Howdy, Stranger!
It looks like you're new here. If you want to get involved, click one of these buttons!
Quick Links
Where is 3D in MMOs and Games?
Elikal
Member UncommonPosts: 7,912
I am 41. When I was a kid in the 1970ies, I imagined the year 2000 WAY different. Today, that I *live* in 2012, I wonder if I am the only one who is disappointed in the future? I expected humanity has the first Moon Colony, Anti-Grav Vehicles, Fusion Power, near-lightspeed space craft and Holodecks.
Ok, Holodecks were more 80ies. But still. If we look at computer games and especially MMOs, I often think "Where the heck is all the progress"? I mean sure you can go to content and we discuss that often enough. But here I want to raise the question of technology. Yes, we have 24 inch flatscreens and no longer 640 x 480 resolutions. But ... where is 3D?
On the Berlin "Funkausstellung", one of the biggest technology fairs, I had seen a sort of 3D video glass I dunno 12 years ago or so? Back then I was so hopeful to see something like that becoming regular soon. Now, we still look at these flat games. Sure, the Holodeck is still far away. And sometimes I feel like progress has been faster once. Or maybe just our dreams about what is possible. I'd love to see games in 3D. I recall like about 20 years ago a small 3d Cafe opened in Berlin, where you could wear a huge helmet and move in a simplistic 3d environemt. I was a student back then, and was so awed and imagined how cool 3d Computer games would look in 20 years. Now 20 years passed, and nothing the like has happend.
Does anyone else know the Anime "Sword Art Online"? If you like Anime and MMOs you might watch it. It's about exactly that, people wearing a "Nerve Gear" and are actually physcially feeling the 3D environment. Wouldn't that be cool? I mean, ok that real brain stimulation would be a bit extreme. But a real 3D environment via some helmet would be cool. Something I definitely would love to experience in a MMO or single player game.
That and the first real AI robot, I always wanted to create. lol. XD
People don't ask questions to get answers - they ask questions to show how smart they are. - Dogbert
Comments
Its coming.
http://www.kickstarter.com/projects/1523379957/oculus-rift-step-into-the-game
Yeah I remember a VR headset and playing Dactyl nightmare in the early 90's on Amiga's. Ive spent the intervening years laughing at Windows machines and their users. At least they are finally starting to catch up - just look at how much backing the Oculus system got
Edit: Maybe we used the same system - here is a link to games it had. Only played the one myself.
http://www.amigahistory.co.uk/virtuality.html
I want 4D and smell-a-vision. We have 4G, they passed over 4D.
I hate wearing anything on my head when watching TV or on the comp, tried those 3d glasses and in best buy last year on those 3d tv's and was not impressed. I guess its a good gimic for people with kids but other than that really not interested.
There are a variety of problems. Nvidia started pushing stereoscopic 3D a few years ago, but it was proprietary and required a special monitor. AMD soon offered their own proprietary stereoscopic 3D. But proprietary means a major pain to support, as you have to do different things to support hardware from different vendors.
OpenGL 4.2 brought stereoscopic 3D to an industry standard API last year. DirectX 11.1 caught up this year. But DirectX 11.1 is exclusive to Windows 8 and doesn't bring anything important, which means it's destined to irrelevance.
OpenGL 4.2 doesn't have that problem, as it only requires Radeon HD 5000 or GeForce 400 series or later graphics and runs on a lot of different operating systems: Windows XP, Vista, 7, 8, Linux, and likely soon Google Android. (There currently isn't any Android hardware that can run OpenGL 4.2, but that is likely to change very soon.) But not Mac OS X, as Apple can't be bothered to write good video drivers. The problem there is that few commercial games use OpenGL at all; mobile games commonly use OpenGL ES, but that's a different critter entirely.
There are also hardware problems. You need to render each frame separately for each eye. That means you need double the frame rate. A lot of games are sufficiently processor intensive that, while it's easy to get the 40 frames per second on any reasonably modern CPU, the steady 120 frames per second you want for stereoscopic 3D is much harder and often impractical.
If you're designing a game to support stereoscopic 3D in the game engine rather than an external vendor trying to hack something together in drivers, there are a lot of things you can do that improve the situation. You can time things so that both eyes see the same frame at the same time, rather than making frames alternate between each eye. This lets you share a lot of the CPU work between the frame for each eye so that it's still a lot more than normal, but a lot less than double what it would be without stereoscopic 3D.
I haven't investigated how stereoscopic 3D is implemented in OpenGL or DirectX, but if you could set up your vertex data and textures and upload uniforms once and then render to both frame buffers, that would mean that the video card work would be a lot less than double, too. Some uniforms would have to still have to be uploaded separately for each eye, though, as a given object would be at a different position and orientation relative to each eye. Some objects would be culled so that they'd only be rendered for one eye. And, of course, the actual GPU computations would be doubled, so you're really only saving on the GPU time of switching vertex data, textures, and uploading some uniforms. But that's still substantial.
I don't know if the APIs are set up to allow that, and if they're not, I don't know if it is practical to do so in the future. Even if they're not set up that way, it's probably possible to hack together something that offers much of the benefit by rendering both eyes to a single framebuffer object and relying on geometry shaders to do manual clipping and draw everything for both eyes. But again, I haven't tried it.
