Support NeoGAF

[Feep]

Late Monday night.

My old college roommate and I had finished watching the final round of the Omegathon, gotten some dinner, and driven twenty-five minutes back to Kirkland. I had a flight the next morning at 7 AM; I wanted some sleep.

Then, oddly, I got a phone call from my buddy asking me where I was. Far away and tired, I replied. He told me unequivocally to come back, despite having just ten minutes before arrived home. I refused. He insisted.

"Come the fuck back," he said, "and pick me up, and we're gonna go see this guy. Trust me."

Back in the car we went.

We went to see Jamie Kelly, the CEO of VRCade. I had briefly heard of them in a GameSpot article awhile back. I'm a big fan of the Rift, having received one back in April, and I've become fairly sensitive to input lag and latency issues, so I didn't think I'd be terribly impressed.

I was terribly impressed.

The VRCade slightly-hacked-together set (you can see the Oakley symbol on the headstrap, having originally come from a higher-end pair of ski goggles) has a better screen than the Rift Dev Kit. It's actually the same resolution, but the screen door effect is greatly lessened and the pixel response time is notably better. I'm sure the Oculus Rift HD prototype might equal or exceed this advance, but I've never tried it.

The real key, though, is the positioning system. VRCade isn't actually built to compete with the Oculus Rift, because the Rift is meant to be an at-home device, affordable for most middle-class families. There are a lot of inherent limitations with that device, notably lack of positional tracking and one that isn't mentioned much, the tethering wire. There's really no getting away from that strict leash, which not only limits freedom of movement, but even prevents you from spinning in place more than three-hundred-and-sixty degrees. Even the folks making the Omni Virtuix treadmill said they were experimenting with an overhead boom to hold the wire, but that it was a fair challenge.

VRCade uses somewhere around $16,000 dollars worth of equipment for even a small setup, but it effectively turns an empty space into a live motion capture arena. Twelve cameras are mounted overhead, around the space in question. The headset, gun, and whatever else you care to stick little pebbles on are then tracked in real time, both in orientation *and* position. As a result, you can move, bob, weave, duck, jump, whatever, so long as you stay within the capture volume, and the results are incredible. The source of motion sickness with the Rift, the sensation of the inner ear not quite matching up to the eyes, disappears entirely, because there is no longer any effective incongruity between the two. You feel free to move, to spin, whatever, because the wire from the headset simply attaches to a worn backpack (they plan to eventually ditch the backpack as well). And it honestly blows the current incarnation of the Rift out of the water.

I still love my Rift, and this isn't really a solution anyone could ever really obtain at home. But its name is accurate...this thing could completely revive arcades as a cultural force. They want to set up large arenas from which you could rent areas that correspond to in-game space limitations, or perhaps try a hybrid approach where certain longer distance travel is done using alternate means. It's just...extremely cool.

HERE IS A VIDEO OF ME DOING STUFF. My old college roommate is a horrible person and recorded it vertically. I am very very sorry. Feel free to ask questions!

Also, I was like, "Why weren't you at PAX?" They apparently just never got back to him.

Their loss.

 

[Kuro]

This will be quite popular in places like malls for about a year but then will end up as popular as laser tag is nowadays.

 

[senahorse]

Interesting, I would love to dedicate a room to this, but not at that price. I have wondered before if you could do something similar with a few kinects for mapping out a room and then applying textures to obstacles such as couches, walls etc.

 

[Riki]

I want to do this so much. Looks amazing.

 

[dallow_bg]

Fun video to watch.
I'd love to try it.

 

[zephervack]

Damn this is exciting but I really dont see anyone making worthwhile software for it.

 

[Zaptruder]

Cultural force nothing.

OR consumer version + Kinect 2.0 > VRcade.

Seriously, this tech is moving fast... and the limitations of walking in a room vs walking on a spot aren't nearly as big as you might think.

i.e. use kinect to track the leg motion - simulate walking on the spot as walking forwards (maybe in combination with the direction of the analog stick).

This is assuming that 1 kinect 2 can accurately track a human from all orientations (assuming that the body is obfuscating itself - but then you'd still want to pair the motion input from the kinect with a hydra style device, so using the local positioning of the hands holding the hydras and inverse kinematics to solve for the occluded body parts, it shouldn't be a problem at all).

