[theora-dev] Development status?

[Ralph Giles]

On Wednesday, April 23, 2003, at 11:15 am, Rodolphe ORTALO wrote:

> Do you mean a RFC-style document concerning the algo.,
> packet and stream format;
> or a man-style set of programming documentation for
> the various functions offered by the codec?

Obviously we need both. I'd suggest that a bitstream decoder spec for 
the first beta and api docs for the second beta are reasonable goals.

> (Corollary: what's the usual on-line documentation
> format for Xiph projects: man, docbook, other?)

Docbook is our usual format, mostly because it supports both on-line 
and offline output. Actual man pages have traditionally been in man, 
but docbook works for that too; they're equivalent because there are 
converters both ways.

ReStructuredText has been suggested as well, and at some point we'll 
want to do an RFC version of the spec. I'm not sure what the best 
source format for that is.

> Is there some provision in the ogg stream format for
> managing such aspects related to multiple media (synchro
> issues, acceptable loss, etc.)? Should it be done at
> this level or in a upper layer? (Doing it all via
> metadata does not seem totally sensible to me.)

Ogg supports multiplexing several independent bitstreams; it's up to 
the player to handle seeking and sychronization with help from the 
decoders for the various substreams. In that sense it's handled at a 
higher layer.

But theora streams must generally be playable, so codecs are designed 
to facilitate the aspects of multimedia you mention. 'Theora' is 
therefore both a codec specification for a single packet stream and a 
profile for embedding that stream in a transport format with other 
data. Vorbis is like this too, of course.

It's the profile level that's concerned with subtitles, et al. The 
'metaheader' mentioned in the  todo is an attempt to address stream 
management. Monty has proposed that we have something like this for 
stream description and to provide demux instructions. Whether it's a 
'separate' part of the ogg stream (which would break the current spec) 
or just another substream is only a technical issue. However, good 
software mustn't rely on the instructions when the stream can otherwise 
be decoded, so those instructions are really only hints, and I remain 
unconvinced that they're worth the additional specification complexity.

  -r

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to 'theora-dev-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.