ogg dev - Feb 2007 - "Any who has ever written an Ogg (de)muxer curses it's name frequently" (sic)

>From some random dude on Slashdot
[ http://yro.slashdot.org/comments.pl?sid=224420&cid=18177356 ]:

"Second, whether AVI is better or worse than Ogg is debatable. Any who
has ever written an Ogg (de)muxer curses it's name frequently. It's
extremely codec-specific, and the format is rather loosely defined,
with no consistent standard way to do much of anything....

AVI has it's limitations, but they are few. Most of the problems
people experience with AVI is due entirely to limited software which
doesn't properly handle AVIs. The rest of the problems tend to be a
result of lack of standards... For instance, Vorbis can fit into AVI
just fine, but unfortunately, Xiph didn't define HOW exactly, so
everyone has started doing it in their own, mutually incompatible way.
Ditto for subtitles, and other meta-data.

So, the biggest problem with AVI is lack of any single official
standards authority. Ogg has the same problem, but worse, since Xiph
have ignored all efforts to extend Ogg to handle other formats, and
now nothing is compatible."

I'm bringing up this message for discussion.

"For instance, Vorbis can fit into AVI just fine, but unfortunately,
Xiph didn't define HOW exactly, so everyone has started doing it in
their own, mutually incompatible way."

I wasn't even aware of this.  I'm sure this wasn't intentional, and
surely there's bigger priorities on what must be done around here, but
if this is true, it's pretty bad.

Vorbis shouldn't be a format only for Ogg.  Since its specification is
under the public domain, we should allow it to work on any other
container, no matter how bad or proprietary it is, because only by
allowing freedom of use to Vorbis will it become mainstream.

Out of curiosity, how hard is to write a tiny API explaining what is
the official method of having Vorbis contained in other formats?

"Ogg has the same problem, but worse, since Xiph have ignored all
efforts to extend Ogg to handle other formats, and now nothing is
compatible."

According to Wikipedia, Tobias Waldvogel has donated all Ogg Media
code to Xiph.  Why then is Ogg still unable to cope with any other
format outside of Xiph?  Are there technical problems?  Will we have
this working one day?

Food for thought, really, but this kind of little buggers are stalling
interest in Xiph's work.  I think we must find solutions and get a
roadmap going, or something in that regard.

-Ivo

On Wed, Feb 28, 2007 at 07:02:40PM +0000, Ivo Emanuel Gon?alves wrote:
> "Second, whether AVI is better or worse than Ogg is debatable. Any who
> has ever written an Ogg (de)muxer curses it's name frequently. It's
> extremely codec-specific, and the format is rather loosely defined,
> with no consistent standard way to do much of anything....
> 
> AVI has it's limitations, but they are few. Most of the problems
> people experience with AVI is due entirely to limited software which
> doesn't properly handle AVIs. The rest of the problems tend to be a
> result of lack of standards... For instance, Vorbis can fit into AVI
> just fine, but unfortunately, Xiph didn't define HOW exactly, so
> everyone has started doing it in their own, mutually incompatible way.
> Ditto for subtitles, and other meta-data.
So it's hard to write an ogg muxer, and the problem with AVI is that 
there's no good software support? Maybe it's just hard to write a muxer?
> "For instance, Vorbis can fit into AVI just fine, but unfortunately,
> Xiph didn't define HOW exactly, so everyone has started doing it in
> their own, mutually incompatible way."
If there's a reasonable way to do this, sure. I assume it can't be too 
bad, since there are lots of xvid+mp3 videos in avi on the net. But no 
one's ever explained how one does this, so I don't actually have an 
opinion. I just know some xiph people have said the proposals aren't 
good technically, and some other people have done it anyway.
> Out of curiosity, how hard is to write a tiny API explaining what is
> the official method of having Vorbis contained in other formats?
specification you mean? Yes, those can go as appendicies to the vorbis 
spec, along with the ogg and rtp embeddings.
> According to Wikipedia, Tobias Waldvogel has donated all Ogg Media
> code to Xiph.  Why then is Ogg still unable to cope with any other
> format outside of Xiph?  Are there technical problems?  Will we have
> this working one day?
Are you talking about tobias' directshow filters? They've not been 
maintained in years, and we've been recommending illi's instead because
they were and seem to work better.

