Re: embed ogg theora update
Benjamin Smedberg
> http://openmediacoalition.org/index.php/Main_Page) projects feel it
> would be ideal if firefox had native support for ogg theora playback. To
> that end we have detailed the features of a browser component
> http://meta.wikimedia.org/wiki/Firefox_Ogg_Support
> and started working with the VLC developers towards that end. There are
> likely elements that could be supported from the Mozilla side of
> development but the technical details have not been fully established.
What is the ultimate goal of this project?
If the goal is to get something into the default download of Firefox, then
we are going to need to start a process to identify
1) what user needs we're trying to fulfill (goals and nongoals)
2) How much it's going to cost
2a) codesize
2b) integration or import of external code
2c) Mozilla developer resources
3) In addition, Mozilla does not ship code that is not compatible with the
MPL/LGPL/GPL tri-license. Since VLC is only available under the GPL, it is
hard to imagine making any VLC code part of the default download. Perhaps
there are other options, such as including the plugin/extension in the
Firefox plugin-finder, and/or doing a network download of the extension for
users that want it.
> These extended features are part of the larger project of making the
> mozilla platform more multimedia friendly.
I strongly suggest posting to mozilla.dev.platform and mozilla.dev.planning
about this project. We currently are working on a product plan for Firefox
3, as well as data gathering for how we can improve the browser for users.
It would be foolish to plan on a large project such as this without
coordinating with the appropriate module owners as well as product planning
teams. If appropriate, we may be able to allocate one of the weekly "Gran
Paradiso" planning conference calls to discuss your proposals.
--BDS
michael
as us...@domain.ivalid just re-installed)
Benjamin Smedberg wrote:
>> We at wikipedia and other open media (open coalition
>> http://openmediacoalition.org/index.php/Main_Page) projects feel it
>> would be ideal if firefox had native support for ogg theora playback. To
>> that end we have detailed the features of a browser component
>> http://meta.wikimedia.org/wiki/Firefox_Ogg_Support
>> and started working with the VLC developers towards that end. There are
>> likely elements that could be supported from the Mozilla side of
>> development but the technical details have not been fully established.
>
> What is the ultimate goal of this project?
>
> If the goal is to get something into the default download of Firefox, then
> we are going to need to start a process to identify
>
> 1) what user needs we're trying to fulfill (goals and nongoals)
> 2) How much it's going to cost
> 2a) codesize
> 2b) integration or import of external code
> 2c) Mozilla developer resources
> 3) In addition, Mozilla does not ship code that is not compatible with the
> MPL/LGPL/GPL tri-license. Since VLC is only available under the GPL, it is
> hard to imagine making any VLC code part of the default download. Perhaps
> there are other options, such as including the plugin/extension in the
> Firefox plugin-finder, and/or doing a network download of the extension for
> users that want it.
>
yea I don't think the short term goal is to make it part of firefox but
as least as easy as flash to install. (so being part of the
plugin-finder would be good). In terms of firefox 3 supporting this
feature set. Libtheora http://www.theora.org/ is essentially public
domain so I imagine it would be compatible with the Mozilla license. It
will be a bit more work to build the plugin from libtheora then to use
what is already working with VLC but would give mozilla a bit more
control over how its implemented. Maybe something for "Mozilla planning"
to consider...
>> These extended features are part of the larger project of making the
>> mozilla platform more multimedia friendly.
>
> I strongly suggest posting to mozilla.dev.platform and mozilla.dev.planning
> about this project. We currently are working on a product plan for Firefox
> 3, as well as data gathering for how we can improve the browser for users.
> It would be foolish to plan on a large project such as this without
> coordinating with the appropriate module owners as well as product planning
> teams. If appropriate, we may be able to allocate one of the weekly "Gran
> Paradiso" planning conference calls to discuss your proposals.
>
will do...
--michael
Jonas Sicking
Personally I think we should really consider shipping ogg support builtin.
