Google ups the ante. H.264, it’s on. Of course, given the entrenchment of existing videos and the Flash plug-in, many are predicting the “die tryin’” scenario is the more likely one. Time will tell. Photo (CC-BY) Linus Bohman.

Today, Google announced it is omitting H.264 support from its Chrome browser, in favor of free and patent-unencumbered VP8 (via the WebM container) and OGG Theora codecs. Simply put, it’s the biggest victory the open video camp has gotten in a landscape that has largely seemed tilted against them. The long-term outcome is, fairly, anyone’s guess. But you can at least mark one in the column of open video advocates, if you’re keeping score.

As I’ve covered on this site ad infinitum, H.264 simply isn’t an option for free and open source software. It’s a proprietary format backed by aggressive patent enforcement – something that is entirely within the legal rights of the parties that own the intellectual property on which it’s based, but which restricts free software projects like Mozilla and kdenlive.

Unlike a plug-in, now that HTML5 includes a video tag, free codecs and containers are essential to the future of video on the Web. Without them, browsers and creation tools – the entire workflow by which people share media – can’t work in a free software ecosystem. Because Apple has chosen to support only H.264 in its influential mobile devices and desktop browser, and because the same mobile devices don’t support Flash, the net effect has been to force video content not supported by Flash to H.264.

Understandably, web designers and content creators care mainly about getting their content visible, not about obscure legal wrangling. (No one can fault them for that.) But that’s the reason that today’s announcement from Google is such big news. Word of the intricacies of patents or licensing schemes don’t really impact Web content. But seeing Google throw its weight entirely behind the free VP8 and Theora codecs in their fast-growing Web browser could indeed motivate change.

No good deed goes unpunished: many publishers and users are criticizing the deal (see incensed commenters on Google’s blog), and Apple-biased writers – who I believe have artificially turned this into a question of pro- or anti-Apple positions – aren’t helping. But then again, I’m beginning to think that you have to tune out these kinds of complaints. Users in general seem to want the benefits of free and open source software like Firefox without any associated cost or sacrifice. And that’s, frankly, to be expected. It’ll be up to people in the free software community to deliver the kinds of benefits in the long run that they expect, at which point these gripes will seem a distant memory.

To put it another way, users may be vastly underestimating Google’s power to make VP8 a hit.

Can Google single-handedly shift video content usage? Well, let’s see: they’ve got the world’s most popular search engine, the world’s fastest-growing desktop browser (Chrome/Chromium), the world’s fastest-growing mobile browser (Android), the world’s fastest-growing mobile platform (Android), and the world’s most popular video site (YouTube). That’s to say nothing of emerging plays in TV and tablets and netbooks. Nor is Google alone. The world’s largest open source browser, Firefox, by necessity supports only open formats and backs Google’s WebM and VP8. You’ll also find commitments from everyone from NVIDIA and ARM to Skype and Winamp (seriously). Now, in fairness, industry partnerships come and go, and people put their logos on all kinds of things. But Chrome, Android, YouTube, and Firefox ought to count for something.

Whatever this means for Chrome, or even HTML5 video, this is a huge victory for Mozilla’s Firefox, in that it no longer has to stand alone. Photo (CC-BY) Channy Yun.

As usual, you can count on John Gruber (daringfireball) to root against Google, against Flash, against the open video formats, and for H.264.

The big problem WebM has versus H.264 is that there are hardware decoders for H.264.

This I agree with in part, but the situation is changing. That’s why Google has been pushing for hardware adoption and solutions for native-side decoding. It’s also important to note that mobile processors are getting faster and more capable.

I think everywhere else, H.264 will continue to dominate, and instead of getting native playback, Chrome users will get playback through Flash. This should be great for Chrome OS laptop battery life.

Burn! But for starters, let’s evaluate what “native” means. Yes, Flash decodes H.264 as a plug-in. But that’s already the reality for an end user on, say, any desktop Firefox browser watching H.264-encoded videos. And underneath the hood, Flash is using what is technically native code (including, very often, GPU acceleration) to decode and display video, just as would a plug-in. Gruber is conflating a whole bunch of variables to lay blame for battery life issues at the feet of Flash H.264 decoding.

Anyway, if anyone thinks the point of this is to steer publishers away from H.264, that’s exactly the point; Gruber can disagree, but he surely can’t feign shock. Given Microsoft has been relatively neutral, the only company I can see aggressively advocating H.264 support right now is Apple. On the other side of the fence, you have Google, Mozilla, major hardware makers, and more or less the entire free software community.

