Khronos To Develop New Standard For VR

[phoronix]

Phoronix: Khronos To Develop New Standard For VR

The Khronos Group is going public this morning with a call for participation of companies that are not yet Khronos members but looking to join a new initiative: developing a new, cross-vendor VR standard to allow for better virtual reality interoperability of hardware devices, game engines, and more...

http://www.phoronix.com/scan.php?page=news_item&px=Khornos-Embargo-VR

[Mike Frett]

So does this take developers etc. away from projects like Vulkan? Or what's the deal here?

[lunarcloud]

This sounds exactly like OSVR, except with buy-in from vendors other than Razor. Cool.

[M@yeulC]

Good initiative, but I still expect it to take quite some time, if Vulkan taught us anything.

This effectively reminded me of Valve's OpenVR. There's hope that the resulting API will be even more "open" (I mean, actual bindings to the various SDKs out there, not just header files).

Either way, this could be great for VR on Linux. We could even imagine an "easy" PSVR port using this API as a front-end for VR application.
The part about controllers is also interesting. I hope they can decouple it from this big "VR API". We need one similar to the Steam controller's API: deal with actions, not controls/bindings; this cannot be stressed enough. And provide hints dynamically, as provided by the system. Then, Valve can make their own layer if they want, which would be great.

Nevertheless, it's great to see big players agree on something on this market. Would Microsoft (Hololens & Co) join the fun, or is it not specified? I know it's AR, but there's definitely some overlap here.

[SaucyJack]

Don't get me wrong I'm glad Khronos is working on this, but holy crap do they take forever to do anything.

[schmidtbag]

Valve/HTC, Oculus, and OSVR/Razer are all doing VR because they want a piece of the market. They created their own standards where each of them were hoping they would be universal. Oculus was sort of taking the Nvidia approach and tried being more proprietary, while Valve and OSVR are both trying to create open standards. So what doesn't make sense to me is why all of these companies are joining together for something that could effectively obsolete their APIs. What was the point of them making headsets if they will eventually all be functionally the same, including competitors like PSVR?

As an OSVR user, I personally am really liking the way OpenVR is turning out. It seems to be very well supported and it works great even on unsupported headsets. I still don't fully understand if Khronos intends to adopt OpenVR, or just use pieces of it in the same way they did with Mantle for Vulkan.

[TheOnlyJoey]

So does this take developers etc. away from projects like Vulkan? Or what's the deal here?

This project will probably be part of new OpenGL and Vulkan specification so they will probably hire additional programmers for this project, or have external companies hire them.

This sounds exactly like OSVR, except with buy-in from vendors other than Razor. Cool.

Well not quite, OSVR just handles abstracting existing drivers and does a bit of frameworking for DirectX and OpenGL for certain features.
The Khronos framework will handle everything rendering related ("direct mode" in base specification, timewarp systems, variable hardware synced refreshrates etc), and might introduce a basic interaction to hook up the vendor drivers, but it will not replace the vendor drivers.

This effectively reminded me of Valve's OpenVR. There's hope that the resulting API will be even more "open" (I mean, actual bindings to the various SDKs out there, not just header files).

Either way, this could be great for VR on Linux. We could even imagine an "easy" PSVR port using this API as a front-end for VR application.

The problem with OpenVR is that they called the framework open, but still providing blob drivers, not making it any different from other drivers out there, most FOSS developers refer to it as ValveVR because of that.

The only current framework supporting a wide range of HMD's with full open source drivers is OpenHMD, which also can be hooked up to OSVR to use in their supported library of games.
Also since you mentioned PSVR, it is already supported by OpenHMD (currently in branch though).

[ssokolow]

Valve/HTC, Oculus, and OSVR/Razer are all doing VR because they want a piece of the market. They created their own standards where each of them were hoping they would be universal. Oculus was sort of taking the Nvidia approach and tried being more proprietary, while Valve and OSVR are both trying to create open standards. So what doesn't make sense to me is why all of these companies are joining together for something that could effectively obsolete their APIs. What was the point of them making headsets if they will eventually all be functionally the same, including competitors like PSVR?

Being proprietary is good if you can gobble up enough of the market fast enough to gain a network advantage. If your competitors match you and balance out the split, you wind up with a situation like Japan in the 80s, where there are enough mutually-incompatible choices that it cripples customer interest until a Microsoft Japan comes along and introduces the MSX standard to provide a common baseline that software can be written against.

[schmidtbag]

Being proprietary is good if you can gobble up enough of the market fast enough to gain a network advantage. If your competitors match you and balance out the split, you wind up with a situation like Japan in the 80s, where there are enough mutually-incompatible choices that it cripples customer interest until a Microsoft Japan comes along and introduces the MSX standard to provide a common baseline that software can be written against.

True, but unfortunately for Oculus, they lost a lot of their potential (and existing) customers due to pricing, software compatibility, broken Kickstarter promises, and the whole Facebook thing. To my understanding, Oculus no longer has any significant advantage over Vive, and OSVR+Leap Motion are slowly catching up (but still a worse overall experience).

[LinAGKar]

Being proprietary is good if you can gobble up enough of the market fast enough to gain a network advantage. If your competitors match you and balance out the split, you wind up with a situation like Japan in the 80s, where there are enough mutually-incompatible choices that it cripples customer interest until a Microsoft Japan comes along and introduces the MSX standard to provide a common baseline that software can be written against.

You're linking to an <a>-tag.