W3C home > Mailing lists > Public > public-html@w3.org > September 2008

Re: Is longdesc a good solution? (was: Acessibility of <audio> and <video>)

From: Leif Halvard Silli <lhs@malform.no>
Date: Tue, 09 Sep 2008 17:20:06 +0200
Message-ID: <48C69426.7040107@malform.no>
To: Henri Sivonen <hsivonen@iki.fi>
CC: Matt Morgan-May <mattmay@adobe.com>, HTML WG <public-html@w3.org>, W3C WAI-XTECH <wai-xtech@w3.org>

Henri Sivonen 2008-09-09 11.36:

> On Sep 9, 2008, at 00:26, Matt Morgan-May wrote:
>> On 9/8/08 1:22 AM, "Henri Sivonen" <hsivonen@iki.fi> wrote:
>>> On Sep 7, 2008, at 22:31, Leif Halvard Silli wrote:
>>>
>>>> <video>: Would you propose the use of <object> instead of <video>
>>>> when HTML fallback for videos is wanted as well?
>>>
>>> No. I would propose that users who don't see the video track play the
>>> video and listen to the soundtrack. When the content provider makes an
>>> additional effort for addressing the not seeing the video track case,
>>> I'd suggest the effort be put into making an alternative audio
>>> description sound track.

Those hindered from listening would need the transcript most.

>> Where's the opportunity cost assessment on this one, then?
>>
>> It is clearly less work to produce a text equivalent and associate it, 
>> than to script, voice and associate a secondary audio description track.
> 
> You are right. Even though I suggested merely listening to the sound 
> track in the common case, I went straight to the rather impractical high 
> end with the more accessible suggestion.
> 
> How far-fetched is audio description authoring in your assessment? Does 
> it even make sense to build browser support for it?

As interesting as the answer to that question indeed is, it is 
also a question very much colored by the (commercial) meaning of 
the word "video". A <video> doesn't necessarily contain a video. 
It can just be a recording of a speech, for instance.

About speech: If <video> and/or <object> represents a streaming of 
a live event, then sometimes the manuscript will be made available 
in the same moment the speech is being held.  After the speech, 
what would the government publish on its web site for "common 
consumption"? The speech manuscript or the speech transcript?

It could make sense to offer the manuscript for public consumtion, 
and the transcript as fallback for the video. If transcript was 
unavailable, it could also make sense to offer as fallback a link 
- in the form of a @longdesc URI for instance - to the manuscript. 
  And I don't think that there will always be natural to place a 
link to the manuscript close to the <video>.

> In any case, a full-text transcript (including annotations recounting 
> significant visual happenings) does not belong in <object> fallback, 
> which is what Leif asked me about, because a full-text transcript--once 
> written--is useful in general. In particular, it is useful in cases 
> where it is *also* useful to make <object> render the media file (even 
> if only for the audio for blind and low-vision users and only for the 
> video for deaf or low-hearing users). The transcript should be available 
> on the same page or on another page via a plain link.


* What if clicking the "plain link" could "open" the fallback of 
the <object> for regular users? Why must <object> be so closed? Or 
perhaps I should say: Why must <video> be so closed?
* What if the AT user would like to listen to the 
transcript/manuscript being read by the familiar voice of the 
software instead of listening to the (unclear) voice from a live 
recording?

Making the manus/transcript available via @longesc doesn't prevent 
  it from being referenced in the juxtaposing text either.

 
>> Which is partly why WCAG 2 Guideline 1.2 offers authors the
>> choice of full-text transcript as an alternative, at level A.
>>
>> What you're proposing means that the fallback that's good enough for WCAG
>> would need to be done as a hack in HTML5, rather than a clean semantic
>> association.
> 
> Would this
> <video src=movie.ogg>Please upgrade to a browser that supports HTML5 
> video.</video><p><a href=transcript.html>Annotated transcript</a></p>
> be a "hack"?
> 
> Is a semantic association between the <video> element and the transcript 
> necessary if the link is very near the video in the document reading order?

A problem is that it might not be "very near" the video. E.g. if 
the fallback was a manuscript (since transcripts are hard work), 
then the author might not see a reason to place a link to the 
transcript  near the <video> if the link was provided elsewhere.
-- 
leif halvard silli
Received on Tuesday, 9 September 2008 15:21:11 UTC

This archive was generated by hypermail 2.3.1 : Monday, 29 September 2014 09:38:58 UTC