Oculus "Rift" : An open-source HMD for Kickstarter

[2EyeGuy]

Making a driver that reliably gives a good VR experience in most games is extremely difficult, and perhaps impossible.

I should point out that the Oculus Rift technically doesn't provide a VR experience by itself, and none of the games with native Rift support will support Virtual Reality.

The definition of Virtual Reality, as most people understand the term, and as the term has been used throughout it's life, requires position tracking of the head, and at least one hand. It means you can move your head around and look at things from different perspectives, and move your hand around to interact with things.

The Oculus Rift is just a Head Mounted Display with orientation tracking (but with mind-bogglingly unbelievable and awesome specs), and that only gets you half-way there.

Heya Guys! I got intimidated by thew 170+ pages - I went to their site, and what I don't like is that they seem to rely on having games/engines "ready" for their games instead of using drivers like Tridef Ignition - this is bad.

Yay! Yuriythebest is back! I wondered why we hadn't seen you in this part of the forum. Not that the other people here aren't great, but it feels like there's something missing without you here.

Don't worry, there are actually several drivers for it already, including Vireio Perception by Cybereality which does actually try to give part of the VR experience in many games. There's also an unprecedented amount of developer interest, including big names and indies, and apparently most developers have ordered dev kits.

But like Palmer says, it's more difficult to just hack Rift support into existing games because the Oculus Rift has 3D, a very high FOV, strange aspect ratio, head tracking, needs things to be rendered life-size, and needs a high frame-rate.

Look at how popular all the game peripherals that require special coding are. What peripherals?

Guitar Hero controller, Wii Remote, Nunchuk, Balance Board, 3D TVs, Xbox 360 Controller, and Kinect are the main popular ones (unless you're old enough to count sound-cards, mice, and joysticks).

[2EyeGuy]

I hope the first release of the Rift addresses these issues because as it is right now, my DIY Rift doesn't even come close to making me feel like I'm 'in the game.' It's more like I have my face pushed up against my old iZ3D monitor, which, like my DIY Rift, I only used for a brief period then set it aside as an interesting but unusable/unpleasing curiosity.

Watching other peoples' reaction to the demo unit Oculus has been showing around makes me wonder exactly what it is they're seeing that I don't. Can my DIY unit be *that* much different than theirs?

Before I abandon HMD technology altogether, I'm going to try Dycus's driver and see if adjusting the image to match my tiny IPD (5.5cm) will help. I also have a real Rift on order, so there's hope there, too. But I'm not holding my breath on either.

Wait... Dycus has a driver?... That we can use?... Has Dycus been doing a Palmer and leaking things?

Does your DIY Rift have (very low latency) head tracking? Because I think that was supposed to be a large part of the feeling like you are in the game.

And isn't your DIY Rift focused at infinity? It shouldn't feel like your head is pressed up against a screen, it should feel like a huge screen several metres away.

[cybereality]

I don't think Dycus has a driver, at least not that I've heard of. He probably meant Emerson.

[Okta]

Making a driver that reliably gives a good VR experience in most games is extremely difficult, and perhaps impossible.

I should point out that the Oculus Rift technically doesn't provide a VR experience by itself, and none of the games with native Rift support will support Virtual Reality.

The definition of Virtual Reality, as most people understand the term, and as the term has been used throughout it's life, requires position tracking of the head, and at least one hand. It means you can move your head around and look at things from different perspectives, and move your hand around to interact with things.

The Oculus Rift is just a Head Mounted Display with orientation tracking (but with mind-bogglingly unbelievable and awesome specs), and that only gets you half-way there.

Look at how popular all the game peripherals that require special coding are. What peripherals?

Guitar Hero controller, Wii Remote, Nunchuk, Balance Board, 3D TVs, Xbox 360 Controller, and Kinect are the main popular ones (unless you're old enough to count sound-cards, mice, and joysticks).

We will dismiss sound-cards, mice, and joysticks because they are standard interfaces that can operate with all/most available content. What do all the game controllers you mention have in common? Console controllers. Not a good comparison. The Rift is a niche within a niche like a novint falcon or hydra, but i suspect even more so because those im guessing those have some sort of legacy control support. The RIft is in the niche of head mounted displays and in its own niche again because it cant display standard content. And as you say above its not even a VR device just a niched,niched display. One thing that also stands out in the VR regard is that the RIft is designed for chair/couch use only.

