[theora] theora and xml

Fri Dec 4 15:23:36 PST 2009

On Sat, Dec 5, 2009 at 9:37 AM, ogg.k.ogg.k at googlemail.com
<ogg.k.ogg.k at googlemail.com> wrote:
>> In any case: what I said earlier still holds - the best way to include
>> such XML into Ogg is through a Ogg Kate track, but you have to
>> serialise the data and write an encoder (or at minimum a transcoder to
>> the Kate input format).
>
> If it is timed (eg, different portions of XML correspond to various
> temporal segments), then I think a good way would be to include it as
> metadata to each of these temporal segments in a Kate stream. Metadata
> attached to timed segments, or other objects, is something that is not
> in a release of libkate yet, however (the bitstream is still backward
> compatible of course).
> If this is done in this way, then there are two possibilities:
> - include each timed portion of XML as a separate metadata (metadatum
> ?) in XML format (upon decoding, one can then retrieve this metadatum
> along with start/end times, and parse it as XML)

If you look at Tom's "CMML" example, you will see that the the "clip"
tags with start and end time exactly meet this requirement.

Everything that is in the <head>, however, would need to go into some
system-wide metadata - vorbiscomment or skeleton.

> - decompose each timed portion of XML into more basic blocks and
> include those as separate metadata for the same temporal segment
> (means you get a set of tag/value pairs, rather than XML, on
> decoding).

I don't think that is possible with the kind of hierarchically
structured markup that the rights description languages use. Creative
Commons: yes, but not the others.

> If it's not timed (eg, it applies to the whole track), then it's best
> added as Vorbis comment(s) to the Theora track.
>
> In any case, whatever you do, you'll need some software to retrieve
> the metadata at decode time, and do with it what it is one does with
> timed metadata.
>

That is a very good point to make. Tom: why do you even want this
stuff inside a media file? And what software do you have that will be
able to do something with it at decode time?

Cheers,
Silvia.