LyleLanley
Banned
I can't imagine even Google would want to invest the ridiculous amount of resources it would be required to make such a thing. Would there even be a big market for such a thing?
-
Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.
Microsoft Gave Take-Two's CEO A Demo Of Its Mysterious VR Tech
- Thread starter Detective
- Start date
I can't imagine even Google would want to invest the ridiculous amount of resources it would be required to make such a thing. Would there even be a big market for such a thing?
Not that I disagree with anything you're saying really... but don't you mean that AR is a superset of VR, rather than a subset?
Ol No Bones
Banned
You guys are having one of the most pointless arguments I have seen.
I guess, actually, yes. What I mean is that AR is derivative of VR, ultimately. So yeah, visualizing a venn diagram, I guess that would be superset, not subset.
Thanks for contributing with one of the most pointless comments, then.
Alx
Member
The thing is AR and VR describe functionalities, not technique. "Augmenting Reality" or "simulating a virtual reality" is the purpose, not the technology.
More generally, a product is defined by its use cases and features. And that's also why companies build them and people buy them. Technology is only a mean to an end.
No they don't. VR is a specific technique, as is AR. They are mediums.
And, in any case, you've lost sight of the point of the conversation - the guy is asking why we are concentrating on VR instead of doing work with AR. Obviously, the reason why is because VR forms the foundation of AR. You don't get AR, until you have VR.
Yea, I got what you were saying, in that we'd essentially have to solve all of VR in order to have solved part of AR. Just the word "subset" was sticking out for me.
I completely agree with your points though.
What do you consider 'very basic looking'? Both PS4 and Xbox one could do VR at 1080p/60 at better quality than PS360 and I think that would be a good enough level for me.
I H8 Memes
Member
That would be good enough for me too (hell, much simpler would be). I think the problem was more omonimo acting as if the Xbox One would be too weak to provide any decent VR application, yet the PS4 just so magically happens to be exactly the point where the tech becomes viable.
AR is the ultimate goal. VR will eventually give way to AR. I am excited about VR specifically because it will eventually lead to AR. The possibilities for true AR are unbelievable.
LyleLanley
Banned
Are you saying they could play PS3/360 games at the proper resolution and FPS for VR? I have no idea if that's possible or not but I have my doubts they could do something like The Last of Us at the required resolution with a stable 60 fps.
That... kinda is 1080p though.
I'm not sure how much more of a performance hit multiple perspectives would cause, but outside of that, isn't something like The Last of Us on PS4, and Halo 4 on Xbox One already somewhat achieving something along those lines at 60fps, without even having been initially made with those consoles in mind (assuming we're talking 960x1080 per eye)?
Would something like split-screen Forza 5 not be a pretty good indicator actually?
NotSmartEnough
Member
Surely this is untrue.
A space sim like Elite or NMS isn't going to render the same frame twice. A lot of screen space is going to consist of extremely distant objects (starfield, planets, sun, etc), which are going to look identical for each eye.
I assume devs will simply clone content between images if its distance exceeds a threshold. Only close objects need to be rendered twice.
Sure there will be an overhead to VR, but it's not a straightforward doubling.
not as cool as cake boss
It isn't. I see people people make that claim all over the board I've gotten to the point where I don't correct it anymore. You're entirely correct - there is an overhead, it's not 0-cost nor is it negligible, but it's also not something that requires double the hardware. It's reprojection, there are ways to reduce the performance hit.
NotSmartEnough
Member
It would be interesting to know at what range objects become identical for each eye, I guess it's somewhere around 20-50 metres?
Also, I guess that objects very close to the user, which need the heaviest additional rendering, are either going to be slow-moving or transient, which probably allows for some sleight of hand to reduce processing overhead.
Also, I guess that objects very close to the user, which need the heaviest additional rendering, are either going to be slow-moving or transient, which probably allows for some sleight of hand to reduce processing overhead.
You lost me on the second line. Why would objects closer to the viewer be slower moving? Wouldn't any displacement be far more significant than that same object at a distance?
NotSmartEnough
Member
Sorry, I didn't make much sense!
I was thinking that anything getting close to a user at constant speed would only be really close and need really detailed rendering for a short time before passing the user and moving further away again.
If something was going to be filling your vision for a long period of time then it can't be moving fast relative to you, and maybe devs can get away with duplicating more content between frames.
I don't have any VR development experience, this is just speculation!
Alx
Member
I suppose it will end in an "agree to disagree" situation like with Figment, but like their name makes it clear, AR and VR are about reaching a given experience, whatever technique you use to do it.
