[whatwg] Thoughts on video accessibility

[Silvia Pfeiffer]

On Thu, Jul 16, 2009 at 11:56 PM, David Singer<singer at apple.com> wrote:
> At 23:28  +1000 16/07/09, Silvia Pfeiffer wrote:
>>
>>  > 2) I think the environment can and should help select and configure
>> type-1
>>>
>>>  resources, where it can.  It shouldn't need to be always a manual step
>>> by
>>>  the user interacting with the media player.  That is, I don't see why we
>>>  cannot have the markup express "this source is better for people who
>>> have
>>>  accessibility need X" (probably as a media query).  However, media
>>> queries
>>>  are CSS, not HTML...
>>
>> Would you mind providing an example that demonstrates the use of media
>> queries? I cannot currently imagine what that could look like and how
>> it could work. Feels free to use CSS in addition to any require HTML
>> (and javascript?). Since I cannot imagine what that would look like
>> and how it could work, I cannot start to understand it as an
>> alternative.
>
> sure. using deliberately vague way of writing the media queries
>
> <video blah blah ... >
>   <source src="xx-O.ers" media="want-captions" />
>   <source src="xx-N.ers" media="not want-captions" />
> </video>
>
> xx-O has open (burned in captions), uses the same codecs etc.  It gets
> selected if the user says they want captions, otherwise XX-N (no captions)
> is selected.
>
> <video blah blah ... >
>   <source src="xx-S.ers" media="want-sign-language" />
>   <source src="xx.ers" />
> </video>
>
> xx-S has a sign-language overlay capability.  It gets selected for those
> users expressing a positive preference for sign language; otherwise we don't
> waste the bandwidth loading that, and we load the plain xx file.  It may be
> that the media part of the UA also detects this user preference and
> automatically enables sign language in xx-S.

I just noticed that the "media" attribute is already part of the
"source" element definition in HTML5. I wonder which browsers have
implemented this attribute.

After having looked at http://www.w3.org/TR/css3-mediaqueries/, my
understanding is that media queries specify the different presentation
media that the html page's different stylesheets were built for and
thus allow choosing between these stylesheets through the "link"
element and its "media" attribute where the query goes. Also, IIUC,
the list of presentation media is currently restricted to ‘print’,
‘screen’ , ‘aural’, ‘braille’, ‘handheld’, ‘print’, ‘projection’,
‘screen’, ‘tty’, ‘tv’, and 'all' and the queries cover only the
features width, height, device-width, device-height, orientation,
aspect-ratio, device-aspect-ratio, color, color-index, monochrome,
resolution, scan, and grid.

This is different for the "source" elements though: instead of
specifying different presentation media and choosing between
stylesheets, the "media" attribute specifies different user
requirements and chooses between video source files. This makes it
independent from CSS, IIUC.

Is the intention to extend the specification of "media queries" to
include generic means of selecting between alternative files to load
into a HTML page? Is there a W3C activity that actually extends the
media queries to audio and video files?

If this is the case, it could also be used for the associated "text"
elements that Ian and I discussed earlier in this thread. The
alternatives there would be based on a combination of languages and
the different categories of time-aligned text. The language would
choose between different text files to load, and the text category
would choose between different default styles to apply.

I can imagine that that would work, but has anyone started extending
existing media query specifications for this yet?

Regards,
Silvia.