The reason i think this is worth considering is that it's lately become
apparent that media formats is a huge and important battle, and I think
we could be a strong influence here. However there are some requirements
that would have to be fullfilled:
1. We need to be able to pull in an existing library with a compatible
license that is well owned and used by other projects. We don't want to
end up in a situation where we have to take ownership of the library
when time rolls around for security patches and such.
2. We need to get other major browser vendors on the same track.
Possibly not all of them, but at least a couple of other big ones. This
can be especially hard since some of the vendors have vested interests
in formats of their own.
3. It needs to be well integrated into the web platform. I.e. web
developers should be able to easily use this in a cross browser fashion
and interact with it to build games and such. So just confining it to
the plug-in box is not a good idea. Then we might as well just treat it
as any other plug-in and not ship it by default.
/ Jonas
squaredancer
to generate the following:? :
if you guys could pull it off to get a "uniform" media format that
actually *WORKS*, you will be ordained as Saints of the Internet during
your lifetime!
As is, both Video and Audio formats are catastrophic - requiring a
browser to have at least 2-3 players for each media - and even then they
often don't function!!
reg
michael dale
Flash video is getting close to becoming the standard for web page based
video services. The problem is wikipedia can't and shouldn't host their
video on youtube. Likewise open media advocates can't and shouldn't use
flash as the basis for an open collaborative media platform. Flash has
enabled some interesting social video projects(jumpcut, motion box) but
to realize these types of projects and future ones in an all open
software environment will require more platform development.
IHO firefox should support ogg audio/video vorbis/theora as a base
media. But handle it in such as way that IE or safari users could render
the same page code with the vlc plugin. (Similar to how SVG is handled
using adobe or native firefox) Firefox users seeking to decode more than
just ogg theora with open source software could install the vlc plugin
which already handles many media formats and their respective embed tags.
--michael
Gervase Markham
> Personally I think we should really consider shipping ogg support builtin.
I agree this would be great. I'm sick of going to websites and not being
able to view audio/video inline. Having built in support would encourage
sites to use the format. And, as the original page suggests, we could do
cool things with integrating it into the page layers.
> 1. We need to be able to pull in an existing library with a compatible
> license that is well owned and used by other projects. We don't want to
> end up in a situation where we have to take ownership of the library
> when time rolls around for security patches and such.
libogg, libvorbis and libtheora fit these criteria. However, including
them would have a significant effect on download size. Perhaps we
wouldn't need all of them but:
gerv@otter:lib$ ls -l libogg.so.0.5.3 libvorbis.so.0.3.1 libtheora.so.0.2.0
-rw-r--r-- 1 root root 15632 2006-05-16 00:46 libogg.so.0.5.3
-rw-r--r-- 1 root root 233868 2006-08-24 12:31 libtheora.so.0.2.0
-rw-r--r-- 1 root root 163980 2006-05-02 14:44 libvorbis.so.0.3.1
gerv@otter:lib$
That's 400k.
> 2. We need to get other major browser vendors on the same track.
> Possibly not all of them, but at least a couple of other big ones. This
> can be especially hard since some of the vendors have vested interests
> in formats of their own.
I know WHATWG has specified a simple Audio object. Do they have plans in
this area?
> 3. It needs to be well integrated into the web platform. I.e. web
> developers should be able to easily use this in a cross browser fashion
> and interact with it to build games and such. So just confining it to
> the plug-in box is not a good idea. Then we might as well just treat it
> as any other plug-in and not ship it by default.
Indeed.
(Setting follow-ups to mozilla.dev.platform)
Gerv
Jonas Sicking
>> 1. We need to be able to pull in an existing library with a compatible
>> license that is well owned and used by other projects. We don't want to
>> end up in a situation where we have to take ownership of the library
>> when time rolls around for security patches and such.
>
> libogg, libvorbis and libtheora fit these criteria. However, including
> them would have a significant effect on download size. Perhaps we
> wouldn't need all of them but:
>
> gerv@otter:lib$ ls -l libogg.so.0.5.3 libvorbis.so.0.3.1 libtheora.so.0.2.0
> -rw-r--r-- 1 root root 15632 2006-05-16 00:46 libogg.so.0.5.3
> -rw-r--r-- 1 root root 233868 2006-08-24 12:31 libtheora.so.0.2.0
> -rw-r--r-- 1 root root 163980 2006-05-02 14:44 libvorbis.so.0.3.1
> gerv@otter:lib$
>
> That's 400k.
