Valve's goal for VR - surpassing "immersion" with true presence inside gameworld

#1
VR research made by Valve have identified what technologies are required for computer generated VR content to be able to fool our brain to completely accept visual stimulus and process it as "real" image. Valve will NOT PRODUCE hardware, they are just interested in R&D that they will share with anyone, and will prepare Steam for this VR revolution.



Valve has issued a challenge to the entire entertainment industry, and that challenge has a name: “presence.” Presence is the magic that vr has always promised, and it is what will transform entertainment.

Presence is the new term for a concept most have never experienced, not even those with oculus kits. It is the result of all their research, it is what vr is and what vr does.

Presence is a state in which, due to a variety of factors, virtual reality becomes imperceptible from reality. A number of technological hurdles must be passed before this occurs – 95hz refresh rate seems to be neccessary (but maybe not!). High resolution is necessary. Latency from motion to last photon needs to cross below 20 ms. Pixel persistence of less than 3ms is a must, and is reasonably the upper limit. A wide fov is necessary, at least 85 degrees (the rift is 90). Positional tracking accurate within 5 mm. Much more advanced optics than currently exist. And all this is just for visual stuff, you also need user input, content, etc.

Once all these meld, however, you are in for a wild ride. Once presence kicks in, you cannot stop your subconscious from feeling like it is real. You’ll feel vertigo. You’ll start sweating under the sun. You’ll get goosebumps in the cold. Once these low level functions of the brain are convinced, for all intent and purpose, you are in the game. The holodeck realized.

The demo valve put together here, they believe, is one of the first vr experiences on the planet to approach presence. You’ve read the impressions, people are blown away even if they can’t quantify why. Presence is why. Immersion is nothing compared to presence.

Valve doesn’t intend to produce a vr headset, theirs is an r&d proto. They might, but it’s not in the works. Instead, they’re freely sharing their research with a number of companies who will bring vr to the masses, she they believe oculus will lead the revolution.

“The doom of vr is coming” said abrash, indicating that soon a transformative experience that will define vr is inevitable. Valve sees vr as the most important venture going forward.
Out will iterate on pc, and it will change the world. They believe a consumer vr experience that invokes presence will be available by 2015. Today is day one, the rules of vr are being written right now.

Palmer luckey came in doing clean up, preaching what he’s learned about vr development, what works and what doesn’t. That stuff is widely available online, so I wont cover it… yet. But valve issued the gauntlet – bringing vr to the masses will take the joined effort of the entire entertainment industry. This is larger than one single company. When we arrive, and we will arrive, it will change not just entertainment, but computing entirely.

Ready, player one?
http://thesonicreblog.com/the-virtual-reality-revolution/



As for Steam compatibility with VR gaming, all games will be compatible!!! Some natively with full immersion/presence, but old games will be projected inside "virtual cinema" mode! They will also release Steam VR SDK.
Steam vr has a built in legacy mode for non vr games that projects onto a vr screen. The current vr mode demos this in bpm. Going forward, steam vr will detect when games are running in vr, and when they’re not, they’ll project onto this vr screen automatically, to provide compatibility across the board.
http://thesonicreblog.com/not-done-yet/


Fun times are ahead of us! Especially for GPU manufacturers. :D
 

gofreak

GAF's Bob Woodward
#4
A lot of the things on that slide are projected to be in OR's model in 2014.

1Kx1K per eye = 2million pixels = 1080p, which is the minimum OR is targeting (albeit in one panel - I don't know if there's a quality difference per eye in having 2 720p panels vs 1 1080p)
sub-20ms latency is on the cards for OR
3ms pixel persistance = what OR showed in Crystal Cove?
110-degree FOV = OR's diagonal FOV

I guess tracking and optics quality are a somewhat fuzzier issue though.

edit - a virtual screen for non-VR games via Steam is great
 
#5
I think that they think 1K x 1K resolution per eye is not for "one half of display", but really for visible area that eye is seeing. Current Oculus Rift lens is focused on smaller surface of one half of a screen, making even 1080p screen grainy.

With current lens Rift structure, they could possibly approach 1Kx1K with 1600p screen.

You'll need to control virtual hands to have a presence in a game world.
That is not the case for driving/flying simulators.
 
#10
Conspiracy theory time.

