Previous Entry Add to Memories Share Next Entry
xiphmont

Xiph.Org press release about WebM

Hooray! The press release:

Xiph.Org announces support for the WebM open media project

The Xiph.Org Foundation is pleased to announce its support of the WebM open media project as a project launch partner. As announced earlier today at the Google I/O Developer Conference, the WebM format combines the VP8 video codec, the Matroska container, and the Vorbis audio codec developed by Xiph into a high-quality, open, unencumbered format for video delivery on the Web. Xiph will continue to contribute to WebM as a whole and collaborate in its further development and deployment.

Success within the Open Source community is vital to the larger success of WebM. Community adoption of a newly opened source base, such as VP8, is traditionally fraught with peril. Xiph was the primary organization to develop and promote the earlier VP3 codec when open sourced by On2, and we know that substantial work lies ahead of us to make WebM a success. Toward that goal, we look forward to working with the established community projects also contributing to WebM including Matroska, ffmpeg, GStreamer, and Mozilla, as well as open-source oriented business leaders such as Google, Opera, Red Hat and others.

Kudos to all for the progress we've made, now we've got to get back to work.

Monty
Xiph.Org

The Xiph.Org Foundation is a not-for-profit corporation dedicated to open, unencumbered multimedia technology. Xiph's formats and software levels the playing field for digital media so that all producers and artists can distribute their work for minimal cost, without restriction, regardless of affiliation. May contain traces of nuts or poorly contained awesome.

Tags: ,

I'm not a video specialist, but I don't think Matroska Container a good choice for video streaming.
Will the WebM project do some hacks to Matroska Container? Shouldn't Ogg be a better format for web video streaming after some work on it?

And, sorry by my poor english.
(Frozen) (Thread)

(Deleted comment)
I only want to know one thing from Matroska.
Does Matroska support http://en.wikipedia.org/wiki/Interleaving ?

And sorry again by my bad english.
(Frozen) (Parent) (Thread)

Re: Matroska Containter

xiphmont

2010-05-20 05:16 pm (UTC)

Yes. Only early revisions did not have interleaving. That was a long time ago.
(Frozen) (Parent) (Thread)

Thanks. Then, Matroska is a good format.
Just waiting for CELT and Dirac now.
(Frozen) (Parent) (Thread)

Re: Matroska Containter

xiphmont

2010-05-20 01:36 pm (UTC)

I think the Matroska format actually is good for streaming, even if that's most often not what it's being used for. Software implementations that only do local file playback will just have to catch up (It's not like every Ogg player deals with streaming properly either). I'm biased and I do think Ogg would have been a better technical choice for WebM, but that doesn't mean Matroska is wrong or bad. It's what Google chose, and as the maker of a competing container, I say it was a good choice.

There's one way in which Matroska is a much better choice than Ogg; in the current climate with one particularly outspoken open source group effectively at war with Ogg, choosing Matroska sidesteps that whole political issue. Everything else being equal, Matroska was a much better choice in this situation, and we're fully behind that. We'll not be dropping Ogg, and I'll be introducing the transOgg spec soon I hope, but we'll be using Matroska ourselves in WebM work. It meets all the necessary criteria with flying colors.

Edited at 2010-05-20 01:37 pm (UTC)
(Frozen) (Parent) (Thread)

Err ... What is transOgg? Is there any page at xiphwiki about this?
(Frozen) (Parent) (Thread)

In the interest of transparency

xiphmont

2010-05-20 01:27 pm (UTC)

Screening now enabled. As usual, this is my blog: Anything that's obvious trolling with no content and no value gets axed at my whim. At least be witty. Jeez.
(Frozen) (Thread)

Re: In the interest of transparency

(Anonymous)

2010-05-20 03:32 pm (UTC)

monty can we have an ETA on the aotuv merge it would be really beneficial as vorbis is an important ingredient of webm.
Thanks for all the effort you guys put into this
(Frozen) (Parent) (Thread)

Re: In the interest of transparency

xiphmont

2010-05-20 06:28 pm (UTC)

None right now, but eventually. A merge is not as simple as running patch.

Edited at 2010-05-20 06:29 pm (UTC)
(Frozen) (Parent) (Thread)

Re: In the interest of transparency

(Anonymous)

2010-05-20 08:06 pm (UTC)

On the topic of Vorbis, is there going to be any big advocacy push for this as the audio format for HTML5?

Wikipedia seem to be leading the charge on this at the moment, but since Vorbis was already playable in Chrome, Firefox and Opera, and it's going to be native in Flash as part of WebM, and in IE9 and Safari as part of WebM plugins, plus there are existing fallbacks to silverlight, java and older Flash versions via an Actionscript based implementation it seems little stands in its way with some appropriately structured javascript to manage the fallback process.

People will need to provide Vorbis anyway for Opera and Mozilla users, but it would be good if the tiny sliver of portable Apple devices were seen to be the ones that you had to do the extra work for and that you maybe didn't bother with if you weren't directly targeting them.

Perhaps even a push to officially bless it in HTML5 if that wouldn't be seen as inflammatory by Microsoft and Apple and prejudicial to WebM in it's entirety being added to the spec.

It would also be tragic if Adobe and Microsoft implemented this in a way that only worked with video. I'm hoping and assuming that audio tag support is planned for in Google's masterplan too, but there's been little clarity on that point.

(A different anonymous)
(Frozen) (Parent) (Thread)

Do you plan on implementing WebM in Fedora 14?
(Frozen) (Thread)

Ok... if Xiph.Org support it the I support it :D

Let's make WebM the most widely and popular choice for HTML5
(Frozen) (Thread)

We can't disagree.
(Frozen) (Parent) (Thread)

Reuse code

(Anonymous)

2010-05-21 09:09 am (UTC)

Hello. Exciting times!

I was wondering, given that VP8 was made by the same guys who wrote VP3 and so on, does it somehow share a common arquitecture that would make it possible to implement some of the improvements already seen in Theora (such as the superb activity masking, for instance). Or you really need to do everything from scratch?

Cheers
(Frozen) (Thread)

Re: Reuse code

(Anonymous)

2010-05-27 12:56 am (UTC)

I had a similar questions which came to mind.
Will there be any code or approaches to encoding which can be drawn from VP8?

Perhaps more on my mind is, given that someday there will likely be a VP9 (perhaps not that exact name), will code or encoding approaches be able to be easily drawn from Theora and used in VP9?
(Frozen) (Parent) (Thread)

Somewhat; there are certainly basic techniques that can be applied to VP8. The code will have nothing whatsoever in common with Theora; Theora itself no longer has a single line of code in common with the original VP3.

There will of course be new codecs past VP8 (Tim's been planning for a long time, just like I've been planning for Ghost for a long time). It's unlikely any code will survive in the end, though we generall favor an incremental approach that swaps things out for new techniques piece by piece.
(Frozen) (Parent) (Thread)

The future of OGG

(Anonymous)

2010-05-22 01:09 pm (UTC)

Hi!

As first I would like to thank you for the excellent work which you and other people at Xiph.org have done over the last years. I think you should be proud of the fact that your Vorbis codec was choosen for the only supported audio codec of the new WebM format.

In one of the previous posts you have mentioned that you are working on the successor of OGG (called transOgg). But why don't you inform us about the current state of this project? I think there are many interested people who may send their ideas which might be useful.
(Frozen) (Thread)

Interesting news!
(Frozen) (Thread)

You are viewing xiphmont