Re: Encrypted Media proposal

On Mon, 05 Mar 2012 18:29:33 +0100, Charles Pritchard <chuck@jumis.com>  
wrote:

> On 3/5/2012 1:43 AM, Philip Jägenstedt wrote:
>> On Fri, 02 Mar 2012 19:53:32 +0100, Charles Pritchard <chuck@jumis.com>  
>> wrote:
>>
>>> On 3/2/2012 3:27 AM, Philip Jägenstedt wrote:
>>>> 3. Rendering must likely use some form of overlay. Opera has this for  
>>>> some platforms, but it limits the functionality of the <video>  
>>>> element, e.g. CSS transforms (other than scale+translate) and CSS  
>>>> opacity won't work. We consider this a platform limitation and it is  
>>>> not something we really want to make a core requirement for <video>.
>>>
>>> I'd like to see this bit added somehow to the CSS Transforms spec.
>>>
>>> Flash is treated as an overlay in Chrome, it's behavior with transform  
>>> has been unstable, and it would've helped to have scale+translate  
>>> specified as an "ok" alternative. Flash is still broken in Chrome.
>>>
>>> I think we're going to need to figure out how to signal to the  
>>> scripting/CSS environment when CSS transforms are limited to  
>>> scale+translate (with rotate missing).
>>
>> There's really nothing to add to the spec, it's just a quality of  
>> implementation issue. The point is that overlays *limit* the quality of  
>> implementation possible, so I think it would be a bad idea to require  
>> it in any Web standard.
>>
>
> It's a recurring issue relating to the technical qualities of transforms.
>
> If a <video> or <object> tag can not respond to rotate, perhaps it  
> should not have rotate return in its computed styles and/or getters.
>
> Doesn't that make sense?

But it *can* be supported. On the platforms where we use overlays, we  
still fall back to passing raw frames in the cases where an overlay  
doesn't work. However, if a CDM refuses to pass decoded frames, then we  
can't fully support CSS transforms. CSS opacity is possible, but requires  
splitting everything into more layers. It would be much better if we  
simply didn't make overlays as part of the Web platform...

-- 
Philip Jägenstedt
Core Developer
Opera Software

Received on Tuesday, 6 March 2012 10:32:18 UTC