I mean... the leg-hip swing that a large room or an omni directional provides is nice, but with walking on the spot, you still get the semi-accurate proprioceptive feedback (i.e. muscles contracting up and down, feet striking the ground) and natural headbob (i.e. the VR headset will literally bob up and down on the user's head as their feet move up and down) which when coupled with good visual feedback should prove to be an extremely immersive sensation.

After all, even with a large room or omni treadmill/board, there are elements of movement that aren't captured properly; such as moving up and down inclines and steps.


So using the kinect, the solution to providing the player with an adequately safe and immersive experience is to calibrate the player's position and create a centered location for them to home in on... as the player moves away from the spot, the software informs them of this (i.e. you are leaving your activity zone, please move back *uses on screen indicators to show them where they are relative to the 'home point'*).

Similarly, the kinect can show them obstacles should they start getting close to them (e.g. moves too close to a chair and might trip over; the chair ghosts into the view of the user - or even a person that moves in too close might be shown so that the user can avoid them if necessary).

 

[senahorse]

You're reading my mind Zaptruder, the next few years could provide a great leap in immersive gaming.

 

[chris121580]

This will be quite popular in places like malls for about a year but then will end up as popular as laser tag is nowadays.

Yes as popular as laser tag...which is still located all over the place and is insanely popular

 

[Lactose_Intolerant]

needs the omnidirectional tread mill and would be perfect

 

[Iced_Eagle]

I live in Seattle. I'd love to try this out!

How did your buddy get in contact with him to get a demo?

 

[Feep]

OR consumer version + Kinect 2.0 > VRcade.

I remain fairly unconvinced that even a Kinect 2.0 could replicate the kind of experience shown here...at least, only one of them. Using inverse kinematics to solve for certain occluding body parts might be okay, but it could never allow for a freeform environment where players can face any arbitrary direction. If the player is facing away from the Kinect, any props are going to be completely blocked. And the more players involved, the rougher it's going to be.

The omnidirectional treadmill and Razer Hydra do solve some of the problems, but now you're up at approximately one thousand dollars worth of tech, for *one person*. If you want to play multiplayer over the net, player avatars could be created using a Kinect 2.0, but that's *another* piece of tech required, and a headset for voice chat, and now you've got a jillion wires of different lengths you've got to manage spinning around in a probably crowded space and anyone who comes to visit thinks you're some kind of insane mad scientist. It's just a *lot* to put together for each individual involved.

For what it's worth, I got to try the Razer Hydra during PAX and came away thoroughly, thoroughly unimpressed. I couldn't do anything that required even the manual dexterity of a three year old.

Damn this is exciting but I really dont see anyone making worthwhile software for it.

Actually, it's just stuff in Unity. Assuming they get sent the source project, they can apparently just replace the Oculus Rift camera prefab with their own, and it'll just work.

Of course, a game should *probably* be designed with full motion and the relevant space in mind, but it's not like they'll need some dedicated team just to create a single piece of software.

 

[Zaptruder]

I remain fairly unconvinced that even a Kinect 2.0 could replicate the kind of experience shown here...at least, only one of them. Using inverse kinematics to solve for certain occluding body parts might be okay, but it could never allow for a freeform environment where players can face any arbitrary direction. If the player is facing away from the Kinect, any props are going to be completely blocked. And the more players involved, the rougher it's going to be.

The omnidirectional treadmill and Razer Hydra do solve some of the problems, but now you're up at approximately one thousand dollars worth of tech, for *one person*. If you want to play multiplayer over the net, player avatars could be created using a Kinect 2.0, but that's *another* piece of tech required, and a headset for voice chat, and now you've got a jillion wires of different lengths you've got to manage spinning around in a probably crowded space and anyone who comes to visit thinks you're some kind of insane mad scientist. It's just a *lot* to put together for each individual involved.

For what it's worth, I got to try the Razer Hydra during PAX and came away thoroughly, thoroughly unimpressed. I couldn't do anything that required even the manual dexterity of a three year old.