There’s no way publishers can drop H.264. To support Chrome, they’d have to add WebM-encoded versions of each video.

Gruber’s entire argument is predicated on the idea that most video content is currently encoded in H.264, and thus it shall be for the foreseeable future. But that ignores an underlying reality: most publishers already rely on a third-party to deal with the task of video hosting. That’s especially true on mobile – many average consumers will be just fine viewing YouTube videos and nothing else. But even beyond YouTube, video transcoding, hosting, and delivery infrastructure is sufficiently challenging that even major publishers often turn to providers like Brightcove.

If you’re self-publishing, changing video formats is a big deal. But if you rely on a video third party to deal with complexities of video publishing, you already depend on them to solve these problems for you. In fact, part of the reason you pay them is to deal with eventualities like, say, a proxy war spiced with patent disputes between two tech giants over obscure codec issues. In other words, those third-party publishers – like YouTube – could well add a new format, and quickly. (I’d say if you’re Brightcove, you might take this as great news.)

Gruber’s assumption that what you see today is what you’ll see this time next year is therefore a big leap.

Then, Gruber questions Google’s sincerity – and seizes an opportunity to complain again about Flash:

If Google is dropping support for H.264 because their “goal is to enable open innovation”, why don’t they also drop support for closed plugins like Flash Player?

Because that wouldn’t make any sense? Codec support has to be baked directly into the browser. Plug-in support is a facility that lets you run whatever you want. In the fully open source Chromium branch of the browser, H.264 is already missing. Whatever Google’s motivations, they’re being explicit about their goal: they want to steer people to open codecs and away from closed codecs.

As it stands now, Chrome not only supports Flash, it ships with its own embedded copy of Flash.

This statement is misleading. Chromium doesn’t ship with Flash, and never shipped with H.264, for precisely this reason.

I don’t see how Google keeps Flash but drops H.264 in the name of “openness” without being seen as utter hypocrites.

You can see Google however you want, but it’s not hard to see why people would see this particular battle as an unadulterated win for openness, no strings attached.

Flash adoption is already a fact of life. And it represents a plug-in, which is different from content that is delivered directly through HTML tags inside a web page. Google’s stated goal is driving adoption of open codecs. They wouldn’t make any headway doing that by disabling Flash – as Gruber himself says, Flash is likely to be the fallback for content that isn’t yet ready for the new codecs. Without that callback, people wouldn’t use Chrome or other tools at all. And there’s too much legacy content to abandon Flash.

The question of Flash’s place on the Web was decided years ago. With HTML5 and the video tag, the Web has a chance to do things differently – a chance that won’t soon come again. So, yes, Google continues to support Flash. But by being aggressive on video, they’re betting that another question – codec support in HTML5 – may not yet be set in stone.

Should you trust everything Google says, and take every intention at face value? Of course not. They’re huge company. You’d have to be insane.

Is this a huge win for openness? Absolutely.

This can no longer be viewed as Flash versus H.264, Adobe versus Apple. It has clearly become VP8 versus H.264, Google versus Apple, and free software versus Apple. As with pro sports, though, I wouldn’t worry about taking sides only to watch your favorite player switch jerseys. The battle here really is open versus closed. That’s been overstated elsewhere (iOS versus Android, for one), but here’s it’s strikingly clear-cut.

Most of all, it’s a huge win for Firefox. Mozilla stood alone with a browser that supported only open codecs and not H.264. But with Google in their corner – and Android, Chrome OS, the Chrome browser, and YouTube with them – that’s another story. The phrase “game changer” is wildly overused today to the point of being a cliche. But I think this may actually qualify. Google’s open source release of VP8 and WebM was a big gift to the open video camp, but it lacked teeth to drive adoption.

With this announcement, that changes. Open video just got a posse, in a very big way.

References:
Google Dropping Support for H.264 in Chrome [Daring Fireball]

HTML Video Codec Support in Chrome [Chromium Blog]

  • http://www.danreetz.com Daniel Reetz

    It is indeed a big win for openness, and might actually cause me to use Chrome for my full-time browser. (the free holiday wifi on flights was nearly bribe enough…)

    Love the nuanced, critical articles you've been writing lately (the VLC one in particular). Please, more like these.

  • http://Wileywiggins.com Wiley

    Vp8.

    Looks.

    Like.

    Shit.