Valve is designing HL3 with ful VR in mind. Its non-VR version will be a shell of the real game.

Valve says VR will have its doom moment in 2015 because Valve is in fact revealing or releasing its doom equivalent in 2015 for VR (read HL3).
 

gofreak

GAF's Bob Woodward
#11
I think that they think 1K x 1K resolution per eye is not for "one half of display", but really for visible area that eye is seeing. Current Oculus Rift lens is focused on smaller surface of one half of a screen, making even 1080p screen grainy.

With current lens Rift structure, they could possibly approach 1Kx1K with 1600p screen.
Oh cool. That makes a bit more sense as a target then rather than something happening now or this year perhaps.
 
#20
The most interesting part about that slide in OP, is that all the hardware is already there. If Sony can make 2096PPI full OLED in large quantities, anything can be made by tons of other companies. Persistence modes with OLEDs are solved thanks to their instant switch time, motion-to-photon latency can be solved, refresh rate, fov, optics and tracking...

All of this is not only already available now, but also very cheap. They just need to take individual parts, and optimize them in ways that nobody has optimized them before.
 
#21
All of this is not only already available now, but also very cheap. They just need to take individual parts, and optimize them in ways that nobody has optimized them before.
People like Carmack are optimising right now. I think the biggest bottleneck will be the software and user uptake which will be slow because of an initial lack of software, Also to push a 2000 x 2000 combined resolution at a good framerate you will need a decent GPU.
 
#22
The vr room sounds cool. I'd love to stand inside a globe with the entire Civ 5 map projected around the globe and I get to issue commands like Enders Game.
 
#23
Also to push a 2000 x 2000 combined resolution at a good framerate you will need a decent GPU.
As long as mobile devices can catch up in rendering simple gameworlds, all will be well. First VR games on PC do not need to be visually complex. First Oculus Rift has already proven that great immersion > great graphics.
 
#24
I think that they think 1K x 1K resolution per eye is not for "one half of display", but really for visible area that eye is seeing. Current Oculus Rift lens is focused on smaller surface of one half of a screen, making even 1080p screen grainy.

With current lens Rift structure, they could possibly approach 1Kx1K with 1600p screen.
.
Yeah, I was going to reply the same. If the are hyping it so much, talking about "presence" as a kind of nextgen VR... I don't think you can reach "presence" until you have 1080p per eye VR, possibly even a bit more.
 
#25
As long as mobile devices can catch up in rendering simple gameworlds, all will be well. First VR games on PC do not need to be visually complex. First Oculus Rift has already proven that great immersion > great graphics.
True, being in VR already gives a great sence of immersion, your brain kinda enhances what you see so it all feels real. Its like looking at a stylized theme park world.
 
#31
Almost certain HL3 will be exclusive to these VR headsets.
I doubt it, there are to many limitations that human beings have that might limit the sense of VR they can experience, and thus you are limiting your target audience even further. I wouldn't be surprised if it has a VR mode trough that is many steps beyond current VR implementations.
 
#32
Almost certain HL3 will be exclusive to these VR headsets.
Action based FPS VR games are far from being perfect right now. Controlling them, users sitting in their chairs while experiencing standing character, especially if gameplay is very fast...

VR will bring a new slate of game genres. HL3 will most surely be traditional FPS game, possibly with great VR support. But it wont be VR exclusive.
 
#40
Wouldn't it be 1000 x 2000?
I think 3200x1600 if you want 1k visible resolution per eye (the optics distort the edges of the screen resulting in a very blurry low resolution image, so the effective area that you actually see is much smaller)

the green area is what you actually are able to see in the rift and is wrapped around a FOV of 90. The rest is blocked trough the plastic optics cone, and the piece of plastic that separates the screen in 2 half's, and the imperfect optics
I do think the optics will improve, and the visible area on screen will get better in the consumer VR devices, but higher resolution is always better.
 
#41
Conspiracy theory time.

Valve is designing HL3 with ful VR in mind. Its non-VR version will be a shell of the real game.

Valve says VR will have its doom moment in 2015 because Valve is in fact revealing or releasing its doom equivalent in 2015 for VR (read HL3).
People have been speculating this for a while now, but looking at the slides, reading blog impressions I have my doubts it's going to be 100% VR.