[Mel]

I hope the first release of the Rift addresses these issues because as it is right now, my DIY Rift doesn't even come close to making me feel like I'm 'in the game.' It's more like I have my face pushed up against my old iZ3D monitor, which, like my DIY Rift, I only used for a brief period then set it aside as an interesting but unusable/unpleasing curiosity.

Watching other peoples' reaction to the demo unit Oculus has been showing around makes me wonder exactly what it is they're seeing that I don't. Can my DIY unit be *that* much different than theirs?

Before I abandon HMD technology altogether, I'm going to try Dycus's driver and see if adjusting the image to match my tiny IPD (5.5cm) will help. I also have a real Rift on order, so there's hope there, too. But I'm not holding my breath on either.

Wait... Dycus has a driver?... That we can use?... Has Dycus been doing a Palmer and leaking things?

Does your DIY Rift have (very low latency) head tracking? Because I think that was supposed to be a large part of the feeling like you are in the game.

And isn't your DIY Rift focused at infinity? It shouldn't feel like your head is pressed up against a screen, it should feel like a huge screen several metres away.

Sorry, I meant Cyber.

For head tracking I use a TrackIR 5. Maybe it's not good enough for an HMD, but I found it didn't really add much to the experience. Also, I do have the lenses focused at infinity. I suppose that when I say it feels like I have my face pressed up against my iZ3D monitor, I actually mean I have my head very close to the monitor so I can still focus on it.

I'd be interested to hear from other DIYers if they really get the sensation that they're truly 'in the game.' If so, I'm quite envious.

[2EyeGuy]

I wasn't suggesting that the Rift was the same as those. Just answering your question in the hope that it would shed some light on what factors affect the success of peripherals.

You're wrong about sound-cards, joysticks, and possibly mice. TODAY they are standard interfaces that can operate with all content. But that wasn't the case originally. In the days of DOS, whenever you installed a game you had to tell it what sound card you had, and the games mostly included separate low-level code for many different sound cards, and often different music and sound effect files for different cards. There was a standard for 2-button 2-axis joysticks, but that wasn't any use for the Gravis Gamepad, or flightsticks with lots of buttons and axes, that had to be supported specially, and they were supported. The mouse is a bit different. It didn't exist until Microsoft decided to make it in 1982 and to add support for it to Microsoft Word (for DOS). Then other companies cloned it and copied Microsoft's INT 33H driver interface (but they added a third button). But games still had to specially add support for the mouse, and early games certainly didn't support it. When John Carmack invented mouse-look in Wolfenstein 3D (or was it Catacomb 3D, or Hovertank?) he was being very creative.

I didn't mention graphics cards, but they also apply. Games had to specially support Hercules, CGA, EGA, VGA, and many different kinds of Super VGA, all at the lowest level. Strangely Super-VGA never really caught on in DOS games, although they still sold well, but the other ones all did.

The other ones I mentioned, were mostly console controllers. Although when I said the Xbox 360 controller, I actually meant on the PC, not the Xbox 360, and on the PC they are implemented specifically, not using the generic joystick code. The exception is 3D TVs, which have a non-gaming function as their main selling point.

So the pattern from all of that is... the game peripherals that are successful (in the non-niche sense) are:

* those that are made and promoted by the console maker (IBM or Microsoft in the case of PCs), and third party clones of them.
* those that add something that was seriously missing, (usually features people had seen on other systems) and were supported by John Carmack and/or other prominent game developers, eg. sound cards, Gravis Gamepad, some other special joysticks.
* those that originally had a non-gaming purpose, eg. mice, 3D-TVs, home networks, modems, (and to an extent, sound cards); but became gaming peripherals after being supported by some prominent game developers.
* And it also helped if they were from a famous company in gaming already (eg. Gravis gamepads).

Currently the Oculus Rift needs to fit into the second category, or start their own success story that hasn't been done before, or just find a niche. But after analysing all this, I'm starting to think it might be a mistake for Oculus not to develop their own console (think "Nintendo On", rather than "Virtual Boy").

[Okta]

I wasn't suggesting that the Rift was the same as those. Just answering your question in the hope that it would shed some light on what factors affect the success of peripherals.

