[theora-dev] using Kate for WebVTT encapsulation

[Silvia Pfeiffer]

On Sun, Apr 17, 2011 at 7:59 PM, ogg.k.ogg.k at googlemail.com
<ogg.k.ogg.k at googlemail.com> wrote:
>> What do you mean by adding baggage? Do you want to translate the
>> WebVTT styling to Kate styling?
>
> No, quite the opposite: I imagine a browser would have a WebVTT parser
> plugged after demuxers, so the best would be to just cut the input in timed
> chunks to mux them at the correct time. What I meant is that using libkate
> for this seems like using OpenOffice to create a small text file. It works,
> but you'd be quicker done using vi :)

Well, it's the same for SRT in Kate. But is it really worth creating a
new text track format to encapsulate WebVTT in Kate? I think Kate
would be the easiest.



> [example snipped]
>
> So that would lend itself well to embedding a whole raw cue per packet.

Yes, including the cue settings. I think that would be trivial (just
like you do SRT right now).


> Now, using libkate gives you working mux/demux code

Yes, that's where I think the advantage is: an existing mapping and
even existing code.

> and support in
> various software already,

Other than for the simplest SRT-like construct, I doubt that's true.

> still needing to plug to new WebVTT parsers,
> but means more complex code/format that creating a new simpler codec
> for WebVTT, which would be a subset of Kate.

I hesitate to do that, because it means we have multiple means of
encoding text tracks in Ogg. I'd much prefer if we can use Kate tracks
as the solution for encapsulating them all. The header will tell what
it actually is. Do you think that's a bad idea?

Silvia.