1. Oculus themselves are targeting seated VR for now. I'm not sure how a seated VR FPS experience can achieve great presence and be a selling point for VR.

2. Valve's tech demo revolved around walking in a room. If they gear up HL towards that it's going to need a big infrastructure, unless they gimp it to being full of very very small rooms and corridors. Like how could you translate the Strider battle from Ep2 to this? Not to mentions as it's HL, it's probably going to have vehicle segments. You livingroom or bedroom isn't going to cut it.

I think it's going to support VR, but won't be a full experience.
 
#42
1920x1080p is on target for day 1 acceptable VR.
Each eye will receive 960x1080 = 1 036 800pixels (vs 1 000 000 which is Valve goal).

PS4 is just above threshold for acceptable VR.
 
#43
I don't understand the current hype with VR. To me, it seems like another "3D" gimmick. And, people complained about how heavy and uncomfortable the 3D glasses are. VR headsets are at least 10X worst comparably. Until VR headsets get down to at least 3D glasses size, it will be going nowhere.
 
#45
I've been thinking for awhile that the next evolution in arcade gaming could potentially be in the VR realm. If we ever reached a point in which full immersion was feasible (controlling run movements, hand movements, etc.), then a big open space would be required; living rooms and bedrooms wouldn't cut it.

Arcades could once again provide an experience that would be impossible to deliver to the average consumer's living room.
 
#47
I don't understand the current hype with VR. To me, it seems like another "3D" gimmick. And, people complained about how heavy and uncomfortable the 3D glasses are. VR headsets are at least 10X worst comparably. Until VR headsets get down to at least 3D glasses size, it will be going nowhere.
Don't knock it until you've tried it dude. 3D on a flat screen is a gimmick. It doesn't feel natural, its just a fun little effect that gets old pretty quick. VR is not like that at all.
 
#48
I don't understand the current hype with VR. To me, it seems like another "3D" gimmick. And, people complained about how heavy and uncomfortable the 3D glasses are. VR headsets are at least 10X worst comparably. Until VR headsets get down to at least 3D glasses size, it will be going nowhere.
VR will blow many mind.
 
#49
1920x1080p is on target for day 1 acceptable VR.
Each eye will receive 960x1080 = 1 036 800pixels (vs 1 000 000 which is Valve goal).

PS4 is just above threshold for acceptable VR.
If you are going that literal with the slide interpretation, then what about 95 Hz?
 

GraveHorizon

poop meter feature creep
#50
But seriously, sounds cool. And two cubed meters of tracking space? Are they expecting people to get up and walk around?
That's roughly 1.26 x 1.26 x 1.26 meters, which isn't really walking around space. Seems like enough space to move your head around in most directions from a seated position, or ducking and dodging from a standing position. I read it as "wiggle room".

EDIT:
I don't understand the current hype with VR. To me, it seems like another "3D" gimmick. And, people complained about how heavy and uncomfortable the 3D glasses are. VR headsets are at least 10X worst comparably. Until VR headsets get down to at least 3D glasses size, it will be going nowhere.
The "3D gimmick" was minor annoyance (glasses) for minor payoff (debatable simulated depth window). I put up with the glasses for every movie I want to see just because it's in3D because I'm simply interested in the effect, and I'll admit that it's not always great or even utilized well (nothing has beaten Avatar and Jurassic Park 3D). For normal people, the depth of an average 3D movie probably isn't worth the cost of comfort of 3D glasses, which honestly isn't that big of a deal for a guy like me who already wears glasses.

VR on a head-mounted display like the Rift is medium annoyance (entire headset) for major payoff (presence in simulated world). You see it as if you are there, not simply looking at a window into a fake world like with 3D movies. You turn your head and the world changes, just like reality. You look all around, and the world is all around you. Sure, a headset is much more weight on your head than a cheap pair of plastic glasses that weigh almost nothing, but the 3D effect is also much more pronounced because it responds based on your movement. As long as the game/simulation is convincing enough to trick a user's brain into "feeling" as if they are in that world, and the quality of the game or simulation is entertaining enough, people will be interested in trying it. And the form factor should get better over time.

It seems like most people with something bad to say about VR either absolutely hates the idea of wearing something like glasses/a headset, or just doesn't really care about VR to begin with (my friend is both).