hosting WebM files on servers

773 views
Skip to first unread message

skylinestudios

unread,
May 24, 2010, 1:17:41 PM5/24/10
to WebM Discussion
How are WebM files served? Do I need something like Wowza Media
Server, or can it reside on a regular web server?

Pavel Koshevoy

unread,
May 24, 2010, 1:22:16 PM5/24/10
to webm-d...@webmproject.org
On 5/24/10 11:17 AM, skylinestudios wrote:
> How are WebM files served? Do I need something like Wowza Media
> Server, or can it reside on a regular web server?
>

Any HTTP 1.1 web server should do.
You'll need to tweak the server config to emit the correct mime info for
webm files.

http://www.webmproject.org/code/specs/container/#naming

Pavel.

Andy Berkheimer

unread,
May 24, 2010, 1:21:59 PM5/24/10
to skylinestudios, WebM Discussion
WebM files can be served using a standard HTTP server.

The HTTP Content-Type needs to be set correctly, for example in apache config:

<IfModule mime_module>
    AddType video/webm .webm
</IfModule>

salsaman

unread,
May 24, 2010, 1:43:10 PM5/24/10
to Andy Berkheimer, skylinestudios, WebM Discussion
For some application servers (e.g. Websphere), you will also need to add the mime type to /etc/mime.types

It's probably a good idea to do this anyway.


Salsaman.


http://lives.sourceforge.net
https://www.ohloh.net/accounts/salsaman

Dean Lee

unread,
May 25, 2010, 12:02:55 AM5/25/10
to Andy Berkheimer, skylinestudios, WebM Discussion
And 'audio/webm' for .webma . :D

Philip Jägenstedt

unread,
May 25, 2010, 3:29:50 AM5/25/10
to Andy Berkheimer, Dean Lee, skylinestudios, WebM Discussion
I don't believe a file extension for audio/webm has been decided. The most
common suggestion has been .weba.

Philip
--
Philip Jägenstedt
Core Developer
Opera Software

Andy Berkheimer

unread,
May 28, 2010, 10:27:30 AM5/28/10
to Philip Jägenstedt, Dean Lee, skylinestudios, WebM Discussion
The proposal of ".weba" for audio/webm has come up a number of times.  Support has been pretty universal, are there any objectors hiding out there?

-Andy

Matthew Heaney

unread,
May 28, 2010, 10:47:05 AM5/28/10
to Andy Berkheimer, Philip Jägenstedt, Dean Lee, skylinestudios, WebM Discussion
On Fri, May 28, 2010 at 10:27 AM, Andy Berkheimer
<andyber...@google.com> wrote:
> The proposal of ".weba" for audio/webm has come up a number of times.
>  Support has been pretty universal, are there any objectors hiding out
> there?

The file extension ".weba" works for me.

--Matt

Gaute Strokkenes

unread,
May 28, 2010, 10:56:43 AM5/28/10
to Andy Berkheimer, Philip Jägenstedt, Dean Lee, skylinestudios, WebM Discussion
Andy Berkheimer <andyber...@google.com> writes:

> The proposal of ".weba" for audio/webm has come up a number of times.
> Support has been pretty universal, are there any objectors hiding out
> there?

Does the world really need another container for vorbis-only audio
files?

--
Gaute Strokkenes

Lachlan Hunt

unread,
May 28, 2010, 11:17:09 AM5/28/10
to Gaute Strokkenes, WebM Discussion
On 2010-05-28 16:56, Gaute Strokkenes wrote:
> Does the world really need another container for vorbis-only audio
> files?

The fact that WebM can be used as another vorbis-only container comes
for free. The need to support audio-only files with a MIME type and
file extension recognises that authors would inevitably create such
files anyway, and at least gives a sane way for them to be supported.

Also, AIUI, WebM has a slight advantage over Ogg when seeking through
the file over HTTP, due to the presence of the Meta Seek information,
whereas Ogg lacks an index and seeking is a little more complex. When
seeking, the cues allows the implementation to more accurately find out
exactly which part of the file to request with an HTTP Range request, if
the file hasn't completely loaded.

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

Pavel Koshevoy

unread,
May 28, 2010, 11:53:45 AM5/28/10
to webm-d...@webmproject.org
On 5/28/10 9:17 AM, Lachlan Hunt wrote:
>
> Also, AIUI, WebM has a slight advantage over Ogg when seeking through
> the file over HTTP, due to the presence of the Meta Seek information,
> whereas Ogg lacks an index and seeking is a little more complex. When
> seeking, the cues allows the implementation to more accurately find
> out exactly which part of the file to request with an HTTP Range
> request, if the file hasn't completely loaded.
>

According to the webm guidelines CuePoints are for video frames only.
http://www.webmproject.org/code/specs/container/#muxer_guidelines

What does this mean for audio-only files?

