Re: Discussion on Caption Model

Silvia,

Thanks for sharing the Caption Model. It is a great start and helps us
understand the different elements that need to be considered.

Perhaps we need to develop a set of performance objectives (in other words,
user/functional requirements). Ultimately consumers who use the captions
need to determine the best way we want to receive our captions. We would
need to know which are features are available for us to control and which
features are determined by the author, and cannot be modified.

To provide one example, your document may assume that the only way to
display captions is to overlap it on the video. Some consumers have
expressed a desire, and preference, to see captions at the bottom of the
screen without overlapping with the video image similar to watching a
letterbox movie with captions.

Shane


On Fri, Oct 21, 2011 at 3:49 PM, Silvia Pfeiffer
<silviapfeiffer1@gmail.com>wrote:

> On Sat, Oct 22, 2011 at 3:04 AM, Olivier Thereaux
> <olivier.thereaux@bbc.co.uk> wrote:
> > On 17/10/2011 12:29, Silvia Pfeiffer wrote:
> >>
> >> I've called the document the "caption model - an abstract model of
> >> captions":
> >>
> >>               http://www.w3.org/community/texttracks/wiki/Caption_Model
> >>
> >> I may have missed some features of international caption models, so
> >> please have a thorough read and see if you're missing anything.
> >
> > Very good document Silvia. I am trying to understand what its breadth
> should
> > be: should it document all the use cases currently covered by all the
> major
> > captioning systems, or only what is common in the linear broadcast world?
> > I'm wondering if, for instance, there is an archived collection of use
> cases
> > which the TTML group used...
>
>
> Ah yes, good question - I should be more clear.
>
> The document is trying to capture all existing uses of captioning and
> subtitles so as to be able to understand what features we require from
> a timed text format.
>
> It would be best to check if the functionality of your favorite
> caption format is fully covered in this model and if not, explain what
> needs to be added/changed. The model should cover both, existing
> linear broadcast uses as well as existing online uses - and all of
> this on an international basis.
>
> I updated the intro to the wiki page to clarify this.
>
>
> As for TTML - I just checked and found this document:
> http://www.w3.org/TR/ttaf1-req/ as the collection of use cases for
> TTML. It doesn't, however, specify in an abstract way how text should
> be rendered, but instead lists a subpart of XSL styling paramters that
> should be usable by TTML. It is not very useful in that respect for
> checking whether all rendering cases that were envisaged to be needed
> for TTML are represented in the model document, since it provides the
> solutions without explaining the needs. There may be another document
> that explains what caption formats were analysed to come to this list,
> but I wasn't able to find it. If you have a link, please share.
>
>
> > Two other minor comments come to mind:
> >
> > 1) It is my understanding that captioning can use visual elements such as
> > color  (symbols, too - can be thought of as "text" for the sake of this
> > document?) to specify meta-information: who is speaking, whether the
> sound
> > being captioned is off-screen, etc.
>
> Yes. I guess the current focus of the document is on layout and not so
> much on text styling. It's good you point this out - I'll add it.
>
>
> > 2) There are use cases for the display of the caption contents outside of
> > the Video viewport (and sometimes, there is no viewport at all - think
> radio
> > transcripts). Is this something we want to add a provision for?
>
> We can mention it, but I'd like to keep this focused on video, because
> I think the devil lies in the details of the visual representation on
> top of video.
>
> When timed text is rendered outside of the video viewport, we end up
> with text oin the Web page and that need is covered by CSS.
>
> Where we want to render text for audio in a viewport, we actually can
> publish it in HTML5 through a video element and end up in the same
> situation as video, so this is covered. If there are special needs for
> positioning and styling there that isn't already met, do explain.
>
> Cheers,
> Silvia.
>
>