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

Re: Acessibility of <audio> and <video>

From: Leif Halvard Silli <lhs@malform.no>
Date: Tue, 02 Sep 2008 19:39:33 +0200
Message-ID: <48BD7A55.3060104@malform.no>
To: public-html@w3.org

Lachlan Hunt 2008-09-02 13.19:

> Leif Halvard Silli wrote:

     [...]

>>>> Justin James 2008-09-01 19.42:
>>>>> I can imagine the furor if we also applied this logic to
>>>>> images, by saying, "if you want accessible images, use a format
>>>>> that natively supports metadata of alternate text, or put a
>>>>> subtitle/caption/legend/etc. in your image."
>>>
>>> No. Unlike video, images have no way embedding accessibility features 
>>> within them, and their meaning is very often depending on the 
>>> context. However, videos do have various ways of including native 
>>> accessibility features, such as subtitles/captions or audio 
>>> descriptions.
>>
>> Our recent debate about EXIF proves opposite.
> 
> EXIF data is just metadata and is not supported in PNG or GIF. It is 


First, we are deviating: Justin took as premise that even if it is 
or eventually - had been - possible to add useful fallback 
directly in photos, we would still have offered a way to offer the 
same things directly in HTML. And ditto for video.

Back to the deviation: EXIF is not the sole point here. My photo 
eidtor allows me to insert metadata via XMP in PNG & GIF as well.

XMP-link: http://wikipedia.org/wiki/Extensible_Metadata_Platform

> certainly not optimised for accessibility purposes and only provides 
> very limited benefit.  Although it can contain descriptions, that's not 
> particularly useful when the alternate text needs to be context sensitive.

Need for context sensitivity is a good point. And it is obviously 
simpler to be context sensitive direcly in the page - in the 
markup - than if you have to edit the file format.

But it also a matter of what possibilities one has (users don't 
control the markup of Flickr) and what one are used to (if image 
meta data was a well known ting with good support, then I suppose 
we would also see some context sensitive use of it).

[...]

>>>> Add to that that movie formats are *very* often used for displaying 
>>>> photos.
>>>
>>> Could you elaborate on what you mean, provide some sort of evidence 
>>> to support this claim, and also explain why it's relevant?


>> Link: http://www.dagbladet.no/kultur/2008/09/01/545443.html
> 
>> One could have taken away the video functionality, and it would still 
>> be a photo. (I, for one, did not care to look at the video.)
> 
> That's just using the poster frame for its intended purpose: providing a 
> visual, iconic representation of the video to indicate what the video is 
> about.  It is not a substitue for actually watching the video.  Since it 


I feel you are avoding the point.

For those who want, there is no need to watch that video, as its 
poster is a photo which itself documents the incident. Presence of 
@alt text on a photo from some kind of "footage" should not depend 
on whether the photo was inserted via <video> or <img>.

> isn't just an image, it's a video, the most suitable additional 
> information to provide is not alternative text for the poster frame, but 
> rather an brief summary of what the video is about.


This smack of a kind of formalism. The poster /can/ be consumed as 
a regular photo, many are satisfied with that. That poster plays 
the role of a photo. Those who care can enjoy the video as well.

I don't speak generally. I speak about this example. I would not 
compare "Pirates of the Caribbean" with a news footage.

> One possibility worth considering is using the title attribute, or the 
> surrounding text (possibly the caption) to provide additional 
> information.  Conceptually, it would be like video description in the 
> sidebar of a YouTube video.

The poster images of Youtube videos almost never have the photo 
qualities - or <img> qualities, if you wish - as the combined 
photos/videos in the online newspaper we discuss. In the latter 
case, the video opportunity can be seen as an added benefit of 
simply looking at the poster photo - the poster photo has been 
"crafted" so you can enjoy it as a relevant photo.

> In other words, the poster frame is like the picture on the front cover 
> of a DVD, and the summary is like the blurb on the back that describes 
> what the movie is about.  Clearly neither of them are a substitute for 
> actually watching it, but both provide advisory information about its 
> content, which is useful to anyone.

Clearly you prefer to simplify. Perhaps Youtube can be compared to 
a video shop, and the posters images of the Youtube images to dvd 
boxes. But this is not the case for the article I pointed to.

However, even if we think about front covers on DVDs, I don't feel 
this adds up. Such covers also often contain text. And, posters 
typically function as PR for the video. Clearly, even those who 
are permamently or temporarely unable to see the poster photo want 
to be targeted by that "PR".

Looking at book covers is a good alternative to reading books, for 
many of us. But even if reading the book is the "real thing", this 
does not allow use to skip @alt for the <img> with the book cover.

>> The Dagbladet.no example I gave above shows that the poster image  can 
>> serve the role of a photo.
> 
> No, it serves as a visual representation of the video in order to give a 
> hint about what the video is about.


I feel you are presssing a premade concept over real life.

 
>> If the poster image shows Putin looking at a sleeping siberian tiger, 
>> then why do you need a separate <img> with the same motif?
> 
> I never suggested that you would.

Which confirms my point. (Not necessarily, but in this case.)

>> And just as I was more than satisfied from seeing the photo, likewise 
>> could anyone incapable of seeing the video be more satisfied with an 
>> @alt text for the poster photo. (Rather than having to be bothered 
>> with listening/seeing the entire video before getting any clue.)
> 
> No, what would you rather have? Alt text for just the poster frame or a 
> summary of the actual video?  Surely, the former is relatively useless 
> compared with the latter, and the summary is most certainly more 
> appropriate and useful for everyone.

What would I rather have: I enjoy the poster of the example we 
discuss - as a photo. I don't need to see the video. And an @alt 
text summary would also be fine if I had a text browser - so I 
could know that the footage was not worth seeing.