Thank you,
Pavel.

Silvia Pfeiffer

unread,
May 28, 2010, 8:58:06 PM5/28/10
to Andy Berkheimer, Philip Jägenstedt, Dean Lee, skylinestudios, WebM Discussion
Total support from me!
S.

Sent from my iPhone

Philip Jägenstedt

unread,
May 29, 2010, 2:53:37 AM5/29/10
to webm-d...@webmproject.org, Pavel Koshevoy

You still get the duration in the header and don't need to seek to the end
of the file as with Ogg (before it recently got an index, anyway).

Steve Lhomme

unread,
May 29, 2010, 4:23:41 AM5/29/10
to Pavel Koshevoy, webm-d...@webmproject.org
Yes but this scenario is only for video files only. For audio-only files mkclean does create Cues for example.

Gaute Strokkenes

unread,
May 30, 2010, 7:58:32 PM5/30/10
to Lachlan Hunt, WebM Discussion
Lachlan Hunt <lachla...@lachy.id.au> writes:

> On 2010-05-28 16:56, Gaute Strokkenes wrote:
>> Does the world really need another container for vorbis-only audio
>> files?
>
> The fact that WebM can be used as another vorbis-only container comes
> for free.

I cannot help but disagree. Sure, from the perspective of an
implementor intending to support WebM, decoding vorbis-only WebM does
come for free. But sometimes, just because you can doesn't mean that
you should.

Vorbis-only WebM is not "free" from the perspective of an implementor
who only wants to create an audio-only client. Nor is it free from the
perspective of users and developers who will have to deal with two
different ways of doing the exact same thing and the resulting
incompatibility and interoperability problems. There will be two
different sets of bugs, quirks, metadata capabilities and so forth.

It is worth noting that there exist Vorbis decoders implemented in
exotic (from the multimedia POV) environments, such as Java and Flash.
Adding support for another demuxer in such environments is substantially
harder since you cannot reuse existing C or C++ code. Wikipedia falls
back to such a method on Internet Explorer, for instance (see
<http://en.wikipedia.org/wiki/Wikipedia:Media_help_(Ogg)>.)

> The need to support audio-only files with a MIME type and file
> extension recognises that authors would inevitably create such files
> anyway, and at least gives a sane way for them to be supported.

Someone, somewhere is bound to eventually create VP8+Vorbis muxed in
AVI. That doesn't mean it should be encouraged. In my view, if a user
accidentally creates vorbis-only WebM it is better that they find out
sooner rather than later about the compatibility issues. More
importantly, muxing software will be less likely to make it some sort of
default or preset.

> Also, AIUI, WebM has a slight advantage over Ogg when seeking through
> the file over HTTP, due to the presence of the Meta Seek information,
> whereas Ogg lacks an index and seeking is a little more complex. When
> seeking, the cues allows the implementation to more accurately find
> out exactly which part of the file to request with an HTTP Range
> request, if the file hasn't completely loaded.

The use case that you mention, which combines progressive download (as
opposed to true live streaming) over an extremely high-latency link with
the requirement for high-precision seeking (as opposed to approximate
seeking) strikes me as rather marginal. In video-land, you are usually
restricted to seeking to keyframes; people seem to get by nonetheless.

Ogg has an advantage of its own in that you can (more-or-less) save a
downloaded live stream to disk and play it back just like any other
file, without having to run it through a special remuxing app to add a
seek table.

--
Gaute Strokkenes

salsaman

unread,
May 30, 2010, 9:28:33 PM5/30/10
to WebM Discussion
Why webm and weba anyway ? Why not mkv and mka ? Mkv has been around for years although generally with divx, xvid or h264 video.

nimesh.c...@accenture.com

unread,
May 31, 2010, 12:16:33 PM5/31/10
to webm-d...@webmproject.org, matthew...@google.com, jkol...@google.com, fgal...@google.com

I’m trying to create an OpenMax component for VP8 to integrate with Android-StageFright. But I am not able to find a parameter structure for configuring VP8 as We have “OMX_VIDEO_PARAM_H263TYPE” ( See Below ) for configuring H.263 params. Is Google working with Khronos group to add a parameter structure type for VP8 ?

 

 

 

4.3.13 OMX_VIDEO_PARAM_H263TYPE

H.263 is a video standard defined by the ITU.  Parameters for this video standard are

controlled using the OMX_VIDEO_PARAM_H263TYPE structure.