  • CARMAI

    Big win for openness? Maybe, but Consumer experience is the Bottom Line here. How your video performs in your device will decide what codec gets popular and widely adopted. The consumer experience cannot take steps back for the sake of openness. Overall better performance, easy of use, features sets, battery life, etc, on your device will greatly influence codec popularity. The codec the offers the best performance in the vast majority of devices will.

    I own 2 PCs, 3 Macs, an iPhone, a PSP, a PlayStation 3, an Xbox 360, and guest what all of them have in common: they all play H.264. 

    The regular non technical consumer will just do the same thing they always do, the simplest and most logical thing, switch browsers, so they can see a video that petty much every other device can decode. 

  • Peter Kirn

    @CARMAI: I'm regularly playing WebM + VP8 videos on YouTube in aforementioned Chrome, and my (fairly modest) computers here don't melt down. Ease of use? Ease of use is being able to play the videos. YouTube substitutes these videos seamlessly and there's not anything noticeable to an end user.

    There are indeed quality differences in encoding VP8 versus H.264, but this stuff is subtle, not the sorts of stuff that's going to matter to consumer experience.

    There are far more variables here than just codec choice. Your PCs and Macs are capable of playing VP8 (and Theora), and with the proper encoding, the content can look good. (The game consoles, sorry, aren't something publishers are going to target. The iPhone, of course, is another story.)

    The battery life stuff, meanwhile, dangerously approaches FUD. I agree with what you're saying, but the real question here is whether Google can get their hardware partners to deliver on the performance issues. I can say this – there's a lot of progress on both the hardware and software sides that the H.264 apologists are ignoring, and Google is making a big bet here. Now, whether that actually plays out remains to be seen – I just wouldn't necessarily write off VP8 completely because of bad past experience.

  • Peter Kirn

    Just FYI, you can test the WebM + VP8 experience on YouTube at youtube.com/html5 via a recent Firefox 4 beta. I find performance on my machines far exceeds Flash, which is otherwise the choice for FF, and that quality matches YouTube's H.264 encoding.

    Oddly, part of what Gruber argues is that Google isn't motivated entirely by "openness."

    But I'd *hope* in this case they aren't only motivated by that. I'd hope the publisher of the world's largest online video site would be doing something that makes sense for them and not shooting themselves in the foot. I think in this case, it's safe to assume they believe the technology works for them as a publisher.

  • http://www.blairneal.com Blair neal

    I guess one of my issues with it is that I don't come across it very often when exporting from Final Cut/Compressor or Premiere…of course I haven't really been looking..but aside from this blog, this is really the only place I hear about VP8. It seems like there needs to be quite a publicity push for it to catch on…but I could've just been living under a rock somehow. Not saying I'm against the open format…just seemed like an abrupt move.

  • Peter Kirn

    @Blair: No, you're right. I think you should be able to expect WebM export in Adobe products soon. I expect Apple is avoiding it for political (and perceived legal) reasons.

    Generally, what's happened so far is we've seen codec support, but not necessarily the full-blown content creator experience where it's an export option (outside free software).

    But it's early days yet. Brightcove believes that WebM will overtake H.264 – and this is way back in May of last year.
    http://gigaom.com/video/brightcove-president-webm

    I think there's a clear chicken and egg issue, so Google decided to be the chicken.

    I would like to see this happen in those kinds of direct export filters, though, not just server-side transcoding as is happening now on YouTube.

  • http://sunset2sunrise.net jesse be

    it's definitely an emotional topic for people, which is understandable as these descisions can change who is making money.

    somebody or nobody, lmao.

    but the question of who we pick should come down to how each of us want it to work.

    google = paid for by advertising.

    apple = paid for by consumers.

    if it was your content, video, music, software, what would you prefer?

    I personally love giving things away, but I don't like invasive ads to be a part of it.

    it also begs the question, does this mean no more YouTube videos on the iPhone, they have to be h.264 to play, it's still true the devices that use hardware have significantly better battery life.

    it's all gravy to talk about the future like it's already the present.   but I'd rather have better battery life and no ads.   

    that's just my 2cents

  • http://xdissent.com Greg Thornton

    I never voted for Google as "declarer of openness." Since when does *cutting* support for software from your platform mean you care more about the freedom of your users?

