Re: Encrypted Media proposal (was RE: ISSUE-179: av_param - Chairs Solicit Alternate Proposals or Counter-Proposals)

On 3/1/2012 10:37 AM, Tab Atkins Jr. wrote:
> On Thu, Mar 1, 2012 at 10:06 AM, Mark Watson<watsonm@netflix.com>  wrote:
>> To clarify:
>> - a browser can have multiple CDMs for different keysystems
>> - what we propose to standardize is the discovery, selection and interaction
>> with CDMs, not the CDMs themselves (not unlike the current situation with
>> codecs).
> One point that has been made multiple times is that the current
> situation with codecs is *horrible*.  It is not a good solution that
> we should attempt to emulate; it's a filthy, painful hack that was the
> best we could do if we wanted a<video>  element, given the standoff
> between browsers on which to support.  If there was a way to go back
> in time and convince everyone to use a single codec, that would be
> *great*.

The feedback I've received on this issue is that vendors do not want a 
baseline codec to be part of specifications. Canvas+PNG is a fluke that 
was reluctantly picked up, for the sake of compatibility.

I surveyed vendors recently to see if we could have a baseline for 
lossless audio recording. Some of the reasoning I heard back was that we 
should not hook ourselves into a standard media format now which would 
be obsolete in the near future. I don't see a way to get around 
ideals/principles. It's not a standoff over one format or development. 
It's about not defining (or over-specifying) these parts of implementation.

Now, I'd hoped with lossless audio, we could have something that's 
simple, good enough, and non-controversial, as it'd be lossless and 
royalty free. But, that didn't float. And because that didn't float, I 
don't expect other items, such as video codecs, to float. We can get 
upset over that position, but it's not going to produce anything of 
substance. The parties involved have opposing views.

Let's keep in mind, <audio> playback/codec issues can be worked around:
https://github.com/JensNockert/aurora.js
http://codecs.ofmlabs.org/
"it is possible to play MP3 and Apple Lossless even in browsers without 
native support"

<video> requires a lot of processing power; hardware decoding is much 
preferred:
https://github.com/mbebenita/Broadway
http://libwebpjs.hohenlimburg.org/vp8/webm-javascript-decoder/

Having a baseline in video means hardware support, and that's going to 
have to happen through market forces, not here.

Mark's position, that the CDM is outside of spec, seems reasonable given 
the existing scene.

I do want to see an example implementation sometime, just as the WebRTC 
folk did.

And as for the market, we'll see where Google takes things, both with 
HTML5 Video over YouTube, and with WebM. They are setting that stage.

-Charles

Received on Thursday, 1 March 2012 19:12:09 UTC