instead of 90 is a 50% gain (with some positional tracking compromises), the FoV with more density in the middle, lower resolution but with full RGB pixels, and in theory a higher lens quality. I suppose all of this should sum up to get surprisingly close in terms of the perceived Image Quality Vs vive, but not quite up to oculus CB.

I don't believe for a second that crescent bay has two separate screen at 2560x1440. Most likely they split them to improve refresh with two controllers, and/or improve yield with two small screen, and/or improve optical properties with a non-coplanar arrangement.
 
They should target more stylized graphics instead of realism. Even with the complaints I raised about Oculus DK2, the impression of depth around you is very engrossing. With a lot of the complaints I raised fixed, it could be stunning. I just think it would be a mistake to chase realism, because the sacrifices to get to higher locked frame-rates will be great. Probably the worst demo I tried on DK2 was that Tuscany demo you always see.
 
The best the PC VR makers can hope for is that Project Morpheus is a success so that big teams will work on VR games & port them to the PC VR sets.
 
The best the PC VR makers can hope for is that Project Morpheus is a success so that big teams will work on VR games & port them to the PC VR sets.
As I've posted before, I think this is going to be the biggest hurdle. On PC VR is going to be an expensive proposition and the Steam Hardware survey suggests that it's going to be a niche platform in what already seems to be a market smaller than exists in the console space.

But even Morpheus is going to be an investment even if you already own a PS4. It's going to need to sell itself to work. People have to try it and need to buy the damn thing straight away. They have to nail it. Then, and only then, will VR work because I think to get the best of out it you need to design games specifically for it and the commercial returns aren't going to be great unless there is a massive install base.
 
As I've posted before, I think this is going to be the biggest hurdle. On PC VR is going to be an expensive proposition and the Steam Hardware survey suggests that it's going to be a niche platform in what already seems to be a market smaller than exists in the console space.

There are about 125 million registered Steam accounts (i.e. PC's being used for some level of gaming) so that's a pretty big potential market compared to the 20 million or so PS4's out there. The difference between Morpheus and OR for me is that OR is likely to be at least, if not more so focused on non-gaming applications as it is on gaming, and that may give it an advantage in terms of consumer uptake. There's also the very real (likely) possibility that Nvidia/AMD will enable conversion of non-VR games to VR through drivers (as they do/did with 3D gaming) which would open up a huge library of VR enabled games that won't be available on consoles. No doubt the quality of the VR experience in those games though will be lower than dedicated VR games to varying degrees.
 
There are about 125 million registered Steam accounts (i.e. PC's being used for some level of gaming) so that's a pretty big potential market compared to the 20 million or so PS4's out there.
Now look at the actual hardware and tell me how many are going to be able to drive 90Hz dual 1080p screens at any appreciable quality. 48% of PCs only have dual core CPUs and a third have 1Gb VRAM or less.
 
Now look at the actual hardware and tell me how many are going to be able to drive 90Hz dual 1080p screens at any appreciable quality. 48% of PCs only have dual core CPUs and a third have 1Gb VRAM or less.

Full resolution/frame rate isn't a requirement to use VR, it's merely the ideal (for a better than Morpheus experience). But as I said above, there will likely be lots of non-gaming applications for OR which won't require any kind of significant hardware.
 
Full resolution/frame rate isn't a requirement to use VR, it's merely the ideal (for a better than Morpheus experience). But as I said above, there will likely be lots of non-gaming applications for OR which won't require any kind of significant hardware.

With FaceBook owning Oculus, you can count on it. Personally I'm it in for gaming (and possible movies) and that's what I'm getting from most people here, hence the focus on resolutions and framerates.
 
Full resolution/frame rate isn't a requirement to use VR, it's merely the ideal (for a better than Morpheus experience). But as I said above, there will likely be lots of non-gaming applications for OR which won't require any kind of significant hardware.
Very high frame rate and very low latency ARE a requirement to avoid vomit, and so far 1080p is the minimum to avoid having the pixel structure too distracting. I agree that apps will easily get away with simple graphics, but I find it hard to imagine a real system seller that isn't a game.
 
