Re: Roll-up captions in WebVTT

On Mon, 28 Nov 2011, Christian Vogler wrote:
> >
> > That's like ten times more annoying to read than the normal captions 
> > on that video. I think if we are to support roll-up captions we need a 
> > far more convincing use case.
> 
> Two points in response:
> 
> 1. Roll-up captions are essential for live captioning where the text is 
> transmitted as the captioner types, especially when steno errors and 
> subsequent corrections are factored in.  Another issue is the sheer 
> number of words per minute during live captioning, where without roll-up 
> you would end up missing the context half of the time. This usage 
> scenario is very common during live broadcasts in the United States, 
> where captions are not only roll-up, but also appear word by word, 
> instead of line by line. I'll check around if I can find a video clip 
> that shows this clearly.

Since we're not supporting live captioning at all currently, that doesn't 
seem like a convincing use case either. If it's the only use case, then we 
should address it at the same time as live captioning.


> 2. The implementation of roll-up in this video is poor. The positioning 
> and alignment jumps around with every roll up, which of course makes it 
> very hard to read.
>
> Now come to think of it, a corollary is that roll-up support by itself 
> is not sufficient. You also need a way to paste on captions live word by 
> word for it to be truly useful.

If the use case is just live captioning, then I agree that just roll-up is 
insufficient. Currently, though, doing live captioning with external video 
text tracks in HTML is very much an issue that's up in the air. This is 
the bug for that topic:

   http://www.w3.org/Bugs/Public/show_bug.cgi?id=14104

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

Received on Monday, 28 November 2011 23:58:44 UTC