[CLOSED] The right way to enter VR mode about webxr HOT 4 CLOSED

Comment by toji
Monday Mar 30, 2015 at 20:33 GMT

I like to think of the current syntax as implying "I want you to display this content fullscreen on device X". The "VR" part of it may be overtly specific, but suits the immediate need.

Thus far, we haven't seen any devices where it's meaningful to render anything other than a fullscreen view for the purposes of VR. Beyond that, however, it serves as a useful base for progressive enahancement that does almost the right thing in several situations. For example: On Android even if you have no support for HMD distortion or anything like that you can render a split screen canvas fullscreen and it looks acceptable in a Cardboard harness. If we then add distortion support the same command will now get automatically "upgraded" to a more correct version while still being backwards compatible with other/older browsers. Same goes for something like the Oculus Rift. Even if "VR fullscreen" isn't natively supported you could at least mirror displays and use fullscreen to see an image that swims around the edges. (Better than nothing!) That's the kind of evolutionary track that the web thrives on.

from webxr.

Comment by borismus
Friday Apr 03, 2015 at 01:02 GMT

OK, any reason why the vrDevice argument to requestFullscreen isn't
captured in the spec? I'd like to polyfill this behavior too.

On Mon, Mar 30, 2015 at 1:33 PM, Brandon Jones [email protected]
wrote:

I like to think of the current syntax as implying "I want you to display
this content fullscreen on device X". The "VR" part of it may be overtly
specific, but suits the immediate need.

Thus far, we haven't seen any devices where it's meaningful to render
anything other than a fullscreen view for the purposes of VR. Beyond that,
however, it serves as a useful base for progressive enahancement that does
almost the right thing in several situations. For example: On Android
even if you have no support for HMD distortion or anything like that you
can render a split screen canvas fullscreen and it looks acceptable in a
Cardboard harness. If we then add distortion support the same command will
now get automatically "upgraded" to a more correct version while still
being backwards compatible with other/older browsers. Same goes for
something like the Oculus Rift. Even if "VR fullscreen" isn't natively
supported you could at least mirror displays and use fullscreen to see an
image that swims around the edges. (Better than nothing!) That's the kind
of evolutionary track that the web thrives on.

—
Reply to this email directly or view it on GitHub
MozillaReality/webvr-spec#3 (comment).

from webxr.

Comment by jbaicoianu
Friday Oct 23, 2015 at 07:13 GMT

One important question this raises is how screen duplication would work with this system. Currently in FF Nightly with Oculus Rift 0.7.0 runtime, upon entering fullscreen with a vrDisplay the canvas is moved to the Rift display and what's left on the monitor is just whatever was behind the canvas.

Would it be possible to send an option to requestFullScreen() to do screen mirroring itself? I believe the Oculus SDK has a way to do this in an efficient way with the new direct driver mode.

Barring that, there's the option of duplicating the canvas element's content at the app level, but probably at the price of performance.

from webxr.

Comment by toji
Tuesday Feb 23, 2016 at 20:57 GMT

Issue is no longer relevant with upcoming API changes. Closing.

from webxr.