  • chris thorpe

    I don't actually care what Codec is used, but I'm wary of Google. VP8 is assumed to be patent-unencumbered because nobody has sued yet. Once it starts to be used, trolls will start suing – and they'll start with smaller companies who can ill affford lawsuits, not Google. Google may or may not offer support. We'll see. The other thing that concerns me is – follow the money. Google is in business. At some point with all their initiatives, they either monetise them, or unceremoniously drop them (as Android device makers will find out at some point). All of which is a roundabout way of saying H.264 is evil today, and VP8 will be evil at some point. Ideally we'd use neither. 

  • humanbulk

    Sorry for Chrome, I will keep in my Safari. VP8 is the worst quality video I ever seen, plus it has no very good color correction.

    But tell the people the truth please… H264 is free to use as a new license comes out. So shut up!

  • Ian Davies

    The whole notion of Google as this big open-source hero is utter crap. Google support whatever they think will give them the biggest business advantage and gain them the biggest advertising revenues.

    That's not necessarily a bad thing, but it's a non-sense to pretend that they are somehow driven by benevolence to the OSS community or anyone else.

    Also, the way that VP8/WebM is constantly framed as "unencumbered" is disingenuous at best. The reality is that its patent status hasn't been legally challenged. Yet.

    Expect that situation to change very quickly.

  • humanbulk

    Anyway following the same rules they must take out also Flash because is not open source… this is political and commercial fight no in the name of the user. Google is become the evil big brother and we are let em do all he wants.

  • Steve Elbows

    WebM didnt gain much traction in 2010, so it surely needed something like this in order to stand a chance. Hopefully any transition will happen in a way that isnt too painful for creators or consumers. Flash will likely gain new legs from this stuff as it continues to have a place in bridging the gap between browser codec support variations.

    From a purely technical point of view, WebM has yet to impress me. And yes, flash has done much in recent years to decode H.264 efficiently so while questions about hardware decoding remain relevant, it hopefully wont be too much of a problem as things pan out.

  • RichardL

    This is huge. It was inconceivable a few years ago. 

    Now let's see some love for OGG Vorbis as an audio format. 

  • Peter Kirn

    @humanbulk: The "free" H.264 license applies to end users, not the decoders and encoders used in software. That's a far cry from being able to distribute code as Firefox needs to do. The reason Flash isn't an issue is that it's an independent plug-in, not an integral part of the browser – and the whole idea here is to be part of the browser.

    Flash, as I said, isn't distributed in the open source version of Chrome, Chromium. I suppose the real hypocrisy, if you want to argue that, would be that Google encourages Flash use in their App Store for Chrome.

    I've seen the calls to remove Flash from Chrome. It seems to me to be a straw man argument. Flash isn't part of Chrome, period; video decoding is part of the Chromium code base.

    If Google removed Flash support from Chrome, people would simply install the Flash plug-in and get it back. Obviously, *blocking* them from using Flash would be pretty awful; there's no reason to do that.

    But by supporting VP8 and Theora in the browser and not H.264 – this is in the browser code itself – Google levels the playing field with Chromium and Firefox, neither of which can legally use H.264.

    Yes, VP8 and Theora are each potentially subject to patent litigation, so "unencumbered" is theoretical. Still, we can say this: the creators of VP8 and Theora have a free, blanket license for their own IP. H.264 and the MPEG LA do not.

  • http://www.blairneal.com Blair neal

    I'm also curious about the 'future-proof'-ness of these current codecs. For the last year or two I've been assuming that h.264 support would be around for a quite a while since it's tied into so many devices and things of that sort.

    Now that my raw video isn't on a tape anymore, it becomes a little more important to be assured that I'll be able to open my DSLR files 5, 10 or even 20 years from now. I know no one can offer that kind of assurance…future proofing data is always a scary prospect.

    I've also run into the issue in wondering what to do with a performance venue's video archive…h.264 can save space, but if these potentially historically significant files eventually lose support, that could be quite a devastating turn of events.

  • Wiley

    The quality difference is not subtle if you are doing anything that involves particles, pixel dithering or anything that standard perceptual compression takes a dump on. Y'know, the kind of stuff tons of visualists work with? Heck, the quality difference between Apple's h.264 encoder and the (imho) vastly superior x264 encoder is not subtle. 

    I think it's a huge drag that the html5 video element is now unusable for people like me, who want to reach across platforms, but also want to use the best possible quality and have my videos continue to play so nicely in iOS. Thanks a lot. Keep patting yourselves on the back.

  • Peter Kirn

    @Wiley: If you feel that way, you should really become involved in patent reform. Seriously.

