Re: Acessibility of <audio> and <video>

2008/8/25 Ben Boyle <benjamins.boyle@gmail.com>

> A transcript is a lot of information to keep in an attribute. Maybe a
> way to associate a transcript would be nice.


The assumption here is that the best way to accessibly represent video or
audio is as a transcript. That may be true, but it is an assumption
nonetheless. Let's unpick it.

Audio poses no significant accessibility problem for blind people who are
not deaf, as long as their browsers are capable of downloading or playing
the audio, and providing access to the controls for doing so. If these
problems are overcome (and they may be significant - try browsing MySpace
for audio with Lynx! - not that MySpace is HTML5 compliant, however...),
then audio is accessible for blind, non-deaf users.

Audio poses a problem for deaf people because they cannot hear it, but this
is important: it is not just the words that they cannot hear - it is the
audio as a whole which poses difficulties. Not all audio contains words.
Some audio may contain incomprehensible speech. How should that extra
information be represented, if at all? Should music, for instance, be
transcribed in musical notation? If so, in what mark-up language and can
that mark-up language be embedded into HTML5?

Video (with audio) poses problems for both blind and deaf people.

Good video captioning systems mitigate problems for visually-impaired people
by using large type, and using colours, italics and other textual cues to
denote the subject being referred to by the transcript (noises off described
in italics, for instance; and alternate colours for each successive person
talking), and for fully blind people some amount of extra description may be
needed, to describe the changing visual imagery - to turn films, for
instance, into radio plays with additional visually descriptive narration.

Video for deaf people typically requires captions or sign-language overlays.
It is important that the information should be displayed over or immediately
adjacent to the video, and in synchronisation with it, because one cannot
watch a video and simultaneously read a long, separate transcript, in any
convenient fashion. (NB. Making the transcript viewable as a whole could
still be useful for those who just want to scan the transcript, but this is
quite a different thing from making a video accessible to a deaf audience.)



> Personally, I would include the transcript in the page content and
> just group or link the two together. I reckon transcripts are useful
> to everyone, I hate wading through videos looking for useful content
> (text is much faster to scan)

I agree. IMHO, the questions for the mailing list are: how? and should this
"how" be specified in HTML5?


> Despite thinking this is a useful approach, and wishing more sites
> employed it, I don't think it should be mandated in the spec.

I.e. you think it should be recommended but not required?


> I'd love
> to see the spec encourage it, but more importantly I need the spec to
> tell me how to do it.

I agree wholeheartedly with this and the rest of Ben's comments.

Sam