@Waggler, that was a really good video and interview. I like how in depth you went when describing everything that you saw in the game since we can't see it ourselves.

Dycus or Palmer,

Maybe you could give us some clues whether the Rift will still be "open-source" as the topic suggests. Will the driver/glsl wrapper shaders be open source and or the design itself?

I also wonder whether you will make the rift compatible with linux. The hillcrest tracker driver is closed source and not supported by linux (afaik).

Kind regards,

Jotschi

Ah! Thanks! That will be fixed in the final vid

@IGameArt - Thanks!

Personally I feel that the head tracker should move the gun and the view (so the same way it's done into Domm3 + the rift) and the pad should allow to move the gun into the view.

@Waggler - Thanks for the video, it was really great to see, and I had no problems with your English. You mentioned that you would prefer to aim with motion controller, but you also mentioned that holding your hand up didn't break the immersion. Do you think that seeing your hand being tracked would add a lot to the experience?

In terms of the problem of drift of the gyro affecting proper positioning of hand tracking (like the hydra razer), I was thinking that the best solution would be to have the reference (base station) actually on the HMD. It doesn't really matter if the view drifts as you will be doing gross adjustments to your direction as you play anyway. The main thing is that your hand position relative to your view must be correct. I wonder if this has been investigated in any depth.

@Waggler: Cool video, thanks.

I would prefer to aim with the hand simply because it would feel more natural. As it is now, it feels like you are aiming with a gun attached to your head basically. Which feels very real, but is hardly natural. Also, aiming with the hand would leave you free to look around in the process.

To answer your question, I guess it depends on how accurately the virtual hand is rendered over my real one. If you can make it so that a virtual hand matches the position of your real hand, it would totally feel like you are wearing a computer generated glove, if you get what I mean, which then again would add to the experience of being there of course. It's another piece of your body being accurately represented in the virtual space after all.

That said, with regards to FPSs, what I'd like to see in the future is motion controllers like the Hydra or the PlayStation Move being implemented as they were the stock of the virtual gun, rather than its barrel. Current implementations have you holding them like torchlights or laser pointers basically, which is fine, but I'd like to try out how it feels like to hold them as they were the actual stock of the virtual gun (ie: to aim straight, you'd need to hold them tilted upwards to different degrees depending on the weapon model).

On a side note, the way I envision the controls scheme of a VR FPS is basically like this:
- A 6DOF motion controller dedicated exclusively to gun control (its movement not affecting camera orientation whatsoever)
- A left stick for strafing / moving forward/backwards
- A right stick (located on the motion controller) for spinning around horizontally (dragging the weapon)
- Headtracking for looking around (without dragging the weapon).

Now, this solution poses a rather fundamental problem of course. Say you are looking 90° to your right via headtracking, and you start spinning anti-clockwise via the right stick, how should the virtual head behave? Should it stay still, like a tank turret, as the body spins below it resulting in anatomically unrealistic neck properties as well as in the player losing awareness of his virtual body orientation, or should it be dragged by the spinning body so that it starts spinning anti-clockwise even tho you are not turning your real head (possibly causing unpleasant vestibular-visual conflict, hence motion sickness?). I honestly don't know, but I'd love to see this kind of stuff investigated and prototyped by those who can.

edit: Thank you Cybereality!

Dycus when you move the stick up and down do you actually shoot in that higher and lower position or does the gun model only move up and down the field of view.

And when you move the gun down and out of the field of view is its just translating down or is it actually dropping/turn down so it is then pointing to the ground and can you still shoot but be shooting the ground or wherever the gun is facing or will it not let you fire till you bring the gun up?

waggler, I think a better idea would be like you said but instead of the right stick use another tracker on the shouders or lower back.

Your video was probably the best I have seen. Even though your English wasn't perfect you managed to give a lot more information than I have seen in other vids. It also helped that you seem to have previous knowledge of the technology and pointed out some of the flaws in the system, rather than just saying how awesome it is.