You're wrong about sound-cards, joysticks, and possibly mice. TODAY they are standard interfaces that can operate with all content. But that wasn't the case originally. In the days of DOS, whenever you installed a game you had to tell it what sound card you had, and the games mostly included separate low-level code for many different sound cards, and often different music and sound effect files for different cards. There was a standard for 2-button 2-axis joysticks, but that wasn't any use for the Gravis Gamepad, or flightsticks with lots of buttons and axes, that had to be supported specially, and they were supported. The mouse is a bit different. It didn't exist until Microsoft decided to make it in 1982 and to add support for it to Microsoft Word (for DOS). Then other companies cloned it and copied Microsoft's INT 33H driver interface (but they added a third button). But games still had to specially add support for the mouse, and early games certainly didn't support it. When John Carmack invented mouse-look in Wolfenstein 3D (or was it Catacomb 3D, or Hovertank?) he was being very creative.



Currently the Oculus Rift needs to fit into the second category, or start their own success story that hasn't been done before, or just find a niche. But after analysing all this, I'm starting to think it might be a mistake for Oculus not to develop their own console (think "Nintendo On", rather than "Virtual Boy").

I do well remember dos drivers, irq,dma pains but it is still apples and oranges. Win95 and dos supported mice, soundcards and pretty much joysticks universally once you simply installed and configured the supplied driver. The Rift is different, it wont be supplied with a one size fits all driver.

A Rift style console could certainly work but i dont like consoles Perhaps a high end Android device similar to the Oyua with VR controllers. Remember the Android helmet with built in hand recognition Palmer tried out, i wonder what has happened to that.

[Owen]

The Rift is really more like a new platform than a peripheral.

Conceptually its just as different from regular PC gaming as PC gaming is from mobile. And just like pc versus mobile most straight ports are going to be bad. Some genres should translate well, most will not without major changes.

[2EyeGuy]

Dos supported mice, soundcards and pretty much joysticks universally once you simply installed and configured the supplied driver.

That's not at all true (except for mice).

[2EyeGuy]

I was just dreaming about a review of the Oculus Rift:

...using the Vireio Perception 3D drivers. So, I have to say, Left 4 Dead looks amazing on the Oculus Rift. And unlike other reviews that just say "Oh, you know, this super high FOV, 3D, and Headtracking is so nice, but we can't show it to you because obviously we are not broadcasting in 360-surround 3D!" Unlike those who are, like, I don't want to say any.... Ah, I have an idea! Unlike those безглуздя! I can show it to you in... oops... no I can't... So here you... can't see Left for Dead as it looks on the Oculus Rift. Sorry.

[youtube]http://www.youtube.com/watch?v=0tASa5odS7E[/youtube]

[Okta]

Dos supported mice, soundcards and pretty much joysticks universally once you simply installed and configured the supplied driver.

That's not at all true (except for mice).

Things pretty much standardised to sb16 and midi for music iirc?

[logicalChimp]

eventually, but not when sound cards were introduced.

As you say tho, eventually the soundblaster API came out on top, so all the other cards ended up including a soundblaster-emulation mode / driver.

In relation to the Rift, if it proves sufficiently popular (and many companies will be watching closely just from the effect of the Kickstarter) then competing HMDs will probably hit the market, with their own APIs etc. One of them will 'win', and the others end up emulating their API, leading us into the 'SB16' era - which is when (imo) games really started to make the most of potential offered by sound cards.

[sth]

At some point, there has to be a standardized interface for developers to use and for hardware vendors to write drivers against.
If VR gets popular in the consumer space, this has to happen eventually – it's one of the important steps for VR becoming more than a niche product.

Look at it this way: 3DFX kickstarted the consumer 3D graphics card market back in the mid-90s. They used their own proprietary API (Glide), which was bound to become obsolete at the point where other big players entered the market and developers became unwilling to rewrite their renderers just to support a single vendor. Eventually, the industry settled on OpenGL and DirectX. But if it hadn't been for 3DFX, the whole 3D gaming revolution would probably have taken a few years longer to come to fruition.


I image that we'll end up having a standardized API, a sort of "OpenGL for VR stuff" ("OpenVR"? ), in order to standardize HMD and tracker access. And even though Oculus is pretty much alone in the gaming-oriented consumer HMD space right now, they should embrace such a development or even initiate it themselves, so they don't suffer the same fate as 3DFX's Glide API. Without a standard, VR adoption by game developers will probably remain very limited.

However, we have sort of a chicken and egg problem right now. In order to determine what such an API should look like and to build it in a universal, futureproof way, we need VR headsets in developers' hands. There are so many open questions – not just about the actual HMDs, but also about how head- and body tracking will eventually be solved*. There are a lot of new developments in that space right now and we'll probably see a lot more once the Rift is out.


