Re: WebVTT

On Mon, Jun 17, 2013 at 8:44 PM, Sean Hayes <Sean.Hayes@microsoft.com> wrote:
> Yes in that specific case it would be possible to map to the wavefront. However in the general case it won't.
>
>  I could write for example
>
> <p begin="11.3s" end="13.4s">
>   <span begin="0.6s">Somewhere</span>
>   <span begin="0.1s">over</span>
>   <span begin="1.1s">the</span>
>   <span begin="1.7s">rainbow</span>
> </p>
>
> Which would come out like:
>
> T1 --> T2
> <c.hidden>Somewhere</c> over <c.hidden>the rainbow</c>
>
> T3 --> T4
> Somewhere over <c.hidden>the rainbow</c>
>
> T5 --> T6
> Somewhere over the <c.hidden>rainbow</c>
>
> T6 --> T7
> Somewhere over the rainbow
>
> Spotting the difference between these two cases doesn't really seem worth the effort.

Right, now I understand what you mean by "generalising the inner
timing model". It's a very unusual case for captions. You'd probably
also have to do that by repeating text in WebVTT cues and marking
individual sections as hidden. Maybe like this:

T1 --> T2
<c.hidden>Somewhere</c> over

T3 --> T7
Somewhere over <T4>the <T5>rainbow<T6>

Cheers,
Silvia.

Received on Monday, 17 June 2013 12:21:32 UTC