RE: Captions for audio clips

At 02:41 AM 12/20/99 -0500, Charles McCathieNevile wrote:
>SMIL doesn't provide for timing inside media objects - you can do that by
>breaking them into pieces and using explicit timing (that's what would be
>good to do).

If synchronization becomes so important should we try to talk to SMIL WG to
be able to do a SMIL file that contains both the text and timecodes? Or
maybe it should be a separate standard? Having one file for each text line
that needs to be synchronized is really cumbersome. And if the
synchronization changes the files need to change as well. A good tool may
help to manage this but still I would prefer the timecodes with the text.

Another thing is that the user should be able to control what media is
shown. The author design can be a default, but if the author has not
thought of every necessary combination, the user should still be able to
change it.

One question: if we say synchronization is necessary then what is the
amount of synchronization that is needed? Every sentence? Every paragraph?

I think synchronization is often at least P3 (often probably more) but in
some cases I would also like to be able to just read the text and not care
about the synchronization at all. Is there a way to have both? A SMIL
synchronization file that refers to different text portions in a text
transcript file (without forcing it to be divided into several files)? Or
just being able to ignore the time codes when reading synchronized text?

Marja

> The W3C validator should now validate SMIL documents (well, XML
>in general in theory, and SMIL is XML).
>
>I agree that it is at least P2, although I am not sure if it isn't in fact
>P1.
>
>Charles McCN
>
>On Wed, 15 Dec 1999, Bruce Bailey wrote:
>
>  Dear Phill (et al.)
>  
>  IMHO, it is a clear case of P2!
>  
>  Populations effected:  Persons for whom English is a second language. 
>   Persons who are not deaf but have impaired hearing.  Persons with
learning 
>  disabilities for whom processing auditory information is difficult (but
not 
>  impossible).
>  
>  The assumption is that ALL of the above persons might very well PREFER an 
>  audio stream for the SAME REASONS everyone else prefers audio over a text 
>  transcript.  Is it a useful exercise for use to delineate why an aural 
>  presentation is better (in some cases) than a textual one?
>  
>  >From this perspective, the situation is very analogous to persons with
VERY 
>  poor vision who STILL PREFER a GUI browser!  We are empathetic / 
>  sympathetic to this orientation.  Just as we accommodate the partially 
>  sighted, so should we adjust for the hard of hearing.
>  
>  For the above populations, "unimedia audio" represents a significant 
>  barrier to their access of content (we are using RealAudio radio
broadcasts 
>  as an example).
>  
>  For the above populations, a separate transcript has so little value as to 
>  be virtually useless -- just as access to Lynx is not well regarded a 
>  viable option for web surfing by most persons with vision impairments (nor 
>  most average people for that matter).
>  
>  It is, of course, important to have techniques on hand, but that should
not 
>  influence the assignment of Priorities.
>  
>  Does anyone have an example of captioned audio?
>  
>  I experimented with some SMIL file on my local hard drive.  I could get 
>  RealAudio (actually a .rm RealMedia file) to play ONLY the sound (with 
>  synchronized captions), but I could NOT get rid of the blank video window. 
>   Probably I am just doing something wrong, but I did look at the W3C SMIL 
>  specifications.  Does the W3C offer a SMIL validation service?
>  
>  Bruce Bailey
>  
>  
>  On Sunday, December 12, 1999 11:52 PM, Charles McCathieNevile 
>  [SMTP:charles@w3.org] wrote:
>  > Phill, if you are just reading it then that is the case. However for 
>  people
>  > who have marginal hearing, having the sound and the captions/score 
>  available
>  > and synchronized is more valuable than one or the other (similarly for 
>  people
>  > who can hear, but have difficulty reading). One of the challenges we
face 
>  is
>  > that there are people who are looking for multi-modal support - there are
>  > more people with poor hearing than there are with no hearing (and 
>  similarly
>  > for other disabilities).
>  
>  On Wednesday, December 15, 1999 11:45 AM, pjenkins@us.ibm.com 
>  [SMTP:pjenkins@us.ibm.com] wrote:
>  > JW:
>  >> It appears to be broadly agreed within the group that a requirement to
>  >> synchronize text transcripts with audio presentations should be
>  >> established, at least at a priority 2 level.
>  >
>  > PJ:
>  > Where is the broad agreement?  Bruce, Jason, and Charles seem to agree 
>  with
>  > P2.  I'm arguing for P3, and Robert and Eric seem OK with either P2 or 
>  P3,
>  > and I haven't heard form others.  I do agree that there seems agreement
>  > that we need to make the distinction between multimedia videos and 
>  unimedia
>  > sounds files in the errata so that WCAG 1.4 doesn't apply to the unimedia
>  > sound only files.
>  [snip]
>  > PJ:
>  > but I've heard no supporting rationale or any convincing evidence that
>  > suggests that the "value" is more than useful and improves accessibility
>  > [P3].
>  >
>  > Because the deaf,  [learning disabled, or those learning a foreign
>  > language] are so comfortable now with synchronized television (and movie)
>  > captioning, does not support the argument that they will be comfortable 
>  or
>  > have significant barriers removed with synchronized captioned audio only
>  > files.  Can anyone even show me an sample example, or better yet, a real
>  > example on the Web or anywhere?  If we don't add a supporting technique, 
>  a
>  > checkpoint requiring [even at P3] synchronized captions for audio only
>  > files shouldn't even be added to the guidelines.  I've seen natural
>  > language courses use techniques of synchronization to TEACH the language,
>  > but we're talking about guideline 1 - equivalent alternative information 
>  -
>  > not "teaching natural languages" or "teaching singing".  We have been
>  > talking about ideas and theories, how can we suppose that it fits the
>  > definition of "significant barriers".  P3 is still "valuable" and 
>  "useful"
>  > and "improves accessibility".
>  
>
>--
>Charles McCathieNevile    mailto:charles@w3.org    phone: +61 409 134 136
>W3C Web Accessibility Initiative                    http://www.w3.org/WAI
>21 Mitchell Street, Footscray, VIC 3011,  Australia (I've moved!)
>

Received on Tuesday, 21 December 1999 03:26:54 UTC