Anyway, I really hope that Oculus will eventually be the initiator of such an open VR API and not just do their own thing (although in the beginning, they may have to).
But no matter what: Without an actual HMD in developers hands, the whole thing is dead anyway.



* Ideally, such an API should of course provide an abstract interface to tracking data, regardless of the actual hardware implementation. But given the vast range of tracking concepts, determining adequate, futureproof representations of the tracking data is not an easy task.

[morgg]

For Windows, it wouldn't surprise me if Microsoft extended its DirectX API to support VR in the near future, maybe integrating Kinect SDK support.

[sth]

For Windows, it wouldn't surprise me if Microsoft extended its DirectX API to support VR in the near future, maybe integrating Kinect SDK support.

I wouldn't be surprised either (although I think it may take at least a year or two [edit:] after VR finally takes off in the gaming space – which it hasn't, so far).
But in my opinion, a cross-platform solution is needed, no matter what Microsoft (or any other individual platform vendor) does.

[cybereality]

Well Microsoft *just* added stereo support in DirectX 11.1, and how long has OpenGL had QuadBuffers? Don't expect MS to pick up on VR so soon.

[jis]

Don't forget that Microsof's strategy is all about proprietary and incompatible solutions. They do everything possible to create technologies that won't be supported in other operating systems. If there is something to be done in VR, they will create their own VR, driver and game compatibilities.

[2EyeGuy]

Well Microsoft *just* added stereo support in DirectX 11.1, and how long has OpenGL had QuadBuffers? Don't expect MS to pick up on VR so soon.

Actually they added stereo support in DirectX 7 (and removed it again in Direct3D 8).

http://msdn.microsoft.com/en-us/library ... 98796.aspx

[coresnake]

What you guys have forgotten is that Windows is derailing towards a tablet future. Anything added to Direct X will primarily be things that are supported hardware-wise by the Xbox, not high end PCs. Hence their relatively slow uptake of new features.

[sth]

What you guys have forgotten is that Windows is derailing towards a tablet future. Anything added to Direct X will primarily be things that are supported hardware-wise by the Xbox, not high end PCs. Hence their relatively slow uptake of new features.

By the time gaming HMDs hit the consumer space, the next XBox will probably already be out, and I don't see any technical reason why they couldn't add HMD support to it.

Anyway... I'd like to see a vendor-independent, cross-platform solution for VR stuff and I don't really care what Microsoft does on their own. I'm just saying that I could imagine them jumping on the bandwagon at some point.

[cybereality]

I doubt Microsoft (or any one company at this point) understands what is needed for a real VR standard. I think this will come over time, after Oculus does its thing and other players come into the space. We need developers to really push the boundaries of what you can do with VR, and then break down what is needed to make that standard. It doesn't work if someone hasn't set the bar yet.

[Owen]

Agreed, and that will be true for any company. They won't move until there is a market, or their partners (such as graphics chipset companies) start asking for it.

[rmcclelland]

This is the best article on the Oculus Rift I have seen in a while.

http://www.gamesindustry.biz/articles/2 ... revolution

[cybereality]

@rmcclelland: Thanks for the link, great article.

I am surprised by what Palmer said about the price of the consumer version. Under $300 sounds optimistic, but would be awesome if that happened.

[Owen]

He was also saying earlier that the consumer version would probably have hand and torso tracking, I'm guessing thats no longer on the table then.

[2EyeGuy]

The article is light on specs, but otherwise quite good.

I think $200 is a bit too cheap. I'd rather have more features. And I know the Chinese workers are taking outrageously long holidays right in the middle of our production schedule, and forcing us to effectively cancel our own Christmas, but they should still be paid something. We don't need to haggle them down to slave labour.

[Okta]

He had tired of reading articles on his favourite tech sites that seemed to be written by “iPhone-toting hipsters” who had an obvious paucity of relevant knowledge.

Love it

[Zaptruder]

This is the best article on the Oculus Rift I have seen in a while.

http://www.gamesindustry.biz/articles/2 ... revolution

Indeed. Facts are right (as far as I can tell) and it alludes very clearly to the scale of Luckey's ambition and plans at least in the mid term.

I think in the long term, many on this forum are aligned with Luckey in our ideas of what Virtual Reality should be... complete sensory immersion.

Using a heads up display as the interim step, it'd be logical to incorporate as much of the body into that reality as possible, using whatever practical means are available.