Holy hand grenade batman! That is very big indeed and would very likely
be a blocker. Anyone know these libraries and would be able to tell if a
significant portion of that size can be removed? Do they contain
features we don't need by any chance?
What are the sizes when compressed? Usually it's the download size we
worry about and the download files are zip compressed.
If 400k download size is what is needed i'm afraid ogg is a nonstarter.
>> 2. We need to get other major browser vendors on the same track.
>> Possibly not all of them, but at least a couple of other big ones. This
>> can be especially hard since some of the vendors have vested interests
>> in formats of their own.
>
> I know WHATWG has specified a simple Audio object. Do they have plans in
> this area?
I bet that the audio stuff could easily be expanded to include support
for ogg vorbis. For ogg theora I think it would be enough to make it
possible to point <img> tags to an ogg theora file. Possibly with the
addition of some API for starting/stopping the animation as well as
jumping to a particular point. Come to think of it, that might even be
useful for gif.
/ Jonas
michael dale
I bet libthora could be a bit smaller than that when packaged... those
400k files sizes are the uncompressed library files. The ogg
theora/vorbis java decoder cortado which i use on metavid is around 180k
or so. Some size saving may come from only including the decoder, but
some interesting possibilities come from including encoding and live
local stream capture... which will inevitably add some weight to
firefox. Perhaps we stick to it as a plug-in only until a meg or so
extra won't matter and the web environment is more open media friendly.
Like I mentioned in a previous post the native support should be handled
in a way that is compatible with other browser plugins. Integration into
hypertext standards is a good idea, but more of a long term strategy.
Overwriting <img tag maybe not the best approach. Adding an API for
stopping jumping to particular point good idea :)
--michael
Robert Kaiser
> Gervase Markham wrote:
>> libogg, libvorbis and libtheora fit these criteria. However, including
>> them would have a significant effect on download size. Perhaps we
>> wouldn't need all of them but:
>>
>> gerv@otter:lib$ ls -l libogg.so.0.5.3 libvorbis.so.0.3.1
>> libtheora.so.0.2.0
>> -rw-r--r-- 1 root root 15632 2006-05-16 00:46 libogg.so.0.5.3
>> -rw-r--r-- 1 root root 233868 2006-08-24 12:31 libtheora.so.0.2.0
>> -rw-r--r-- 1 root root 163980 2006-05-02 14:44 libvorbis.so.0.3.1
>> gerv@otter:lib$
>>
>> That's 400k.
>
> Holy hand grenade batman! That is very big indeed and would very likely
> be a blocker. Anyone know these libraries and would be able to tell if a
> significant portion of that size can be removed? Do they contain
> features we don't need by any chance?
>
> What are the sizes when compressed? Usually it's the download size we
> worry about and the download files are zip compressed.
>
> If 400k download size is what is needed i'm afraid ogg is a nonstarter.
On most Linux systems nowadays, libogg and libvorbis are already
installed, esp. if you're using a newer desktop environment (IIRC, KDE
depends on vorbis for its system sounds, GNONE might do so as well).
Theora is not that spread yet though, AFAIK.
I don't know what's the story for Windows and Mac though.
Robert Kaiser
michael dale
Theora integrates into mac and pc respective audio video systems. With
QuickTime components for mac [http://www.xiph.org/quicktime/] and direct
show filters for PC http://www.illiminable.com/ogg/
... but these extensions to QuickTime and DirectShow are not widely used...
--michael
Jonas Sicking
> I bet libthora could be a bit smaller than that when packaged... those
> 400k files sizes are the uncompressed library files. The ogg
> theora/vorbis java decoder cortado which i use on metavid is around 180k
> or so. Some size saving may come from only including the decoder, but
> some interesting possibilities come from including encoding and live
> local stream capture... which will inevitably add some weight to
> firefox. Perhaps we stick to it as a plug-in only until a meg or so
> extra won't matter and the web environment is more open media friendly.
