webm metadata

[nzasch]

Hi

anybody knows if there are any plans toward metadata tags inclusion
inside webm?

Also mkvmerge refuses to insert tags when creating webm files (-w). I
think matroska specification for Tags is ok, and that a modern media
format must carry a little of metadata.

Anyway, firefox 4.0 seems to play correctly Tagged mkv files
containing vorbis+vp8 with extension .webm

regards
Nzasch

[Philip Jägenstedt]

Huh, mkv files as in files with the "matroska" doctype? I thought all
browsers had dropped support for that already.

--
Philip Jägenstedt
Core Developer
Opera Software

[nzasch]

hi

Mozilla/5.0 (X11; Linux x86_64; rv:2.0b3pre) Gecko/20100721 Minefield/
4.0b3pre plays http://storage.arav.ventuordici.org/test/pippo6.webm

mkvinfo pippo6.webm | grep 'Doc type'
|+ Doc type: matroska
|+ Doc type version: 2
|+ Doc type read version: 2

I see that, when encoding in webm, ffmpeg 6.0 put the 'title'
metadata (-metadata title=) inside 'Segment information' (the 'General
name of the segment' as http://www.matroska.org/technical/specs/index.html
states).
If one also specifies -metadata description=xxx the metadata doesn't
appear into mkvinfo output.

cheers
nzasch


On 22 Lug, 12:16, Philip Jägenstedt <phil...@opera.com> wrote:

[Philip Jägenstedt]

Reported at https://bugzilla.mozilla.org/show_bug.cgi?id=580982, let's
hope the Mozilla developers fix this.

Philip

[Vladimir Pantelic]

Philip J�genstedt wrote:
> Reported at https://bugzilla.mozilla.org/show_bug.cgi?id=580982, let's
> hope the Mozilla developers fix this.

fix? you meant, let's *break* video playback....

[Philip Jägenstedt]

On Thu, 22 Jul 2010 14:30:51 +0200, Vladimir Pantelic <vlad...@gmail.com>
wrote:

Yes, matroska files should not work in browsers, or we shouldn't have
changed the doctype to webm to begin with. (As it happens I was opposed to
a new doctype, but that's all history now.)

[Vladimir Pantelic]

Philip J�genstedt wrote:
> On Thu, 22 Jul 2010 14:30:51 +0200, Vladimir Pantelic<vlad...@gmail.com>
> wrote:
>

>> Philip J�genstedt wrote:
>>> Reported at https://bugzilla.mozilla.org/show_bug.cgi?id=580982, let's
>>> hope the Mozilla developers fix this.
>>
>> fix? you meant, let's *break* video playback....
>
> Yes, matroska files should not work in browsers, or we shouldn't have
> changed the doctype to webm to begin with. (As it happens I was opposed to
> a new doctype, but that's all history now.)

pardon my ignorance, but why should they not work?

[Philip Jägenstedt]

On Thu, 22 Jul 2010 15:05:02 +0200, Vladimir Pantelic <vlad...@gmail.com>
wrote:

> Philip Jägenstedt wrote:
>> On Thu, 22 Jul 2010 14:30:51 +0200, Vladimir
>> Pantelic<vlad...@gmail.com>
>> wrote:
>>

>>> Philip Jägenstedt wrote:
>>>> Reported at https://bugzilla.mozilla.org/show_bug.cgi?id=580982,
>>>> let's
>>>> hope the Mozilla developers fix this.
>>>
>>> fix? you meant, let's *break* video playback....
>>
>> Yes, matroska files should not work in browsers, or we shouldn't have
>> changed the doctype to webm to begin with. (As it happens I was opposed
>> to
>> a new doctype, but that's all history now.)
>
> pardon my ignorance, but why should they not work?

The reasoning was that people should only use tools that know about WebM
for <video>, to increase the chance that the files will conform to the
WebM constraints (about keyframes, etc). For that to be effective, no
browsers should support the matroska doctype. If they do, some pages will
only be tested in that browser and the other browsers (like Opera) will
have to add support for matroska, making the webm doctype pointless.

[nzasch]

they should not... because http://www.webmproject.org/code/specs/container/#demuxer_guidelines
states:
'The demuxer must only open webm DocType files'

This is a natural and silly consequence of adopting a subset of a
container crafted to be flexible. I don't' understand why do not adopt
the matroska container stating that <video> only supports matroska
files with extension .webm containing vp8 + vorbis. Mandatory elements
into matroska specs are not so much.

As I maintain an embeddable media archive where ALL data is obtained
from metadata tags inside files, seems that if I hadn't pointed out
this bug, I would have had a way to serve vp8 encoded videos via the
<video> tags, at least on firefox (joking :). As I think that
important metadata (author ? license ?) must travel with data, I will
stay on ogg(vorbis,theora) with this archive.

Anyway the webm specs states "At initial release, WebM supports a
subset of the Matroska specification. Support for additional Matroska
functionality will be under consideration as the project matures." and
http://www.webmproject.org/code/specs/container/#tagging is a cut and
paste of http://www.matroska.org/technical/specs/tagging/index.html,
so I hope that Tagging features will be considered soon.

PS: interesting lecture -> http://sjohannes.wordpress.com/2010/04/28/matroska-tagging-support/

cheers
nzasch

[Lachlan Hunt]

On 2010-07-22 15:56, nzasch wrote:
> they should not... because http://www.webmproject.org/code/specs/container/#demuxer_guidelines
> states:
> 'The demuxer must only open webm DocType files'
>
> This is a natural and silly consequence of adopting a subset of a
> container crafted to be flexible. I don't' understand why do not adopt
> the matroska container stating that<video> only supports matroska
> files with extension .webm containing vp8 + vorbis. Mandatory elements
> into matroska specs are not so much.