 -r
> Food for thought, really, but this kind of little buggers are stalling
> interest in Xiph's work.  I think we must find solutions and get a
> roadmap going, or something in that regard.
We've got roadmaps a plenty. What we need is people doing things. :)

 -r

On 2/28/07, Ivo Emanuel Gon?alves <justivo@gmail.com>
wrote:
> >From some random dude on Slashdot
> [ http://yro.slashdot.org/comments.pl?sid=224420&cid=18177356 ]:
>
> "Second, whether AVI is better or worse than Ogg is debatable. Any who
> has ever written an Ogg (de)muxer curses it's name frequently. It's
> extremely codec-specific, and the format is rather loosely defined,
> with no consistent standard way to do much of anything....
>
> AVI has it's limitations, but they are few. Most of the problems
> people experience with AVI is due entirely to limited software which
> doesn't properly handle AVIs. The rest of the problems tend to be a
> result of lack of standards... For instance, Vorbis can fit into AVI
> just fine, but unfortunately, Xiph didn't define HOW exactly, so
> everyone has started doing it in their own, mutually incompatible way.
> Ditto for subtitles, and other meta-data.
>
> So, the biggest problem with AVI is lack of any single official
> standards authority. Ogg has the same problem, but worse, since Xiph
> have ignored all efforts to extend Ogg to handle other formats, and
> now nothing is compatible."
I'm not going to defend Ogg here - writing Ogg code turns out to be
very difficult, as it works fundamentally differently from essentially
every other container format in widespread use, and requires a much
tighter coupling between codec and container.

But to suggest AVI is just unreasonable - AVI is more or less
obsolete, and for good reasons. It's extremely limited, and works
poorly for a wide range of purposes. Existing implementations also
tend to be poor. For the simple "I have a single audio and a single
video track in a file on my hard disk", it works reasonably well, but
that's about all. Even microsoft stopped using it (they've moved to
ASF, which isn't widely used outside of the windows media formats, but
is certainly more capable than AVI).
>
> I'm bringing up this message for discussion.
>
> "For instance, Vorbis can fit into AVI just fine, but unfortunately,
> Xiph didn't define HOW exactly, so everyone has started doing it in
> their own, mutually incompatible way."
I have been told by people much more familiar with the details of AVI
that this is simply not true (there are some horrendous hacks
possible, but these don't allow you to put vorbis into AVI, they allow
some subset of vorbis. I think I've also seen _ogg vorbis_ embedded in
AVI).

I believe the  primary issue is that AVI permits any of:
 - constant duration, constant packet size
 - constant duration, variable packet size
 - variable duration, constant packet size
but not variable duration, variable packet size - a core requirement
of vorbis. I may be wrong in some details there, but it amounts
roughly to that.

We're not opposed, in principle, to standardising an embedding of
vorbis into _other_ more capable container formats. But the core
developers don't really use those very much - after all, they're most
commonly seen in use for using non-free video codecs with vorbis, so
we'd need a proposal from the community - brought to us, for reasoned
debate and discussion. Nobody's done that.

Mike

On 2/28/07, Ralph Giles <giles@xiph.org> wrote:
> specification you mean? Yes, those can go as appendicies to the vorbis
> spec, along with the ogg and rtp embeddings.
That seems a reasonable approach.
> Are you talking about tobias' directshow filters? They've not been
> maintained in years, and we've been recommending illi's instead
because
> they were and seem to work better.
What I gathered about OGM was from the Wp article on it.  It says, OGM
is a hack of Ogg to get other codecs (MPEG 4, etc.) inside.
> We've got roadmaps a plenty. What we need is people doing things. :)
Word.


