[Vorbis] .ogg extension and Theora

[Stephen So]

<004c01c456e8$09d56d40$62389c3f at computername>
Message-ID: <40D62688.1050109 at griffith.edu.au>

Personally, I'm all for an extension change but since Xiph is refusing
to budge on this, then standing still is simply not an option and we
have to make it work somehow.

vorbis-bounces at xiph.org wrote:

>From: "Stephen So" <s.so at griffith.edu.au>
>
>
>
>>With regards to using *.ogg for everything (audio, video, or both), new
>>tools for detecting the content of ogg files is definitely needed for
>>Windows.  This is probably a trivial task as ogginfo probably can report
>>
>>
>
>It's not a "Windows" problem.  (From what I gather from some of the Linux
>users in here, it has the problem too.  And probably the Mac as well.)
>
>

I never said it was just a Windows problem.  But since a large
proportion of computer users out there are Windows users, and
considering the current lack of tools for Windows, I think it is a good
platform to start with.

>The same problem effects the web.  Anytime you see a .ogg file on the web or
>a ftp site, or a p2p program, you still wont automatically know what kind it
>is.  All of those apps would have to be updated as well, just to satisfy the
>desires of Xiph.org
>
>

As I said above, the solution I'm proposing is not in anyway trying to
justify or support the single extension camp.  It doesn't solve every
problem (eg. the one you speak about) but it's just an idea to make do
with what Xiph will eventually offer.

>Considering the internet can't even agree on using ipv6, or on running
>patched up to date servers, or firewalls, or close their SMTP ports, etc.
>etc., I think it's pretty close to zero that everybody is going to install
>.ogg handlers for their web based programs.  And for the OS and all their
>applications.
>
>
>Xiph just made a mistake releasing Vorbis with the .ogg extension oh so long
>before releasing anything else to go into that container.  In nearly
>everybody's mind (users and developers), .ogg means vorbis audio, and
>nothing else.  Not even other codecs such as speex.
>
>

I'm curious about Microsoft's *.asf extension for their advanced
streaming format.  It seems that some portable players claim they
support *.asf as if it were some audio container, when most people
recognise it as containing mostly video (and audio).  I bet the reason
Microsoft gets away with *.asf is because their players are under their
control, they produce the specs, and others (hardware manufacturers)
follow that spec.    Since the specs for the ogg container rests with
Xiph.Org, then I'm sure there is an inherent responsibility for third
party developers to follow the Ogg standards.   Luckily for us,
Microsoft doesn't do support for Ogg so they won't do a
'pseudo-standards' approach like they did with J++ or WWW standards.

>They didn't bother making clear that it should be able to handle other types
>of data.  Even a vorbis audio decoder should be able to handle text (song
>lyrics etc.), multiple pictures (album art) etc.  (This is actually a good
>argument for trying to talk everybody into using a new .oga audio extension.
>It could be designed from the beginning to handle things like that.  But I
>don't think it'd be easy to talk users into changing.)
>
>Trying to patch it by requring OS and app handlers to be written to try and
>sort out the types is a kludge.
>
>
>
>>clicking on each file would start the appropriate player.   Basically
>>what I am saying is we should *ditch Windows explorer* and have our own
>>"Ogg explorer" (hey, that's another name to try :) ) which will open the
>>
>>
>
>I'm sure you can talk Microsoft into adding that to XP, WinMe, Win98, etc.
>etc....
>
>

Did I ever suggest that?  I detect a sense of irrationality here.  When
I say *ditch Windows explorer*, I mean *we* (the user) should ditch it
for organising our ogg files, not Microsoft.

>
>
>>If we make Ogg Explorer downloadable (and supported) by Xiph.Org, and
>>
>>
>
>Most users aren't going to want to install a large program just so they can
>play a media file.  A codec is one thing.  A larger file explorer is a
>little bit different.
>
>

Ever heard of Jukebox software?  Users seem to like iTunes a lot for
handling their m4a files and categorising based on artist, genre, album,
etc.  And no-one said it has to be big.  Even an interface of some
simple program like EncSpot would do and it was quite a small app.  It
is all about adding some abstraction on top.

>Then you have to deal with all the bugs in the program.  (There will be.)
>And maintain it.  Forever.
>
>

Every program has bugs so its a trivial point.  Considering how simple
this explorer program is (does sorting, uses existing libraries to read
Ogg headers, exec external programs), it should be a bit less buggy than
full blown internet suites like Netscape 6 lol.

>And you have to accept the performance penalty involved in opening each and
>every .ogg file to see what it really is.  That involves not only quickly
>opening it and reading the header, but reading enough of the file to make
>sure that there isn't video or such at the very end.
>
>

Music organisation programs like iTunes, Winamp's media library, etc.
can do this type of thing and get information by reading the headers and
tags.  It is not such a big deal reading the header of an Ogg file and
seeing what type of streams are inside.  Ogginfo does it, VorbisExt does
it too and they read Vorbis comments as well.

>And you have to deal with bugs in it.  Which might cause it to crash or lock
>up, or just get very slow when it reads a damaged .ogg file.  (Like how XP
>does with damaged .avi files.)
>
>This is not an elegant solution.
>
>

That is moot point.  Bugs exists in all software.  And they all can
crash and lock up our systems, so  I don't see why we don't use this
excuse to quit using every complex program (or OS) because it is not
elegant enough.

>
>The idea of a generic container is a little attractive, but there are a lot
>of problems involved.  And Xiph didn't address any of them back when they
>would have been easier to solve.
>
>
>
>>boast about it its efficiency in organising your ogg files via content,
>>then I'm sure the average user will want to give it a try (when they
>>
>>
>
>I wouldn't.  I doubt the 'average' user would either.
>
>My 'organizing by content' of media files usually involves moving video
>files to a seperate directory from my music files.   I think most people do
>about that and little else.
>
>

If average users can use jukebox software, I'm sure this type of program
for organising their ogg files, just as iTunes organises their m4a files
based on genre, artist, album, etc., is not rocket science for them.
File abstraction is a clever option to make life easier, should Xiph go
ahead with *.ogg for all.