OMX_VIDEO_PARAM_H263TYPE is defined as follows.

 

4.3.13.1  Parameters

The parameters for OMX_VIDEO_PARAM_H263TYPE are defined as follows.

•  nSize is the size of the structure in bytes.  This value shall be specified when this

structure is used as either an input to or output from a function.

•  nVersion is the version of the structure.

typedef struct OMX_VIDEO_PARAM_H263TYPE {

    OMX_U32 nSize;

    OMX_VERSIONTYPE nVersion;

    OMX_U32 nPortIndex;

    OMX_U32 nPFrames;

    OMX_U32 nBFrames;

    OMX_VIDEO_H263PROFILETYPE eProfile;

    OMX_VIDEO_H263LEVELTYPE eLevel;

    OMX_BOOL bPLUSPTYPEAllowed;

    OMX_U32 nAllowedPictureTypes;

    OMX_BOOL bForceRoundingTypeToZero;

    OMX_U32 nPictureHeaderRepetition;

    OMX_U32 nGOBHeaderInterval;

} OMX_VIDEO_PARAM_H263TYPE;

This message is for the designated recipient only and may contain privileged, proprietary, or otherwise private information. If you have received it in error, please notify the sender immediately and delete the original. Any other use of the email by you is prohibited.

Lou Quillio

unread,
May 31, 2010, 12:22:48 PM5/31/10
to salsaman, WebM Discussion, John Luther
+jluther

On Sun, May 30, 2010 at 9:28 PM, salsaman <sals...@gmail.com> wrote:
> Why webm and weba anyway ? Why not mkv and mka ? Mkv has been around for
> years although generally with divx, xvid or h264 video.

http://www.webmproject.org/code/specs/container/#webm_guidelines

Because WebM is a Matroska subset, it has fewer allowed stream types
and features than Matroska.

A distinct nomenclature gives us a way to easily refer to the subset,
and gives users a way to associate a piece of content with a client
having known-good support. Google Chrome, for example, may or may not
be able to decode filename.mkv, but it will absolutely play
filename.webm -- provided filename.webm conforms to the WebM
guidelines.

* * *

jluther: Shall I add this as a FAQ item? Probably here:

http://www.webmproject.org/about/faq/#webm_video_file_format

LQ


--
Lou Quillio
Tech Writer
WebMProject.org
+1 518.285.0003   <= Mobile (gV)
+1 518.881.4256   <= Office

John Luther

unread,
Jun 13, 2010, 8:51:49 AM6/13/10
to WebM Discussion, Philip Jägenstedt, Andy Berkheimer, Dean Lee, skylinestudios
It turns out that weba is already trademarked for media-related
products in many countries. Are there any other ideas? I saw that
someone suggested audio/webm and .webma.

--JL

Pavel Koshevoy

unread,
Jun 13, 2010, 11:19:02 AM6/13/10
to webm-d...@webmproject.org

On 6/13/10 6:51 AM, John Luther wrote:
> It turns out that weba is already trademarked for media-related
> products in many countries. Are there any other ideas? I saw that
> someone suggested audio/webm and .webma.
>
> --JL

How about .wmka, or .mkaw?

Pavel

Silvia Pfeiffer

unread,
Jun 13, 2010, 4:29:36 PM6/13/10
to John Luther, WebM Discussion, Philip Jägenstedt, Andy Berkheimer, Dean Lee, skylinestudios
Have you tried "webs" for web sound file?

It would be nice to somehow stay in the same naming scheme as .webm
but it may be hard...

Silvia.

Lachlan Hunt

unread,
Jun 14, 2010, 4:48:17 AM6/14/10
to John Luther, WebM Discussion, Andy Berkheimer, Dean Lee, skylinestudios, Philip Jägenstedt
On 2010-06-13 14:51, John Luther wrote:
> It turns out that weba is already trademarked for media-related
> products in many countries. Are there any other ideas? I saw that
> someone suggested audio/webm and .webma.

Does this really matter? "weba" is just the abbreviated file extension
for WebM Audio.

"Court Says File Extensions Not Eligible For Trademarks... Sorta"
http://www.techdirt.com/articles/20100518/2343349484.shtml

John Luther

unread,
Jun 14, 2010, 10:04:06 AM6/14/10
to Lachlan Hunt, WebM Discussion, Andy Berkheimer, Dean Lee, skylinestudios, Philip Jägenstedt
Good point, but IANAL, I will ask.

--JL
Reply all
Reply to author
Forward
0 new messages