But the hardware to actually view stereoscopic 3D images is pretty dismal. Polarity glasses mean you lose half of the vertical resolution, which renders text unreadable, among other problems. Active shutter glasses don't work very well, either, as it means that only one eye can see anything at all at a time. What's really needed is to have a separate "monitor" for each eye, like the Oculus Rift that atuerstar linked, and we'd need it to actually work well. And we'd also need for it to not be a major pain to wear. And we'd need it to hold still and stay in the proper position, as being a quarter inch off to the side will completely break the effect.
Is that practical? Eventually it probably will be. But not necessarily soon.
But there are other problems with stereoscopic 3D that aren't so easily solved. For example, what do you do with fundamentally 2D things, such as the HUD? The headaches that it often gives people will probably diminish as the images get closer to "correct", but how long will that take?
When guys like John Carmack are gushing about the Oculus Rift, I think they may be on to something. At the very least, they are pushing it forward further than anyone else has, and I don't think we are far off at all.
I also believe that once we have the right hardware, the 3D internet will start taking shape finally.
A sure sign that you are in an old, dying paradigm/mindset, is when you are scared of new ideas and new technology. Don't feel bad. The world is moving on without you, and you are welcome to yell "Get Off My Lawn!" all you want while it happens. You cannot, however, stop an idea whose time has come.
Thank you for the technical viewpoint Quizzical - a pleasure as always.
Ive used shutterglass's and stereoscopic and never found anywhere near the level of immersion I got from twin monitor (one for each eyeball). I think its power lies in tricking the mind in a way, something the other presentations of 3D seem to fail at. No matter how much tomfoolery occurs between you and the screen you are always aware there is you - something in the middle - and then screen. The eyes never seem to really relax into it and therefore neither does the mind.
With the full headgear your field of view is eclipsed and neither eye is aware of what the other eye is seeing, its two seperate feeds that composites in the mind, and I feel this is the reason it succeeds so immensely well at its job. Its leaving nothing for the mind to do but its usual job, combining seperate 2d streams of information into one 3d perception of reality.
I think the real answer here its because they still making money with obsolete technology, VR probably exist for a long time by now but why big corp will spend money on it if they can still sell us the old things.
Look the theater, they still use the movie projector a old piece of techno who has what? 100 hundred years? Some years ago they began to use the digital cinema projection , again this is old school. Why we dont have theater with big LCD screen? cause they still making a load of money with this old scrap and will squeez every bit of money before invest on some better techno.
The same things happen with video game or any techno. They sell us the old and cheap stuff so they will sell us the new stuffs many years from now when it will be old and cheap.
The way that the human brain naturally does stereoscopic 3D is that both eyes see the same scene at the same time, but from different viewpoints. Neither active shutter glasses nor polarity glasses can do that.
With active shutter glasses, one eye sees a scene while the other eye sees nothing, then the other eye sees the same scene a bit later while the first eye sees nothing, and then the first eye sees the next frame of the scene while the second eye sees nothing. I don't know the internal details of how human vision works, but it strikes me as likely that either one eye seeing things happen 1/60 of a second before other or each eye seeing nothing half of the time would seriously mess it up.
With polarity glasses, the problem is that the eyes don't see the same scene at the same time, but rather, different portions of the same scene. With a high enough monitor resolution, this likely wouldn't matter. But when you start with a monitor resolution that isn't that large and cut the number of vertical pixels in half, anything that is one pixel tall (e.g., horizontal lines in the text that I'm typing right now) is completely invisible to one eye. And that's probably going to cause problems.
A separate "monitor" for each eye avoids both of those problems. But it could easily create others, depending on how it's implemented. If it weighs a pound, it's not going to be comfortable to wear it for hours on end.
-----
Upon further thought, I think the ideal way to do stereoscopic 3D would be to have both left and right framebuffers active at the same time, and the geometry shader would specify which framebuffer a primitive it emits should correspond to, and then the fragments produced by rasterization would write to that framebuffer and ignore the other. That would avoid any need for manual clipping in geometry shaders, which is never going to be as fast as the fixed function clipping that follows. Or if you're using post-processing, then the geometry shader would specify which (set of) framebuffers a fragment shader should write to.
It would also mean that there's absolutely no additional load whatsoever imposed by stereoscopic 3D in the first three pipeline stages (vertex shaders, tessellation control shaders, and fixed-function hardware tessellation), and that you do much less than double the GPU load for the next stage (tessellation evaluation shaders). Well, maybe not vertex shaders; that depends on how you implement it, and it's something I'd have to think some more about the optimal way to do that. I do think you'd want every object tessellated in exactly the same way for both eyes. But regardless, if you're doing tessellation, vertex shaders don't put much of a load on the GPU; then again, neither do tessellation control shaders.
Would it be practical to implement that? I don't know. It might already be implemented that way, as I haven't looked into it. But even if it isn't, many things that are impractical to implement after the fact in video cards already on the market are practical in future generations if you design the silicon around wanting to implement it.
Velika: City of Wheels: Among the mortal races, the humans were the only one that never built cities or great empires; a curse laid upon them by their creator, Gidd, forced them to wander as nomads for twenty centuries...
I actually bought one of the first sets of Nvidia 3d glasses nearly a decade ago. I remember playing Eq2 Beta and Eve using them and it was awesome. What made me put them away was that I "upgraded" to LCD monitors that didnt have the refresh rate to support them. =/
I too am watching the oculus rift with great anticipation. If it delivers I will own one the day they go retail.
Ahh yes this 1000CS looks like the one I used 20 years ago or so. ^^
Cool, that Rift looks nice. I hope it will be affordable.
People don't ask questions to get answers - they ask questions to show how smart they are. - Dogbert