What you say implies that a photo is useless, that photos rather 
should have been videos. But of course, this is not the case.

>> This has me asking: Why couldn't the poster image just be a regular 
>> <img> element, rather than an attribute?
> 
> I can't recall the rationale for that.

<img>s have fallback.

>>> * Transcript of spoken content.
>>> * Textual descriptions of relavant non-spoken content.
>>>   e.g. descriptions of significant actions or sounds in the video.
>>> * Still images illustrating significant moments from the video.
>>>   e.g. images of presentation slides, if the video was of someone giving
>>>   a presentation.
>>
>> "Still images illustrating significan moments" - that is exactly the 
>> usecase we have in the Dagbladet.no newsarticle I sited.
> 
> Correct, and it does it's job quite well.  But what you seem to be 
> asking for is transcripts, textual descriptions and a short summary, but 


No, you misunderstand. I want fallback for the very poster image. 
Transcripts etc for the video itself, is another issue.

(Even if I link the two issues by suggesting that both the 
<poster> and the <fallback> should be inside the <video> element. 
In fact, I think that both <source> and the hypothetical 
<transcript></transcript> should be considered "fallback".)

> I already explained why all of that should be available to everyone, 
> rather than hidden away from all but assistive technology, which would 
> be the practical effect of including it within the element or its 
> attributes.


I agree that it is problematic that e.g. an <object> is kind of 
closed - it is a drawback that there is no way to prefer the 
alternative over e.g. the video. So, <object> and <video> needs 
rethinking (and so I offer my 2 cents ...)

However the example we discuss demonstrates that it is highly 
practical to combine the role of photo and video. It saves space, 
and can be good design. Very focused for the users.

Hence it should also be highly practical to place the link to the 
transcript - or whatever alternative version one would provide - 
above the poster photo. Clicking the transcript link, the <video> 
element could "open" and show the transcript.

 
>> The poster image should be viewed as <video> fallback!
> 
> Sure, it has some benefit for accessibility too, but it's not fallback. 


It should be viewed as fallback - technially. But not as /the/ 
fallback. It is a "permanent, yet temporary fallback", as most 
users prefers the photo before playing the video.

> It's available to everyone and shown regardless of whether or not they 
> choose to watch the video.  That's like arguing that seeing a movie 
> poster is fallback for going to the cinema and watching the film.

Unless the <video> poster is an <img> (or unless <video> has an 
@alt for the poster photo), then it is not available to everyone.

Secondly, of course, if the <video> contains the full video of 
"Pirates of the Caribbean", the poster is not a substitute for the 
video. Which is why the poster needs fallback in its own right.

The photo of Putin and the tiger is also not a substitute for the 
video from the same event. And that is why the photo needs a 
fallback of its own.

>> It is when looking at the fallback, that we should be offered the 
>> choice between how we want the video served: as slides, transcript or 
>> the very video.
> 
> Right, and I'm ok with providing all of that in a way that is available 
> to everyone, and providing it or a link to it from within the 
> surrounding content seems most appropriate for that.

It seems most focused to provide links to the alternatives in the 
very <video> element. Of course, sometimes it is also practical to 
spread out the links at different places on the page. But /having/ 
to do so might interfer with the intended design and might create 
a less focused experience for the readers. (Repetition is a 
problem which Ian has described as problem for AT users.)

>>> * A link to download the video, possibly in alternative formats, to
>>>   watch in an external media player, perhaps in several formats.
>>
>> When it comes to getting the video in another video format, then - if 
>> we are talking /fallback/ - this requires that the user /knows/ that 
>> he does not want the default format - and that he also knows that he 
>> doesn't want any of the (textual) fallback either.
> 
> What? I really don't understand what you are trying to say.  In 
> practice, it's somewhat common to provide links to media in alternative 
> formats.

I was referring to the problem you mentioned, namely that authors 
write <object data=film>Upgrade your videoplayer</object>.

> There are lots of podcasts that provide high quality or low quality 
> alternatives, in MP3, Ogg, streaming via a flash based player in the 
> page or subscribing via itunes.  Similarly, with videos, it has been 
> quite common to provide links to various formats targetted at QuickTime 
> or Windows Media player, and to provide high and low quality alternatives.
> 
> For example, Ask a Ninja provides both Flash and MOV (h.264/AAC), and 
> Diggnation provides both high and low quality streaming in the flash 
> player; links to small, large and high def versions in various codecs 
> and an audio only MP3.


Like I said, sometimes it is fitting to spread out links to 
alternative variants of the featured media material. Especially so 
if you want the user to load another page with that other format.

 

But <video> may contain "one or more source elements" so you can 
avoid those links, instead letting the UA to select the format.

Why not allow users to select the format themselves, from within 
the <video>, instead of forcing them to load a new page? It would 
also be more focused to keep everything in the same place instead 
of having to look outside the <video> element.

You jumped over the point that <video> could keep the poster image 
inside the element as - technically - fallback. (Though, from an 
agumentative authoring point of view, it would not be 'fallback'.)

In its most simple variant, it could be <video><img></video>. But 
perhaps something like this would be better (one could optionally 
change <figure> with a similar <poster> element):

<video><figure> ...markup... </figure>
        <source src="video-format" >
        <source src="another-video-format" >
        <transcript>Video transcript</transcript>
</video>

And I don't see why a poster should only be allowed on the <video> 
element. Since you have compared with DVD-boxes and cinema, we 
also have CD boxes - with their own "posters".

As Justin said: "Consistency goes a long way towards adoption."
-- 
leif halvard silli
Received on Tuesday, 2 September 2008 17:40:20 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:23 GMT