Am I the only person that has noticed that people using this thing look and sound that they are in a chemically altered state? I supose it is kind of the same, the mind in a different place from the body.

Edited for poor English. Doh!

I think that it should work as if you're on a platform that is controlled by the right stick. So if you have turned 90 degrees to the right and then turn left with the stick, then it would be the same as if you were facing forward and turning left with the stick ie. turn right 90 degrees with your head and 90 degrees left with your stick and you would end up looking in the same direction you started. Also, if you turn your head 90 degrees to the right and strafe left, then to your view, you should be moving backwards, since your platform is moving left, not your view.

Very nice video I like the attention to some of the stuff/questions that keep coming up all the time. I really got to say I hope Palmer and co. get some sleep soon lol. They all look very tired.

No idea, to be honest. And I'm not sure how much I'd be allowed to say if I did. You can wait till Palmer comes around, though.


It shoots where the gun is pointing, so yes and yes. You can always shoot, and it shoots where the gun's laser sight is pointing.

All these problems are solved if you are free to stand and turn around 360 degrees wirelessly. Then the headtracking can map 1:1 with the game's view angle, and then the gun can use a motion-controller with independent 1:1 mapping. You can almost achieve this now with current HMDs/trackers and something like the Wiimote, but the problem is no games support this setup. But if you are working on your own custom applications, or modding an open-source project, its totally possible.

I totally agree with this. This tracker would only need to measure yaw for torso direction, and maybe vertical positioning for crouching and jumping. You could easily have a tracker that clipped onto a belt. If the hand tracking was going to be optical, this belt tracker could also house the cameras for the hand tracking too.

That's a good idea, in fact I've been thinking about it for a few days.

The problem is, if the HMD drafts, the "gun" controller will draft along with it. Correct me if I'm wrong, but I think that until someone invents a true, standalone, draft-less 6DOF tracking device, at least one stationary base station is needed.

With that in mind, I came up with this simple setup (I know it's nothing new or fancy):
By using gyroscopes and accelerometers on the HMD and the "gun" controller for orientation and the method in the embedded video for position,you can get a 6DOF tracking setup for less than 100 dollars.


Here's a dumb little diagram i drew in MS paint to illustrate it.


Camera 1 and 2 would track the controller (just to correct draft) and 3 and 4 would track the player the same way.

You can counteract the gyro drift with a magnetometer. 3 axis magnetometers are very cheap and pretty simple too. Something like this does it without the gyro's http://www.oceanserver-store.com/osevkitsorus.html. The old InterTrax2 had 3 x gyros; 3 x accelerometers and 3 x magnetometers and was an excellent 3DOF tracker. Not around any more but it should be pretty easy for someone like Hillcrest to come up with something as good for a fraction of the price these days.

Part of it might be related to the fact that one naturally tends to shout a little bit when he's wearing headphones.

Given the Rift's current sensor package, I think the best approach for an FPS would be to loosely connect the gun to the view like in Red Orchestra. That way turning your head more than around 20 degrees will drag the gun along and vice versa. It would allow you to make small motions with your head or gun independently, without needing an absolute forward look reference.

Yeah. But since forward/backwards movement would occur relative to your lower body orientation (rather than the head's one like in Doom III) you'd need to provide the user a visual cue or something about it so that he always aware in which direction he's gonna move the moment he starts using that left stick. Like when driving a tank in BF3 where you can see a top-down representation of the tank + turret at the bottom of the screen.



They were indeed, lol. They have been meeting people non-stop, and traveling around the world to show off the Rift to devs and whatnot.
I should have known better and scheduled the meeting in the morning rather than at 6pm



Of course if you play standing up there would be no problem, but I tend to ignore that option because I believe it's not gonna happen. That's for two reasons, mainly:

1 - Most gamers don't really want to be standing up when playing their games. It's ok for brief gameplay sessions, but I can't see myself playing 3 hours of Skyrim while standing.
2 - Safety issues. It's very easy to loose track of where you are standing in a room even if try to stick to one spot.