To make a comparison with a similar medium, "cinematograph" is a medium, like AR and VR. It's about displaying animated scene to an audience. There are many techniques to create that experience (projected film, cathodic tubes, LCD, Plasma,...). Cinematograph is not a technique, and is not defined by them either. Other mediums could benefit from the development of films, cathodic tubes etc, but they're not necessarily linked to cinema as a product.
One is not the result of the other though. Because they share common technologies, any progress in one will benefit the other, but there is no condition that makes it necessary for one to be fully developed for the other to exist. As a matter of fact, both of them exist already, in different fields. Even if they're still rough around the edges, they both follow their own path, while using the same tools.
SniperHunter
Banned
Vr now in 720p
Your eyes will be sucked
Your eyes will be sucked
No problems, what you're saying is more clear now. I didn't say anything against "transient" for that reason.
That being said though, I think there's a decent amount of complication added in to the equation by the user being able to look around freely. Even if the object was stationary, me turning my head quickly 70 degrees is going to make it anything but slow moving to my field of view.
That said, I also have zero VR (and very little game development) experience myself. So I don't know what I'm talking about here either.
Alx
Member
You're still having a hard time separating product from technology. AR benefits from the technology developed from VR. But it is not an extension of VR, their applications barely overlap.
It's like saying planes are the result of car development. They're not, even if the development of engines that started with cars made planes possible. And both planes and cars could evolve alongside, we didn't need to have the perfect car to start flying in the air.
I don't think this is a very good comparison. I can't think of a single thing in regards to AR that to be sufficiently implemented, wouldn't have required a similar (nigh-identical) problem to have been solved in VR first. Form factor and object awareness are literally all I can think of, and they're only really reasons why VR would happen first.
Because VR isn't a product. VR is a technology. It seems you are having the problem separating product from technology. Oculus Rift is a product. Sony Morpheus is a product. VR is a technology. Microsoft Fortaleza is a product. Google Glass is a product. AR is a technology. I can't walk into a store and buy VR. I walk into a store and buy GearVR. GearVR is a product which uses VR technology. Google Glass is a product which uses AR technology.
What is VR? VR is perspective corrected projections of 3D models rendered with fedility such that they become indistinguishable from tangible, physical products to the degree that they affect subconcious systems in our minds. What is AR? AR is perspective corrected projections of 3D models rendered with fedility such that they become indistinguishable from tangible, physical products to the degree that they affect subconcious systems in our minds, abstracted over reality.
Yes. All this '2x' stuff isn't anything special - if it was a 1080p screen then basically you're rendering side-by-side 3d. It's still just one 1080p screen with one side per eye.
PS3 did stereo 3D last gen on a few games at 720p per eye @60Hz. That's a similar number of pixels being moved around (1280x720 vs 960x1080, approx 1m pixels per eye). And games like motorstorm 3D Rift or apocalyose didn't lose too much detail.
PS4 and morpheus will be fine. So would Xbox one.
And PC will have its own issues - when a stable frame rate is critical and you're targetting so many varied setups, there will be potentially lots of tweaking going on.
I H8 Memes
Member
I foresee disappointment in your future.
Alx
Member
Ok so if we're playing semantics, for me Oculus, Morpheus etc. are each branded product, but they're implémentations of a generic product that is "VR headset". A soda is a product, a bottle of coke is one instance of that product.
Ok so if we go into finer classification and consider VR in general, you can separate it from the object itself and call it a medium. That's still not a technology.
Anyway, by your own definition, VR is defined by giving the illusion of being in a fake 3D world. That is a what, not a how. Whether you do that using technology like inertial sensors and LCD displays, or fictional brain implants, as long as you reach that result and match that definition, you're doing VR.
If AR is defined by inserting fake objects in our perception of the real world, then by definition it is different from VR. It just happens that the current implementation uses similar technologies as VR (inertial sensors, LCD displays etc).
Now semantics aside, there is no reason to consider both applications being dependant from each other. Like I said both will benefit from the evolution of the technology, but both can and do exist at the same time. They don't even have the same hurdles and limitations, latency and resolution not having the same consequences in AR and VR experiences for example.
The how VR is done is the same as how AR is done.
This isn't even remotely true. AR is still constrained by resolution and latency. We can already demonstrate AR, the entire reason it's not ready for the lime light is precisely because of latency and resolution.
When you try high latency AR, you get sick just like you do with VR.
Katzenjammer
Member
Isn't that a very, very narrow description of AR though? You can augment reality completely without the use 3d models, and you don't even have to fool the brain that what it is seeing is real to have augmented reality. I'm thinking of applications such as rendering overlays in real-time sports broadcasts, or even looking at a real-time feed of night-vision video.
What you're describing is more of a hud than augmented reality. Most working with ar would not classify what you described as ar. Going back to my calculator example, you could call a calculator a personal computer, but when people refer to pcs, everybody knows what they're taking about.