Not including a decoder just because you can't also fit an encoder seems
like a very bad idea to me. The web has done fine without a gif or jpeg
encoder for years, with very little demand to have it included, and I
would think the same applies to ogg.
Could anyone find out what the size would be of compressed libraries
that only support decoding?
> Like I mentioned in a previous post the native support should be handled
> in a way that is compatible with other browser plugins. Integration into
> hypertext standards is a good idea, but more of a long term strategy.
> Overwriting <img tag maybe not the best approach. Adding an API for
> stopping jumping to particular point good idea :)
We can certainly come up with syntax that would be compatible with
browsers that only support ogg as a plugin. However in addition to that
I think we should ensure that the firefox support integrates well with
the rest of the web.
We're all about moving the web forward.
/ Jonas
michael dale
basic native decoder support would be awesome!
#theora says that libtheora decoding can be integrated at around 75k,
and libogg at around 15k. Synchronized multimedia playback that
_actually works_ on multiple platforms will add a bit of weight to it
beyond the decoding.
let me know if I can help push the native ogg decoding forward in any way.
--Michael
Robert O'Callahan
back to plugins on other browsers.
We could also support the WHATWG Audio object and SVG audio and video
elements. The latter integrate with SMIL to control playback and
synchronization.
This is hugely imporant and would be really nice to have, I'd strongly
support it.
Rob
Vladimir Vukicevic
> Gervase Markham wrote:
>
>> including them would have a significant effect on download size.
>> Perhaps we wouldn't need all of them but:
>>
>> gerv@otter:lib$ ls -l libogg.so.0.5.3 libvorbis.so.0.3.1
>> libtheora.so.0.2.0
>> -rw-r--r-- 1 root root 15632 2006-05-16 00:46 libogg.so.0.5.3
>> -rw-r--r-- 1 root root 233868 2006-08-24 12:31 libtheora.so.0.2.0
>> -rw-r--r-- 1 root root 163980 2006-05-02 14:44 libvorbis.so.0.3.1
>> gerv@otter:lib$
>> That's 400k.
>>
> Holy hand grenade batman! That is very big indeed and would very
> likely be a blocker. Anyone know these libraries and would be able to
> tell if a significant portion of that size can be removed? Do they
> contain features we don't need by any chance?
>
> What are the sizes when compressed? Usually it's the download size we
> worry about and the download files are zip compressed.
As folks have mentioned, these are the full uncompressed binaries, including
support for encoding, not just decoding. The download size shouldn't be
an issue.
I'm very interested in figuring out ways to improve the media platform that
Firefox provides to the web. However, it would have to integrate well with
HTML; the WhatWG Audio spec is a decent early start, but its goal is really
simple audio playback, for sound effects and the like, and isn't really usable
for even simple streaming or other solutions. We need a somewhat more complex
representation for this, but /not/ a full media framework (e.g. no encoding,
no complex "pipelines", etc.). I'd planned to take a stab at this at some
point in the near future.
- Vlad
Ludovic Hirlimann
> simple audio playback, for sound effects and the like, and isn't really usable
> for even simple streaming or other solutions. We need a somewhat more complex
> representation for this, but /not/ a full media framework (e.g. no encoding,
> no complex "pipelines", etc.). I'd planned to take a stab at this at some
> point in the near future.
At least for audio there is already something on the mozilla tree. zmk
which is on the zap branch.
Ludovic
--
http://perso.hirlimann.net/~ludo/blog/
Ognyan Kulev
> libogg, libvorbis and libtheora
I would recommend including Ogg Speex <http://www.speex.org/> decoder too.
Regards,
ogi
Gervase Markham
> At least for audio there is already something on the mozilla tree. zmk
> which is on the zap branch.
That uses Speex, doesn't it? Speex is only suitable for voice, not music...
Gerv
Magnus
Did someone ever file a (meta) bug to track this? It would really be sooo sweet!
(I guess http://meta.wikimedia.org/wiki/Firefox_Ogg_Support is the wiki page for it)
-Magnus