Spinning by rotating the torso right? Good idea. That might actually work. I guess a depth camera can easily take care of that. Then again I'm not sure mapping too many controls to body movements is a good idea. Gamers generally want to move as little as possible

Maybe, the easiest solution is to have the direction of movement based on the direction of your hand. You have two hands... the left can be the movement direction, and the right, your gun direction. Or I think my preference is that the movement is relative to the gun pointing.

We recently came up with a demo for this: http://www.lp-research.com/lpms-head-tracking-demo/
The 3D graphics is a bit old-schoolish (+ a few bugs), though. But the main thing is that latency is low and here is no drift (thanks to the magnetometer). Anyway we combined the system with an optical tracker as well: http://www.lp-research.com/simultaneous ... -tracking/

Could make it really cheap, too. So it might be interesting for low cost VR.

Speak for yourself, man what I'd give for the ability to play FPS while doing my cardio...

I just found all your Move videos by following the links you posted earlier. There was a lot of interesting stuff and thoughtful analysis in those videos. Have you made any more than the SOCOM4 video where you look at the Sharpshooter?

The reason I ask is because I got myself a Sharpshooter with the intent of using with a PC. (As I've seen other people here on the board have done.) One thing I've concluded so far is that you don't want to use a weapon to control body aim. Having to aim towards the edges to turn becomes a shore IMHO and makes it very difficult to do moves like circle strafing.

However, I think that what the Sharpshooter does best (at least in Killzone 3, which is the only compatible game I have) is when you bring up the iron-sights. This makes it possible to aim and fire pretty much as in a light gun game. I imagine that mode combined with head and gun tracking would be pretty amazing.

To control the players body I think the simplest solution could be to add a second thumbstick to something like a Sharpshooter. So you have one to control your movement and one to control your "body-aim". I'm also thinking that if you're standing up it could be neat to have completely separate head, gun and body aim. But if you're sitting down you'd probably want to have head and gun aim by relative to your body. (In that case "straight ahead" of your physical body would be the direction of your virtual body, and you'd turn using a control stick, similar to the Doom3 method.)

The demo is really impressive. Great job.
That last part is quite important. The FSRK-USB-2 can be bought for $99,- at stock price.
The LPMS-CU packages on the LP Research website start at $689,- (incl. shipping)
I'm sure that developers would be interested in combining the Oculus Rift with your sensor, but it doesn't seem to be feasible until it can get to a similar price-range.

Would you be able to tell any (current) drawbacks to using it compared to the FSRK-USB-2 which is shipped with the Rift?

Thanks!


Basically, if we could make the board a little bigger it would be no problem to get the price down to say 80US$. The components are not so expensive. It is the 6-layer PCB + 3D printed case that drives the price up. But if the board could be a little bigger we could use only two layers and therefore become much cheaper. If we knew that there is a certain number of people who are interested in this tyoe of low-cost board we could build it in no time (means, a few weeks).


I haven't used it, but I think the FSRK-USB-2 has a major drawback compared to our thing. It doesn't contain a magnetometer for yaw drift compensation. That means that the yaw angle measurement will drift over time. For example I would guess that if you did some fast turning movements (like we do in the second half of our demo video) with the FSRK-USB-2, the scenery would eventually not turn back to its original position. This is because the error from the gyroscope integration gradually becomes bigger which results in an offset.

I would think that having some additional heading reference (either magnetometer or optical) is unavoidable for head tracking.

No, I didn't publish any other SS related videos other than the SOCOM 4 one, the reason being I hate that thing for exactly the same reasons you bring up. It's a great attachment, mind you, but operating it to control the camera is hardly convenient. It's much easier to control the camera with the wrist (that is, with just the vanilla Move) rather than with your upper body, because with the latter you have to deal with added inertia.