On 2/28/07, Michael Smith <msmith@xiph.org> wrote:
> But to suggest AVI is just unreasonable - AVI is more or less
> obsolete, and for good reasons.
I don't think anyone suggested AVI as a replacement for Ogg (good
heavens!).  That guy on Slashdot was making a comparison between them
both, in a reply to someone else.
> We're not opposed, in principle, to standardising an embedding of
> vorbis into _other_ more capable container formats. But the core
> developers don't really use those very much - after all, they're
most
> commonly seen in use for using non-free video codecs with vorbis, so
> we'd need a proposal from the community - brought to us, for reasoned
> debate and discussion. Nobody's done that.
I haven't looked into the proposal of the MPlayer guys yet (linked in
a message after yours), but maybe it suffices our needs.  If not, I
think it's better that Xiph proposes its own Vorbis hack for AVI.  I
guess we have to take in attention Matroska and MP4 as well.

On 2/28/07, Eric Dantan Rzewnicki <eric@zhevny.com>
wrote:
> Where would a person wanting to be doing things find these roadmaps?
Theora ToDo list on which anyone can contribute to get things done:
http://wiki.xiph.org/index.php/TheoraTodo

Google Summer of Code projects:
http://wiki.xiph.org/index.php/Summer_of_Code

On 2/28/07, Ralph Giles <giles@xiph.org> wrote:
> Well, there are todo pages at wiki.xiph.org, but I meant more in the
> community folklore sense. My point is a roadmap doesn't help much
unless
> there are people committed to making things happen. That's been the
> problem with a lot of this stuff, and why it's been so nice to see the
> ambisonics work happening.
The situation on Ambisonics is tricky, because it depends on someone
coding a whole API for the different Xiph projects AND Monty being
available to apply whatever changes are need in Vorbis.
> If you're one of those people, ask on irc or the mailing lists if you
> want an orientation.
For those who don't know:
irc://irc.freenode.net
#xiph
#theora
#vorbis
...a few more channels

A list of all of our mailing lists may be found at:
http://lists.xiph.org/mailman/listinfo/

Infodump post.

On 2/28/07, Gregory Maxwell <gmaxwell@gmail.com>
wrote:
> Official Xiph endorsement of patented codecs inside Ogg will cause
> problems for those who accept Ogg because it doesn't contain patented
> codecs. (Well there is xvid, but at least they call their files OGM).
It's not an endorsement.  Ogg is a container.  People are free to
throw whatever they want inside, and we should make sure that there's
some kind of standartization for this.  After all, if Ogg can only use
Xiph's formats, it's no better than proprietary containers.
> ... I fail to see the attraction from an adoption perspective. If
> someone is handing out oggs already then they are expecting the user
> to have special software to play them.
We dont want people to have special software.  Xiph's formats are
supposed to be built-in on operating systems.
> If the user has special
> software it's not unreasonable to expect that software to support the
> codecs that we offer.
The argument is about compatibility between that kind of software.
Imagine VLC and MPlayer render Ogg MPEG 4 Video differently.  They
will end up doing the same thing that guy on Slashdot was doing: blame
Xiph, because it didn't specify how they should interact with Ogg.
> Frankly, I see being non-encumbered a primary aspect of the Xiph brand
> identity. It would be foolish to step away from that just to remove a
> bullet point from the list of someone who is determined to complain.
> :)
Xiph will never step away from that goal.  That's the whole point of
why we are here.

Out of curiosity, are you the same GMaxwell from
Wikipedia/Commons/etc?  I salute you, then.  I follow your/his work
(occasionally), and you/he are/is really active there.



-Ivo

On Wed, Feb 28, 2007 at 09:45:54PM +0000, Ivo Emanuel Gon?alves
wrote:
> On 2/28/07, Eric Dantan Rzewnicki <eric@zhevny.com> wrote:
> >Where would a person wanting to be doing things find these roadmaps?
> 
> Theora ToDo list on which anyone can contribute to get things done:
> http://wiki.xiph.org/index.php/TheoraTodo
> 
> Google Summer of Code projects:
> http://wiki.xiph.org/index.php/Summer_of_Code
Ok, thanks. Those lists have many interesting projects. 

Does anyone still have interest in icecast or ices? the above lists
don't mention them and the TODOs in svn trunk are 2 and 3 years old.

-ERic Rz.