Re: WebVR and DRM

We have discussed the potential of having a Video Layer in a future version
of the spec. The primary purpose of said layer would be ease of use and
throughput efficiency, but if we were going to have any DRM support that
would be the place for it. Please note that use of a layer mechanism like
that would carry with it some significant restrictions, and then DRM
content would add several more on top of that:

   - WebGL content wouldn't be able to sample the video, so things like
   refraction effects in overlaid content wouldn't be possible.
   - Occlusion is possible through careful stenciling and alpha channel
   use, but not easy.
   - Videos would have to conform to one of only a couple known layouts.
   - Synchronization of WebGL content over the top of the video may be
   difficult.
   - If using a DRM encrypted video mirroring capabilities would likely be
   disabled.
   - DRM encrypted media support may not be universal, since it relies on
   support from both the VR system, the graphics drivers, and the OS.

In other words, expect any such mechanism to primarily be useful for very
simple 360 video or 2D-video-in-a-theater style scenarios.

And please recognize that DRM is not a magic bullet. If pixels can be
reproduced on screen they can be copied, full stop. Providing value through
your online services rather than any single piece of content is typically a
better route to ensuring users are willing to pay for your product, though
I recognize that's a fairly situation-dependent thing.

--Brandon

On Mon, Jul 10, 2017 at 2:13 AM Bassbouss, Louay <
louay.bassbouss@fokus.fraunhofer.de> wrote:

> Dear Florian,
>
>
> Thx for your quick reply. I am asking this question based on a discussion
> we had in a VR Workshop at 6th FOKUS Media Web Symposium
> <https://www.fokus.fraunhofer.de/go/mws/detailedprogram> where the Work
> of WebVR CG was also introduced (By Francois Daoust from W3C). Through EME
> API you can play encrypted video on the Web so the question is valid why
> you cannot play encrypted 360° video on the Web. I know there are technical
> limitations as you mentioned.
>
>
> BR,
>
> Louay
>
>
> ------------------------------
> *From:* Florian Bösch <pyalot@gmail.com>
> *Sent:* Monday, July 10, 2017 10:21 AM
> *To:* Bassbouss, Louay
> *Cc:* public-webvr@w3.org; Pham, Stefan
> *Subject:* Re: WebVR and DRM
>
> Dear Louay,
>
> DRM is an ineffective and usability hindering measure that serves no
> purpose whatsoever at all. In order to render anything in WebVR you have
> got to render it in WebGL. To render a video in WebGL the plain texture has
> to be uploaded to WebGL (i.e. to the GPU). That means that a
> gl.texture2D(yourvideo) call would decrypt the obfuscated video frame and
> put the plain video frame into GPU vram, where it can be readPixel'ed or
> otherwise be read out. Since DRM attempts to obfuscate the plain content,
> that can obviously not work, so it will never be possible to render DRM
> content inside a usable WebGL context. The best you could hope for is
> convince a GPU manufacturer to include functionality in drivers to overlay
> a portion of the screen with DRM'ed content. However, such content could
> not be shaded, depth-tested, stencil'ed, transformed, etc. It would be
> entirely outside of the normal rendering pipeline, and therefore be quite
> useless in 3D applications.
>
> If you'd like to render video in WebVR (i.e. WebVR) please consider not
> using user hostile technologies because as just demonstrated, they destroy
> usability and make everything extremely awkward, for no real gain. The
> presence of DRM on a video does not preclude it landing on TPB in any way.
> As demonstrated in plenty of examples the main purpose of DRM is to
> introduce vendor-lock-in and make it harder for the competition by
> dishonest means. If your offering cannot exist without the inclusion of
> user-hostile technologies, perhaps this is a clear sign that your offering
> is garbage.
>
> Thanks,
> Florian Bösch
>
> On Mon, Jul 10, 2017 at 9:15 AM, Bassbouss, Louay <
> louay.bassbouss@fokus.fraunhofer.de> wrote:
>
>> Dear Group members,
>>
>>
>> I am not sure if this topic is already discussed in the group. Is there a
>> way to render DRM content for example encrypted 360° videos using WebVR? If
>> not, do you plan to address this issue in the future maybe in WebVR Working
>> Group Charter?
>>
>>
>> Thanks,
>>
>> Louay Bassbouss
>>
>
>