To be fair, the current hydra is dev-kit quality that just happens to be sold to consumers. I expect the next iteration to be significantly more solid. It would be like consider the OR dev kit only on its existing merits without giving it lee-way for what it will be in a short period of time.

But given that this is a technology and an idea that is playing the long game (i.e. a 12 month fad until affordable home tech supersedes it can't really be considered a revival), we can't consider the technologies in this area without some degree of lee-way for what it might hold in stock for us in the future.

Despite the cost, I think the bigger cost component is still the PC itself, but I don't think that's a real detriment to home adoption.


Having said that, pairing up hydra (or whatever future iterations are called) with kinect does kinda bone the finger dexterity part of the immersion equation. But I think it's a reasonable trade off to provide the players with very necessary traditional gamepad sticks and buttons and have it pair up with motion tracking.

Still, ideally, all displays and inputs would be workable in an ideal VR platform/environment. As a designer, I can perceive solutions for most of these eventualities...

 

[Unicorn]

In feep we trust.

 

[majik13]

I need to try this

 

[whitehawk]

That looks so fun. Yeah this doesn't look feasible for home use, but I would love to see arcades open up with a few of these in it.

I've only been able to try one VR experience. I'm extremely jealous that you've got to try not only the Rift, but this thing as well. This year at the Ontario Science Centre there was a video game history exhibit. In that, they set up one of those old hamster balls with a gun and VR headset.

Basically this:

However it was awful. I think it was from the late 90s or possibly earlier. The headset had a god awful screen that was constantly flickering due to a poor connection I'm guessing. The in-game visuals looked like a early PC FPS game, again likely from the late 90s. The head tracking barely worked. The hamster ball didn't work well. It would also get stuck occasionally. The gun you held was also useless. And of course not only was there lots of black around the shitty screen, but I could also see outside of the actual goggles into the real world.

End the end, it was more fun to just take the headset off and just enjoy walking around in a human sized hamster ball.

 

[Das-J]

You pair that thing with the device in the OP and we're talking!

As a side note, I truly enjoy Feep's writing style.

 

[Feep]

You pair that thing with the device in the OP and we're talking!

As a side note, I truly enjoy Feep's writing style.

<3

I tried one of those hamster balls at E3 in 2009, it was hilarious

 

[JamieVRcade]

Interesting, I would love to dedicate a room to this, but not at that price. I have wondered before if you could do something similar with a few kinects for mapping out a room and then applying textures to obstacles such as couches, walls etc.

In our experience, Kinects have neither the range, or fidelity to do what we are doing. The new Kinects are much better, but range and absolute position are still problems.

When you are controlling an avatar on screen, there is a lot of room for error. When you are replacing the user's view of the world through an HMD, there is no room for error. Additionally, the use of props is very important and that isn't something that either Kinect does well in our experience.

The real problem is software development. If you are doing to be putting together your own system and mounting 4 Kinects on your walls, running cables, making your HMD wireless, integrating a prop and prop tracking, wiring in a locomotion system...who is going to make games for you? Who is going to support your exact configuration? You need a standard to develop against and that is what VRcade provides.

 

[JamieVRcade]

Damn this is exciting but I really dont see anyone making worthwhile software for it.

It takes a few minutes to port most games made in Unity over to the VRcade. Additionally, there are hundreds and hundreds of development houses which create content in Unity. You don't need a Call of Duty or GTA to guarantee an amazing experience. A wide range of fantastic content already exists. What is even more exciting is the content that currently DOESN'T exist. By having many deployments around the world and proving the business, we can create an alternative version of your free to play Facebook FPS, for example, for the VRcade with minimal work.

 

[JamieVRcade]

Cultural force nothing.

OR consumer version + Kinect 2.0 > VRcade.

Seriously, this tech is moving fast... and the limitations of walking in a room vs walking on a spot aren't nearly as big as you might think.

i.e. use kinect to track the leg motion - simulate walking on the spot as walking forwards (maybe in combination with the direction of the analog stick).

