Developers › A Journal from Behind the Trenches
Google removes H.264 support from Chrome
In news that's almost too good to be true (still waiting for a retraction, but hoping that it won't come), Google just announced that they will remove support for the patent-encumbered H.264 codec from Chrome:
With Google's powerful and well-oiled advertising machinery backing it and ensuring its growth, Chrome will now contribute to a true open web. Along with Opera and Firefox, we may soon find that the majority of the browser market supports open formats like WebM and Theora, while H.264 supporting browsers will make up a smaller and smaller part of the market.
Great move, Google. Mad props, as they say.
There are already comments in their blog criticizing the decision, but these people do not seem to realize what great news this is for the open web. After all, the web needs to based on open standards, not patent-encumbered technologies. H.264 was threatening to hold back parts of the web like Internet Explorer did before it.Though H.264 plays an important role in video, as our goal is to enable open innovation, support for the codec will be removed and our resources directed towards completely open codec technologies.
With Google's powerful and well-oiled advertising machinery backing it and ensuring its growth, Chrome will now contribute to a true open web. Along with Opera and Firefox, we may soon find that the majority of the browser market supports open formats like WebM and Theora, while H.264 supporting browsers will make up a smaller and smaller part of the market.
Great move, Google. Mad props, as they say.
Readers' comments 40
Left under the original article, as archived.
About this article
| Published in | A Journal from Behind the Trenches |
|---|---|
| Date | January 11, 2011 |
| Author | Haavard |
| Tags | open web, google, chrome, h264, webm, open standards |
| Original | Internet Archive |
Martin KadlecTuesday, January 11, 2011 10:08:11 PM
Joonas LehtolahtiTuesday, January 11, 2011 10:16:30 PM
DillonTuesday, January 11, 2011 10:38:28 PM
ChrisTuesday, January 11, 2011 10:43:14 PM
Originally posted by Astrophizz:
Maybe you should get the facts right. Here's a quote you missed about what Google is talking about:"We expect even more rapid innovation in the web media platform in the coming year and are focusing our investments in those technologies that are developed and licensed based on open web principles. "We expect even more rapid innovation in the web media platform in the coming year and are focusing our investments in those technologies that are developed and licensed based on open web principles."
As you probably know, the W3C states that open web standards must be free of patent claims. H264 does not match that requirement, so it cannot be based on open web principles by definition.
DillonTuesday, January 11, 2011 10:46:15 PM
ChrisTuesday, January 11, 2011 10:49:36 PM
DillonTuesday, January 11, 2011 10:56:15 PM
Anyways I hope Google will improve vp8 more before they finalize this. For on the encoding toolkit needs a lot of improvement for one and it's currently slower to encode than x264. Also as it stands the without psychovisual optimizations vp8's quality isn't that great, even compared to theora.
seriousTuesday, January 11, 2011 11:14:05 PM
Artur „Jurgi” JurgawkaTuesday, January 11, 2011 11:28:20 PM
Daniel HendrycksTuesday, January 11, 2011 11:34:57 PM
d4rkn1ghtTuesday, January 11, 2011 11:42:41 PM
ChristopherTuesday, January 11, 2011 11:47:41 PM
WebRiderTuesday, January 11, 2011 11:52:28 PM
btw think about ARM devices with h/w H.264
Cutting SpoonWednesday, January 12, 2011 12:19:07 AM
But this means Linux browser users have the best format support?
Daniel HendrycksWednesday, January 12, 2011 12:49:19 AM
Originally posted by Cutting Spoon:
WebP can already be: http://badassjs.com/post/1313384987/weppy
João EirasWednesday, January 12, 2011 1:01:09 AM
Originally posted by Dillon:
The h.264 format is not open. It's not available under RAND terms and the W3C has a strict royalty free policy for acceptable standards.
According to the h.264 licensing terms, even the end using watching a youtube video must pay license fees, although the mpeg-la has made an exception for the case of watching video on the web, for the sake of achieving market-share online, so if it got a critical mass and everyone would be dependent on it, the mpeg-la would start relicensing h.264 again and collect royalties. Although vp8/webm is not as good as h.264, it's good enough as a free codec, and the web needs free formats not owned by big patent trolls.