All the arguments were made for and against this change prior to WebM
going public. Despite objections, it was decided that browsers should
only support the "webm" DocType, and should reject "matroska". As
Philip said, Opera argued against it, but we accepted the decision in
the interests of interoperability.

> Anyway the webm specs states "At initial release, WebM supports a
> subset of the Matroska specification. Support for additional Matroska
> functionality will be under consideration as the project matures." and
> http://www.webmproject.org/code/specs/container/#tagging is a cut and
> paste of http://www.matroska.org/technical/specs/tagging/index.html,
> so I hope that Tagging features will be considered soon.

The webm-discuss list is the correct place for such discussions. Feel
free to raise the issue there, or chime in on an existing thread on the
topic.

--
Lachlan Hunt - Opera Software
http://lachy.id.au/
http://www.opera.com/

[Lachlan Hunt]

On 2010-07-22 16:59, Lachlan Hunt wrote:
>> Anyway the webm specs states "At initial release, WebM supports a
>> subset of the Matroska specification. Support for additional Matroska
>> functionality will be under consideration as the project matures." and
>> http://www.webmproject.org/code/specs/container/#tagging is a cut and
>> paste of http://www.matroska.org/technical/specs/tagging/index.html,
>> so I hope that Tagging features will be considered soon.
>
> The webm-discuss list is the correct place for such discussions. Feel
> free to raise the issue there, or chime in on an existing thread on the
> topic.

Oops. Thought I was responding to a different list. Ignore that
comment. :-)

[Vladimir Pantelic]

Philip J�genstedt wrote:

>> pardon my ignorance, but why should they not work?
>
> The reasoning was that people should only use tools that know about WebM
> for<video>, to increase the chance that the files will conform to the
> WebM constraints (about keyframes, etc). For that to be effective, no
> browsers should support the matroska doctype. If they do, some pages will
> only be tested in that browser and the other browsers (like Opera) will
> have to add support for matroska, making the webm doctype pointless.

So, to sum it up:

a "playable" file must have

1) mimetype webm
2) .webm extension
3) doctype "webm"
4) VP8+Vorbis inside

right?

failing 4) will make in unplayable regardless of 1)-3) because
of possibly missing codec(s).

failing any of 1)-3) will make it unplayable although the
demuxer/decoder is perfectly capable of playing it...
(since doctype "webm" support is one strcmp()...)

[Lachlan Hunt]

On 2010-07-22 17:23, Vladimir Pantelic wrote:
> So, to sum it up:
>
> a "playable" file must have
>
> 1) mimetype webm

Yes. video/webm or audio/webm.

> 2) .webm extension

The extension is not strictly relevant in an HTTP context where the MIME
type is specified by the Content-Type header. It is relevant in a local
file system context, where the MIME mapping is done based on file
extension conventions. It may also be relevant to the server
configuration, if you're .htaccess file says:

AddType video/webm .webm
AddType audio/webm .weba

But, as far as the browser is concerned, those extensions could be
anything you like, or nothing at all.

> 3) doctype "webm"
> 4) VP8+Vorbis inside

Yes.

> failing 4) will make in unplayable regardless of 1)-3) because
> of possibly missing codec(s).

Yes.

[Steve Lhomme]

Thanks for your work on that Matroska tag work. I'm also trying to push Matroska tags on my (Matroska) side of things. There are currently not much players that can handle the depth of Matroska tags, nor simple tools to create them. I'm going to work on that because I strongly believe it is necessary to tag video files the same way people usually do it for audio. When you have 100s of video files lying around in various place it's better if you can know quickly what's inside and possibly organise them in a DB without having to move them or rename them in a priorietary way (that will never be able to handle all the possible tag information).

As a reference for comparing existing tag systems, you can look at this:

http://hobba.hobba.nl/audio/tag_frame_reference.html

--
You received this message because you are subscribed to the Google Groups "WebM Discussion" group.
To post to this group, send email to webm-d...@webmproject.org.
To unsubscribe from this group, send email to webm-discuss...@webmproject.org.
For more options, visit this group at http://groups.google.com/a/webmproject.org/group/webm-discuss/?hl=en.

[Silvia Pfeiffer]

In the Web context, you may also want to look at the work of the media
annotations working group in defining a relevant set of metadata
elements, see http://www.w3.org/TR/mediaont-api-1.0/ and
http://www.w3.org/TR/mediaont-10/ . I wouldn't want the tags to be
restricted to their set, but it makes sense to include them.

Cheers,
Silvia.

[Vladimir Pantelic]

Lachlan Hunt wrote:
> > On 2010-07-22 17:23, Vladimir Pantelic wrote:
>> >> So, to sum it up:
>> >>
>> >> a "playable" file must have
>> >>
>> >> 1) mimetype webm
> >
> > Yes. video/webm or audio/webm.
> >
>> >> 2) .webm extension
> >
> > The extension is not strictly relevant in an HTTP context where the MIME
> > type is specified by the Content-Type header. It is relevant in a local
> > file system context, where the MIME mapping is done based on file
> > extension conventions. It may also be relevant to the server
> > configuration, if you're .htaccess file says:
> >
> > AddType video/webm .webm
> > AddType audio/webm .weba
> >
> > But, as far as the browser is concerned, those extensions could be
> > anything you like, or nothing at all.

granted, but since I assume the "playability" of the file should not change
after e.g. downloading it, so the extension better be .webm...

> >
>> >> 3) doctype "webm"
>> >> 4) VP8+Vorbis inside
> >
> > Yes.
> >
>> >> failing 4) will make in unplayable regardless of 1)-3) because
>> >> of possibly missing codec(s).
> >
> > Yes.

and the other case I stated?