This is assuming that 1 kinect 2 can accurately track a human from all orientations (assuming that the body is obfuscating itself - but then you'd still want to pair the motion input from the kinect with a hydra style device, so using the local positioning of the hands holding the hydras and inverse kinematics to solve for the occluded body parts, it shouldn't be a problem at all).

I mean... the leg-hip swing that a large room or an omni directional provides is nice, but with walking on the spot, you still get the semi-accurate proprioceptive feedback (i.e. muscles contracting up and down, feet striking the ground) and natural headbob (i.e. the VR headset will literally bob up and down on the user's head as their feet move up and down) which when coupled with good visual feedback should prove to be an extremely immersive sensation.

After all, even with a large room or omni treadmill/board, there are elements of movement that aren't captured properly; such as moving up and down inclines and steps.


So using the kinect, the solution to providing the player with an adequately safe and immersive experience is to calibrate the player's position and create a centered location for them to home in on... as the player moves away from the spot, the software informs them of this (i.e. you are leaving your activity zone, please move back *uses on screen indicators to show them where they are relative to the 'home point'*).

Similarly, the kinect can show them obstacles should they start getting close to them (e.g. moves too close to a chair and might trip over; the chair ghosts into the view of the user - or even a person that moves in too close might be shown so that the user can avoid them if necessary).

You are introducing a lot of algorithms and processing power to just guess where the player is and what they are doing with their legs. This will no doubt affect latency, which is the absolute killer of VR. Additionally, simply bobbing up and down in place is no replacement for actually moving. Your eyes tell you that you are moving forward and you aren't...that's what VR sickness is all about. Our system removes that limitation by providing you with the space to actually move. All of the hocus pocus needed to fake it goes right out the window because you don't need to fake it anymore, you just move. This results in extremely low latency and the most natural experience, VR or otherwise, that the user has ever had.

The question of walking in one place shouldn't be about whether or not it's possible. It is. The Omni does this. It's about user experience. Do they get sick? Do they have the full range of human movement? Is there a learning curve to learning how to walk? Do they get tangled in cords? Will they trip on their cat? We have solved all of these with the VRcade. Yes, you have to leave home to experience it, but you have to leave home to experience go-karts and paintball and bowling as well. There is a reason for that; the limitations of the home, space, and consumer grade technology seriously hampers those experiences. Especially VR.

 

[JamieVRcade]

You pair that thing with the device in the OP and we're talking!

As a side note, I truly enjoy Feep's writing style.

The Virtusphere would obstruct the tracking of the cameras, so head and prop tracking would have to be handled by unreliable, drifting IMU signals. Also, the momentum of running keeps the ball moving after you are done, so when you want to stop, you don't quite stop. Getting all of your wireless signals through a massive metal ball would also likely introduce some problems.

Nothing, and I mean nothing, beats good old flat out, open floor running, jumping, and human movement across thousands of square feet. That's what we offer.

 

[coldcrush]

I have said this before, but I will say it again, unless of some travesty, then VR in whatever form is the future of entertainment.

For gaming, commercial purposes and also for movies, imagine being able to walk around a movie scene, imagine having to hide behind an object in a horror movie/game,,,, completely mind blowing.

The space issue for movement when playing any VR device is a big hurdle,. the omni move , hoop you can walk in still feels kind of odd, and having a huge room to wander around in is not practical,,,,, this to me seems the number one hurdle,
I am so excited for the day VR becomes affordable and developers actively support it,
I know a basic form of this is very close (within the next 2 years)

 

[mrklaw]

The Virtusphere would obstruct the tracking of the cameras, so head and prop tracking would have to be handled by unreliable, drifting IMU signals. Also, the momentum of running keeps the ball moving after you are done, so when you want to stop, you don't quite stop. Getting all of your wireless signals through a massive metal ball would also likely introduce some problems.

Nothing, and I mean nothing, beats good old flat out, open floor running, jumping, and human movement across thousands of square feet. That's what we offer.

How do you deal with multiple people not bumping into each other? Or are these single player game spaces?

 

[Feep]

How do you deal with multiple people not bumping into each other? Or are these single player game spaces?

Ostensibly, you would see their avatar in the game space just as you'd see any other object. Don't want to collide with another player? Don't collide with another player. = D

Also, hi Jamie!