The technology called ar will be what I described above going forward.
NotSmartEnough
Member
There needs to be a balance in the definition of AR: I would say that the rendered objects need to be integrated into the real world. Text overlays from Google glass or similar aren't AR, are they? If they are how are they different from subtitles on a TV, or a real-world signpost?
Alx
Member
That's still the original definition of augmented reality. The Playroom or Invizimals are in the augmented reality category too. As long as you're adding to a real stream information that is relevant to its content, you're doing AR. A fighter pilot helmet that points enemy planes is doing AR too, even if it's only drawing tiny squares on his visor.
Full 3D integration may not be a requirement, but context-sensitive localization usually is. Like in my fighter pilot example above, displaying "warning, plane detected !" at the bottom of the visor is just basic HUD, but outlining the plane in the visor image is AR.
In Katzenjammer's example, displaying a football score on screen isn't AR, but drawing the free kick distance as a circle is.
Katzenjammer
Member
I (and apparently Wikipedia) don't agree with this classification. VR is about creating an artificial reality for the human senses. AR is about adding things to the "real" reality and feeding back the results through human senses. The only thing the terms necessarily have in common is that both contain the word "reality".
You seem to be discussing a very specific AR application of realistically rendering 3d elements into a real-time video feed from a camera attached to your face, if I understand correctly? In that case that is only a subset of AR.
NotSmartEnough
Member
Yes, it's technically augmented reality by the definitions of the words "augmented" and "reality." Just as a calculator is technically a personal computer.
When people talk about augmented reality, they are not referring to watching their tv across the room filming their table and putting text over it. When people talk about augmented reality, they are describing what I described.
Do me a favour an break down the process by which you believe you accomplish virtual reality, and the process by which you accomplish augmented reality. They are the same process with the difference being one is put against a black backdrop, and the other is put against a clear screen.
Alx
Member
I think you're too focused on your own activities... there are many people working and focusing on augmented reality in the traditional sense, and not only "what you described".
ReplacementPelican
Member
Nice. What does this have to do with VR?
Alx
Member
Ok, let's try it another way then. I just googled "siggraph augmented reality" to find a state of the art example, and this was one of the first results :
http://www.siggraph.org/discover/news/ar-sandbox-cutting-edge-earth-science-visualization
https://www.youtube.com/watch?v=g6fSS3cynDo
It's definitely augmented reality (adding information to a real image, which in that case is 100% real), and in fact uses technology that is entirely different from what is used for VR.
So what now, are they wrong in calling it AR, or will you claim that is still an extension of VR ?
http://www.siggraph.org/discover/news/ar-sandbox-cutting-edge-earth-science-visualization
https://www.youtube.com/watch?v=g6fSS3cynDo
It's definitely augmented reality (adding information to a real image, which in that case is 100% real), and in fact uses technology that is entirely different from what is used for VR.
So what now, are they wrong in calling it AR, or will you claim that is still an extension of VR ?
To be honest, his black screen vs clear screen example was a pretty good summarization. If we're simply talking HUD elements, then then we're only considering 2D information, in which case every scenario relies on previously solved 2D screen technology (such as the current HUDs we have in games today). As soon as 3D information comes into play (such as mapping the distance between two point of a viewers vision) you're immediately reliant on the techniques that need to be solved for VR implementations, as these elements with require the same kinds of updates to match the users view as any other objects in VR would. The only difference is that you can see our world behind them, and the device you use to view (and interact) with it needs to be a damn sight more portable.
Katzenjammer
Member
That's a leading question. You can accomplish AR and VR without the use of vision altogether.
For example, take an empty apartment building, fill it with synthetic sweet smells and play noises of rustling people. Put someone in there with a blindfold. And bam, you have artificially created an environment from which a person can feel "presence" of being in a candy shop. VR. (Badly thought out example but I really need to sleep, so it will have to do to convey my general thought).
On the other hand, take a pair of shoes and put a vibrator in each. Connect them to a GPS, and program it so that the GPS tells a shoe to vibrate when you need to make a turn in that direction. And suddenly you get AR. No screens involved.
NotSmartEnough
Member
No. I think that this makes the definition of AR too wide.
Forget the vibrating shoes. Just put up a sign with some text to tell you when to turn. That's augmenting reality with extra info, right? What's the qualitative difference from your example?
Even without the use of vision, these both seem like common problems and solutions. Both of these would apply to VR and AR. Solving scent in VR would then solve the same situation in AR, only this time the object in consideration would be placed into our world rather than rendered against a virtual backdrop. A vibrator telling you when to turn would apply both to the AR situation you describe, as well as a VR walking simulator, or even a standard walking simulator on any random screen today.
hey if you don't want to do it I'll just start the company and try to do it and if successful give you 40% of the profits.