I think the reason why it feels better when bringing up the iron-sight is because you generally don't move the camera at all in that mode. Like you said, it works pretty much like in a lightgun game. I do still prefer using the wrist because it's faster, but yeah I agree that mode + headtracking + gun tracking would be amazing.

Yeah, I found that when I was only walking around the Sharpshooter was very annoying and it even made my hand hurt enough that I could only play for about an hour. (The reason for this is that you have to twist the entire gun left and right to turn and this feels very unnatural.)

Ideally I think you'd want to be able to bring the weapon down across your chest when moving around as that would make it possible to relax your arms and hands. (Much like you would with a real weapon.)

Yes, but you can't do that cos otherwise the camera would pitch downwards. In a VR environment, with the gun detached from view, you could totally do that

Would this be something Palmer and Oculus crew could implement into the Rift?

I already do: I play GTA IV on my treadmill. When a new mission starts, instead of 'jacking a car, I hop on my treadmill, controller in hand (with an elastic band wrapped around it such that it depresses the 'X' button (run)), and I jog myself to the mission destination. Helluva workout.

It also helps that I have the whole thing set up in front of my 150" projector system.

HD version of my Oculus Rift video is live. Hopefully there are no more typos!

While I agree that a drift compensated sensor is better than one that is not for the general case, the truth is that for most of the content that the Rift will be used for - namely first-person-shooters - yaw drift is a non-issue. For unidirectional/seated play the gamepad is used to swing the character around so a fixed reference frame is meaningless. For standing-360 play, a 1:1 turning angle is not perceptible to the player so drift also doesn't matter. It only matters in the case of cockpit simulators (where the forward direction in the virtual world has to align with the forward direction in the real world), augmented reality (where the virtual and real world overlay), and some free-movement scenarios.

Thanks Waggler, really great video

Really like the video, Waggler. Much more information than most other reviews. I take it you enjoyed the demo?

@Dycus, maybe you could shed a little light on the subject, but i'm thinking i'll have to wait til my devkit arrives to get my answer, but doom 3 bfg comes with everything DOOM that id has created, from doom 1+2, to doom 3+ ressurection of evil. Now i know that D3 +ressurection will support the rift, but what about d1+2? I'd love to be immersed in those 2.5d environments that i've been modding for the past few years

Once comprehensive motion detection/control systems become available (and I hope the Rift is the impetus for this), there should be all kinds of possibilities in the sports gaming realm. Imagine playing football with the Rift, putting you right in the middle of the action! Too awesome!

I'm so excited for the future of VR gaming. Thanks Palmer, et el, for forging the path ahead.

Oh absolutely. And that was in spite of the blurriness (due to my nearsightedness) and the lack of head position tracking. I can almost taste the awesomeness when that stuff gets fixed/implemented.

Then, it will be interesting to see how control interfaces evolve and developers adapt to this new HID.

I hope the industry as a whole embraces this as soon as possible because this is how you achieve true realism: not by increasing polygon count and texture resolution, but rather by tricking the senses so that brain believes what you are seeing is real, no matter the quality of the actual graphic content. Heck, headtracking and FOV were enough for me to believe I was there even tho it was all blurry. It felt like being in Doom III for real without my prescription glasses on basically

Come to think of it, this could actually be beneficial for the industry in terms of development budgets since you don't need to invest that much into graphics to make something believable. But I digress.

Agreed. Sports is a really untapped area when it comes to immersive gaming. Particularly American football (quarterback position only), golf, bowling, and boxing would work amazingly well in a relatively confined space with 1:1 motion sensor technology and immersive viewing.

If you don't mind me asking... How nearsighted are you? I'm pretty extreme myself (about -10, -10) so I'm curious how much work would have to be done to make it useable for me. (Particularly if I'm in a situation where I'm demoing to other people.)

I'm about -2 per eye, with slight astigmatism. I'm afraid you wont be able to see much with the current set up.