ogg dev - Jan 2008 - Ogg/Kate, a karaoke and text codec for embedding in Ogg

Hi there,

I'm not sure what the right place would be, please  direct me to where is
most appropriate
if not this list.

I am currently working on a codec designed for karaoke and text information,
to be embedded
in an Ogg container. It includes text, motion information, and
style/placement.

At the moment, the bitstream design is still in flux, but draws heavily from
Vorbis and Theora,
while leaving scope for backward compatible future changes.

As a proof of concept, I have the codec working for subtitles in MPlayer
(text only for now)
and in some unreleased software of mine, as well as patch to oggmerge to
multiplex it with
Vorbis and Theora (BTW, I added Theora support to oggmerge, which oddly
didn't have it,
maybe I was looking at an old version ? It was from the xiph svn).

I would like to get feedback on the design, as this is the first step
towards having a complete
spec for the bitstream format, and the Xiph people are the best people to
ask.

Is this the right place to post such a thing ?

Thanks in advance
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://lists.xiph.org/pipermail/ogg-dev/attachments/20080114/2e0cba48/attachment.htm

On 14/01/2008, ogg.k.ogg.k@googlemail.com <ogg.k.ogg.k@googlemail.com>
wrote:
> Hi there,
>
> I'm not sure what the right place would be, please  direct me to where
is
> most appropriate if not this list.
Hi,

yes this is an appropriate list.
> I am currently working on a codec designed for karaoke and text
information,
> to be embedded
> in an Ogg container. It includes text, motion information, and
> style/placement.
ok, that sounds very interesting :-)
> As a proof of concept, I have the codec working for subtitles in MPlayer
> (text only for now)
> and in some unreleased software of mine, as well as patch to oggmerge to
> multiplex it with
> Vorbis and Theora (BTW, I added Theora support to oggmerge, which oddly
> didn't have it,
> maybe I was looking at an old version ? It was from the xiph svn).
great, please either send that patch here, or create a ticket at
trac.xiph.org and attach it there.
> I would like to get feedback on the design, as this is the first step
> towards having a complete
> spec for the bitstream format, and the Xiph people are the best people to
> ask.
probably the best thing to do would be to document your design so far
on wiki.xiph.org, and link from:

http://wiki.xiph.org/index.php/Work_In_Progress

Then we can discuss that here :-)

cheers,

Conrad.

Yes, this is the right place.

On 1/14/08, ogg.k.ogg.k@googlemail.com <ogg.k.ogg.k@googlemail.com>
wrote:
> I am currently working on a codec designed for karaoke and text
information,
> to be embedded
> in an Ogg container. It includes text, motion information, and
> style/placement.
Cool.  Have you looked at CMML so far?  What you describe is very
alike what CMML is supposed to do.
> as well as patch to oggmerge to
> multiplex it with
> Vorbis and Theora (BTW, I added Theora support to oggmerge, which oddly
> didn't have it
When possible for you, please post this patch in the tracker.  It's
most welcome.
> I would like to get feedback on the design, as this is the first step
> towards having a complete
> spec for the bitstream format, and the Xiph people are the best people to
> ask.
Feel free to post more information about it.

-Ivo

> yes this is an appropriate list.
Good, good :)
> great, please either send that patch here, or create a ticket at
> trac.xiph.org and attach it there.
Alright, I'll post it. Since Theora seems to be the currently pushed codec,
I was doubting I had the correct version.
> probably the best thing to do would be to document your design so far
> on wiki.xiph.org, and link from:
>
> http://wiki.xiph.org/index.php/Work_In_Progress
>
> Then we can discuss that here :-)
OK, I'll post what I have probably tomorrow.

Thanks for the quick answer.

> Cool.  Have you looked at CMML so far?  What you describe is very
> alike what CMML is supposed to do.
Yes, I saw this a little after I'd started (I'd only seen Writ
originally). I didn't have an
in-depth look at it however.

The things that made me discount CMML were:

 - it seems to be more about meta description of other streams in a
container than
   about being data by itself - Kate streams can be used as subtitles,
but do not have
   to be, they're intended to be self contained. An e-book, for
instance, could be a
   single text stream. There will be more in the documentation I will
post later.

 - it seems unnecessarily complex for my use (allowing links to the
web, for instance,
   which is just fine for meta description of other content for which
it is designed) and
   would need extra definitions for semantics. However, since the Kate
specification
   does not cover "external" file formats, it would be very possible
to create a Kate
   stream from a CMML XML file (for the moment, I build my streams
from a custom text
   based description, which is meant to be temporary, but I planned to
have encoding
   from existing formats (SSA, etc), so CMML would be another possibility).

 - it seemed to be designed for something else (Annodex, which I
didn't have a look
   at, but which seems to be a kind of archiving/retrieval system for
audio/video), and
   discussions about embedding it into Ogg seemed to be old and not
going anywhere
   and peter out (though that's only through Google searches, I may be
missing a lot).

Well, there's another thing also, which is that it's fun to write your
own :)