Very high frame rate and very low latency ARE a requirement to avoid vomit,

Those are terms that need defining, however I'm quite sure that neither the 1440p resolution nor the 90fps frame rate will be standard in PS4 VR games. Therefore regardless of whether you require "Very high frame rate and very low latency" for VR, as I stated in my previous post "Full resolution/frame rate (by OR standards) isn't a requirement". And since it was the specific OR maximum capability you made reference to in your post about the hardware requirements for such an output, the point does stand.

and so far 1080p is the minimum to avoid having the pixel structure too distracting

For the sake of argument I'll agree with that but the original point you made was: "how many are going to be able to drive 90Hz dual 1080p screens". So it seems we agree. You do not need the full resolution of OR to justify VR. Therefore the question of "how many are going to be able to drive dual 1080p screens" as an argument against OR seems irrelevant.

I agree that apps will easily get away with simple graphics, but I find it hard to imagine a real system seller that isn't a game.

Have you ever tried p0rn in VR?
 
That would also kill many PC configurations. :)
If valve is seeing such nice increases in performance with the new crossfire for vr then we might only need a 290x in crossfire to play. By the end of the year that performance should be under $200 a card or we could always upscale
 
Have you ever tried p0rn in VR?
Nope and I don't get this personally. Porn typically exists to satisfy different needs/urges than games. In an intimate situation part of that being satisfying is being able to uh, 'interact'. I can't see how VR porn is anything other than look but don't touch. While each to their own, for me personally this sounds more frustrating than stimulating. I can already do frustration without technology ;)

In a game I don't want to feel what like it's like stab or shoot somebody and that isn't what VR brings to the table for me. It's about removing the barrier between me and the environment my avatar is in. Improving my situation awareness. I don't want to have to press a button to lean around a fall, I want to just lean and see around the wall. The Heist demo on Morpheus shown at GDC looked awesome. There are people ducking and crouching behind furniture only they can see. That's what VR is for me - better immersion.
 
Maybe not port but a LovePlus vr edition might work. Heck, Live2D already doing wonders for bringing drawings into 3D
 
"only 290x in crossfire" :confused:
I don't think it will take that much, honestly. If "The Order 1886" is rendered in 1920x1080 30fps, then that level of graphics should be achievable on the speculated VR hardware with less than 290x in crossfire.
Furthermore, the VR hardware is not out. Both nVidia and AMD are poised to release new high-end cards that perform 50% or so better than the 290x, while still on the 28nm node. Which is nice, because one of those will do well alone, and when nVidia/AMD move to 14/16nm FinFet, we can probably assume that those products will perform better still, in a time frame closely aligned with VR hardware release.

I'm not buying any new gfx hardware on 28nm. The next lithographic step will be a game changer in performance/power, and even if it is largely first taken advantage of in power draw, that is still fine with me. One such card should do very well indeed with first generation VR hardware, but if VR product go a bit higher than expected (for instance higher res) then you could consider adding a second card for better performance. Mind you, at that point we'd be talking 3-4 times the performance of the 290x or some 10 times the performance of the PS4....
I really can't see that we will have a problem with reasonably priced graphics horsepower to drive our headsets, one year out from now.
 
Last edited:
Porn typically exists to satisfy different needs/urges than games. In an intimate situation part of that being satisfying is being able to uh, 'interact'. I can't see how VR porn is anything other than look but don't touch. While each to their own, for me personally this sounds more frustrating than stimulating.

Never been to a strip club then I see. Trust me, there's a big market out there for this. That said, the porn example was only 1 semi-serious example of how there could be many non-gaming applications for VR. Gaming will, at least initially, likely be the main application but on PC at least, I expect considerably greater diversity than just gaming.
 
Back
Top