DillonWednesday, January 12, 2011 8:10:38 AM
ChrisWednesday, January 12, 2011 9:27:36 AM
Originally posted by Astrophizz:
See my quote from Google's blog again. They did not make a blanket statement. You misinterpreted their statement, and didn't see the part of the text where they clarified what kind of "open" they were talking about.Originally posted by webrider:
What innovations are you referring to, and in what way are they relevant to Google's removal of h264 from Chrome?WebRiderWednesday, January 12, 2011 9:46:47 AM
It's pretty obvious that Google harms HTML5 video initiative this way.
ChrisWednesday, January 12, 2011 12:44:48 PM
And I don't think your claim is obvious at all. In fact, I would argue that h264 was a threat to the web, since the web needs to be based on open web standards. That means no patents blocking it.
netwolfWednesday, January 12, 2011 1:01:52 PM
ChirpieWednesday, January 12, 2011 1:09:44 PM
Originally posted by Astrophizz:
H.264 is not an open standard, and will never be an open standard as long as it's patent-encumbered. That blanket statement is perfectly valid. An open standard must meet certain requirements, and h.264 fails the IPR/patent requirement:http://www.robweir.com/blog/2010/09/recipe-for-open-standards.html
FransWednesday, January 12, 2011 4:41:36 PM
Mad Scientist (عادل)Wednesday, January 12, 2011 6:42:23 PM
Knut Remi "DrLaunch" LøvliWednesday, January 12, 2011 7:44:09 PM
Originally posted by webrider:
One of the intentions of HTML5 video from the beginning was to use an open video codec as part of the browser, allowing for a royalty-free implementation of video, independent of any corporations, in a browser on any platform.
During the initial work on HTML5 video, it was decided to use (OGG) Theora as it was the only open format available for implementation at that moment. But as WebM has gone open recently it's not harmful to use WebM in addition to Theora in HTML5 video now.
I respectfully disagree on your statement that Google harms HTML5 video this way. Removing support for H.264 only empowers HTML5 video as video services will be required to provide videos in open formats in Opera, Chrome and Firefox. Requiring the use of open formats means HTML5 video can be supported by more developers and on more platforms.
On another note, great news!
WebRiderThursday, January 13, 2011 12:53:56 AM
Originally posted by drlaunch:
HTML5 describes the structure of web content, not data compression algorithms.
Originally posted by drlaunch:
Browser should use system multimedia framework.
I do not want to take part in these dumb corporate wars.
>>> ODIN Blog
ChirpieThursday, January 13, 2011 9:30:51 AM
Originally posted by webrider:
HTML5 was originally supposed to define a baseline coded, but then companies like Apple threatened to sabotage the whole thing.Browsers should browse the web, and the web must use open standards. H.264 is not an open standard.
DillonThursday, January 13, 2011 8:50:03 PM
Originally posted by xErath:
OK I looked into it and h.264 IS in fact licensed under RAND terms in accordance with ITU and ISO standards. Specifically ITU-T as the h.264 standard and ISO as the ISO/IEC 14496-10 MPEG-4 Part 10 Advanced Video Coding standard. I reread what you wrote and it is FUD to say that mpegla would just reinstate the royalties. There's no evidence to support it. As far as anyone is concerned, h.264 on the web is entirely free. The issue Opera has is with paying royalties to include a decoder.
Here's an interesting article: http://arstechnica.com/web/news/2011/01/googles-dropping-h264-from-chrome-a-step-backward-for-openness.ars
I'm not sure it's entirely right but it's worth reading I think.
It's poignant to note that vp8 hasn't been submitted to any standards body like ISO or ITU, whereas both jpg and png (as examples) have. Hopefully Google is just waiting to fix the bugs inherent in the way vp8 was originally written by On2 before submitting it to a standards committee or releasing control of it. Of course fixing issues and finalizing the implementation is something a standards board could help them with. As such, it doesn't seem like Google intends to submit vp8 to the ISO or ITU. It's possible that they want to maintain control over the codec - which I think would be bad for the the codec and for internet video if webm is to become popular on the web.
I still don't see why Opera doesn't allow users to use system codecs (well except on linux). Then they wouldn't have to worry about paying licenses for decoding software and the video tag could become more like the img tag in allowing various types of media to be used with it. For me personally, h.264 is the better option as it saves bandwidth being more efficient than webm and is the codec I use everywhere else on my computer for video. I envision that Microsoft's h.264 decoding plugin is going to become pretty popular for people who feel the same way (and want to save bandwidth on video), or who want a control over what their browser can do/decode.
ChrisThursday, January 13, 2011 11:33:26 PM
Originally posted by Astrophizz:
There is ample evidence that once someone has established commercial dominance (monopoly), they will tighten their grip to make more money, unless something prevents them from doing so. And the kindness of their hearts is unlikely to do that.H264 is patent-encumbered, and therefore a closed standard.