[quietboy]

Palmer said " .....I mean, our developer kits are $300, and I actually think our consumer version is going to be cheaper than that - almost certainly. But even at $300 it's really quite a bargain.”

[MSat]

Palmer said " .....I mean, our developer kits are $300, and I actually think our consumer version is going to be cheaper than that - almost certainly. But even at $300 it's really quite a bargain.”

I totally missed that part of the article. Thanks for pointing it out. I think he's spot on about the pricing, and I'm glad he's not looking to make another $1k device. $200 would be fantastic, but I think paying $300 for an HMD with something like the Leap Motion integrated into it would be an even better deal. Unless some VR API comes to fruition soon, relying on software developers to individually support a wide range of VR peripherals from various manufacturers would probably be a big detriment to users and developers alike.

[coresnake]

Definitely agree on adding features over bringing the cost down. A helmet with built in motion tracking would be a revolution for the games industry.

[PalmerTech]

Important features will not be cut in favor of achieving a low price. Doing more with less is the name of the game!

[TheHolyChicken]

I'm really glad you hold that opinion, Palmer. I think a really amazing device has more potential than an "ok" device that's a bit cheaper, provided that it's not priced out of the market.

[Parity]

While it is nice to read that the consumer version is aiming to be reasonably priced, I still have my hopes up for a as-good-as-possible variant of the rift concerning resolution. I wouldn't want to see a lower res display, just because it will enable Oculus to keep the targeted price-range.

This will inevitably lead to a new section in the forums: "Hack your Rift with custom panel XYZ"

[MSat]

Important features will not be cut in favor of achieving a low price. Doing more with less is the name of the game!

I don't think development of a consumer HMD could be in better hands.

U DA MAN!

[sth]

While it is nice to read that the consumer version is aiming to be reasonably priced, I still have my hopes up for a as-good-as-possible variant of the rift concerning resolution. I wouldn't want to see a lower res display, just because it will enable Oculus to keep the targeted price-range.

Realistically, I don't expect to see anything other than a FullHD panel (either 1080p or 1920x1200, depending on what panels can be sourced) for the first consumer version of the Rift.

Nothing less, since 5"-7" FullHD panels seem to pop up everywhere these days, so hopefully, they will be widely available by mid- to end 2013.
Nothing more, since supporting anything beyond FullHD will exclude a lot of potential customers from the Rift (no console support, much higher GPU requirements, etc.)

Of course, FullHD is still far from what's actually needed, but I'm not worried. We're just at the beginning and I'm pretty sure that HMDs will improve greatly over the next few years as display technology is moving forward.

[tahustvedt]

In my opinion higher than 1080p isn't a bad thing, and doesn't exclude consoles. The controller will probably accept lower resolutions and scale them to fit the panel, like just about all current LCD displays do.

[sth]

In my opinion higher than 1080p isn't a bad thing, and doesn't exclude consoles. The controller will probably accept lower resolutions and scale them to fit the panel, like just about all current LCD displays do.

I'm not saying it's impossible (in fact I'd love to see something like Toshiba's 2560x1600 panel in the consumer rift), but it makes things more complicated.
The scaler would need deliver really good quality but of course without introducing any additional latency. Does anyone know if this is still a problem? (I know that it used to be, but I'm not really up-to-date on these things)
Not to forget that scaling always looks particularly bad on low resolutions (which is effectively what you have in an HMD).

Furthermore, depending on the game, maintaining a constant 60fps at 2560x1600 is very demanding even for high-end systems, and that's not even counting the additional overhead for a stereo rendering path.

That's why my "realistic" expectation for the first consumer version of the Rift is FullHD.
Of course I'd like to see more than that.

[illizit]

Wow! Just stumbled across this website/post, I've been looking for an HMD and this sounds incredible! Can't believe I missed the kickstarter!

I browsed through a few hundred pages lol... will this headset be able to work with most games without the need of being specifically programmed for the headset? Sorry if this is a noob question!

[mahler]

Wow! Just stumbled across this website/post, I've been looking for an HMD and this sounds incredible! Can't believe I missed the kickstarter!

I browsed through a few hundred pages lol... will this headset be able to work with most games without the need of being specifically programmed for the headset? Sorry if this is a noob question!

Unfortunately they will have to be programmed for it specifically. However, there are some projects going on to make current games work through a 3d-driver. These will not fix the issues that games needs to be truly immersive, but it's a start.