Re: Support for advanced caption features (inc rollup)

On Wed, Dec 12, 2012 at 11:31 AM, Ian Hickson <ian@hixie.ch> wrote:

> On Tue, 11 Dec 2012, Christian Vogler wrote:
> >
> > Basically it says: if you distribute TV programming on the Internet, you
> > must also send the closed captions, and the software or hardware that
> > you provide to users for viewing your programming must display those
> > captions.
>
> Publishers don't generally provide such software. Who's to blame if ABC
> puts a WebM and VTT file pair on the Internet, Jane Doe creates a Web page
> with <video> elements pointing to it, and the user runs Opera to read it,
> and the captions don't show? Who should the user sue? On what grounds?
>

I think ultimately it goes back to the programming provider. So, the idea
is to make it technically possible and as simple as possible to provide
compatible caption displays across different browsers and hosting sites and
publishers, such that even if you need to customize your video player, at
least you can make use of the caption functionality of the browser
rendering engine on all devices that use such a browser rendering engine
(which encompasses more and more white goods and similar devices that are
hard to upgrade).

I think in practice it might be better to just have the browsers that feel
> they need to follow these regulations implement straight 608 captions
> (that is, just a binary file consisting of the raw caption byte pairs, or
> maybe binary files in the 708 wrapper), and the people who want to put out
> TV programming on the Web include 608 captions, and then for the 608
> captions to be entirely ignored by everyone, with the "real" captions
> transmitted in VTT in a layout optimised for the Web. Then the regulations
> are satisfied, and yet we can still get on with doing sensible stuff on
> the Web (rather than worrying about whether we have to support "serif
> monospace" or text with sunken edges or scrolling captions).
>

That would require browsers to implement a second format, which is
something we've wanted to avoid.

Instead, there would probably be extensions to WebVTT. Hopefully they would
be specified such that if not implemented, they would still result in
acceptable rendering in Web browsers. At least, this is how I approached
the "Region" spec.


Cheers,
Silvia.