Originally posted by Astrophizz:
The article is mostly a red herring, and doesn't address what it claims to address: http://my.opera.com/haavard/blog/2011/01/13/opennessWebM is owned by the WebM Project, an open-source project sponsored by Google and others. They have also issued an irrevocable license for free use and distribution of WebM. It is in Google's interest to get as many eyes as possible on the web as well. On the other hand, it is in the MPEG-LA's interest to do a bait and switch to maximize license revenue.
The plugin won't do because it won't integrate with html, css and other web standards. It'll be completely separate from the rest of the browser.
FransFriday, January 14, 2011 12:37:36 AM
Originally posted by Dillon:
I'll notify those with more than 100,000 subscribers. They are part of anyone, right?
Knut Remi "DrLaunch" LøvliFriday, January 14, 2011 5:42:54 PM
Originally posted by webrider:
I made no such statement. I did however make a statement about how HTML5 originally specified which codec was supposed be used in the browser. My statement stands.Originally posted by webrider:
Yeah on Windows that basically means Windows Media Player. No thanks. Then I'd rather go back to using media plugins.Originally posted by webrider:
SourceBasically this means H.264 is only supported on devices where it's built in. By default, the desktop browser doesn't support H.264 except on *nix systems.
WebRiderFriday, January 14, 2011 6:48:19 PM
Originally posted by drlaunch:
http://en.wikipedia.org/wiki/Media_Foundation
http://en.wikipedia.org/wiki/GStreamer
Originally posted by drlaunch:
for example, Windows 7 powered pc...
DillonFriday, January 14, 2011 8:10:41 PM
Originally posted by drlaunch:
Windows 7 has h.264 built in, but Opera doesn't support h.264 on Windows 7.
Edit: oops didn't notice I got beat to saying that.
Knut Remi "DrLaunch" LøvliSaturday, January 15, 2011 11:13:42 AM
Originally posted by webrider:
Originally posted by drlaunch:
It was made perfectly clear which platforms had H.264 in earlier posts. You're only hindering the discussion by reading other people's posts wrong, and misquoting them.FransSaturday, January 15, 2011 11:00:50 PM
Originally posted by WebRider:
Media Foundation is exactly the kind of thing that was referred to ("No thanks"); GStreamer is what Opera already uses and, at least on Windows, that's certainly not the "system media framework." Not sure what your point is.
WebRiderSaturday, January 15, 2011 11:15:35 PM
Originally posted by Frenzie:
any technical reasons?
> GStreamer is what Opera already uses
yea, non configurable, not integrated with system updater. very few formats, jerky playback, no CUDA support... It Just Works
well nvm
seriousSunday, January 16, 2011 10:30:14 AM
* You have to maintain multiple backends instead of just one - which also means you have to introduce an additional abstraction layer, which generally is associated with some overhead. Maybe this is small (eg. if you just have to hand a file pointer and the target drawing surface to the framework), but it could also be bigger (eg. if the output format of the framework needs to be converted in some way before displaying it)
ChrisSunday, January 16, 2011 2:57:06 PM
Originally posted by webrider:
Cross-platform?FransSunday, January 16, 2011 8:54:53 PM
Originally posted by WebRider:
Well, that was already answered, but I want to add a bit to that.
Originally posted by serious:
And even more backends so things can work on older versions of Windows as well.
Originally posted by WebRider:
Opera already has a built-in updater to work around the shortcomings of Windows, and besides, if that's such a big problem then IE9 (and Microsoft software in general) is the only viable software one can use on Windows. Aside from jerky playback I don't see an issue with any of the things you mentioned. I think Wikipedia is the perfect (but certainly not only) use case for VIDEO and AUDIO elements, and for any serious video viewing I'll utilize a dedicated player. That doesn't mean the occasional HD video shouldn't play smoothly (and for me, it plays smoothly), but it's just a browser. (PS I bet IE9's video playing capabilities are at best as good as WMP, which is not very.)
Cutting SpoonMonday, January 17, 2011 7:24:33 PM
Originally posted by DanielHendrycks:
Well yeah, I have the userjs to automatically make the conversion on any page.
Also, I believe the Opera Devices SDK is capable of h.264 playback...? Though apparently the licensing burden falls upon the licensee that chooses to use it? Some sort of compromise for Opera-powered devices that decode HDTV streams and online videos.