    Why? Because the quality differences you describe as part of "H.264" are not so much the codec as a collection of lots of little tweaks that improve quality. As different vendors amass patent portfolios around the technology, it becomes difficult to impossible to share those techniques. I'm not opposed to intellectual property, but it gets to a point where no one can do anything. (Imagine, in particular, this scenario: Jo X works for company A. She switches to company B to work on a new project. Short of erasing her entire brain, just building on her experience at company A with something like video encoding could expose company B to patent litigation. This to me makes no intuitive sense.)

    As you say, different encoders can result in different quality. That means assuming this is an H.264 versus VP8 thing is probably premature. VP8 is a work in progress, and it's improving. It'd improve faster if not for these nasty patent issues.

    When I say the quality doesn't matter, it's because a lot of what you see online is transcoded again for delivery anyway. So the problem for those of us who care about quality is more complex than just this particular battle. And it is indeed messy.

    As Blair says, none of this stuff is future-proof, either. Conventional wisdom says the debate we're having now will probably be irrelevant in a few years.

  • Wiley

    I guess I'm crying out for a minority of people who want to do their own html5 video elements on their own encoded and own hosted sites. The reason that users like me are a minority is because the html5 video element has been so politicized and so gang-borked that no one wants to use it, and we all give up quality and presentation to social video sharing sites.

  • Peter Kirn

    No, it's absolutely a good point … and it makes me interested in doing some more tests with these new VP8 encoders. (It's possible to do an A/B test, too, even with free tools. Because the patent concerns are largely restricted to the US, there are plenty of H.264 encoders out there. I can pretend I'm not in the US to test them legally. I could wear a beret.)

    I'm not unsympathetic, either; I think it's just important to note that folks like Mozilla *aren't* doing this for political reasons. It's not ideological or philosophical. It's something they have to do to comply with the law. And for all of us who appreciate the practical, non-theoretical benefits of using free software like Firefox, we're bound by this set of rules.

    So, the political wrangling is largely the Web getting into Google and Apple fanboy fights. But some of these questions aren't just political; they're real. As you say, that leaves the rest of us to pick up the pieces and try to actually publish video.

  • Wiley

    I also think this is endlessly naive to think this is about openness and not about a FU to the iOS. 

  • Peter Kirn

    @Wiley: Sorry, but the facts don't back you up. As Gruber (ironically) himself pointed out, Android still supports H.264. In fact, WebM support has only just arrived in 2.3's mobile Safari. I'm not sure Android developers even yet have access to a VP8 decoder. And the Chrome team is completely separate from the Android team.

    I don't think this is Android versus iOS, or the Android team would've been involved. And it was Apple's decision – against the wishes of many people involved in the actual HTML5 development process – to standardize exclusively on H.264.

  • Peter Kirn

    Anyway, I don't pretend to know Google's motivations, beyond what they claim. As I said, the actual *upshot* for free software developers is good, which is the only thing you can evaluate with any degree of certainty.

    It's possible Google is doing this as part of a covert plan to prepare the human race for enslavement by alien invaders who don't like H.264 and don't own iPhones. I don't know.

  • Wiley

    We're talking about a future landscape, where neither one of our arguments can be defended, because we simply don't know. Neither you nor I know what the landscape will be like in terms of available hardware decoders by the time Google actually leverages, say YouTube, to switch to WebM.

    Either way, cheering on as some sort of revolution for openness is premature to say the least. And it blows for anyone who was an actual content creator who supported html5 video. 

  • Peter Kirn

    Let me simplify, then:

    Status quo: open source browsers can install Flash. Of course, we all generally don't like Flash, particularly on mobile.

    Future in which H.264 + HTML5 video win out: Open source browsers die. It's the end of Firefox.

    Future in which VP8 + HTML5 video win out: Open source browsers do just fine – and with all the advantages of video being part of an HTML Web page, rather than segregated to a plug-in.

    That's the issue. H.264 with HTML5 was never a solution. Of course, yes, the latter scenario is still a messy one. But the former wasn't much of a solution, unless you live in a universe where everything is made by Apple.

  • Wiley

    The best implementation of html5 video was the way chrome did it previously. codec checks with fallbacks. 

    They have downgraded the application and you are throwing a party.

  • http://pixlpa.com Andrew Benson

    I'm with Wiley on all his points. H.264 isn't going anywhere anytime soon. It's used by numerous consumer video devices for hardware encoding and has been championed by the industry as the way forward (I personally disagree, but who cares). My point is to say that throwing all your support behind any one codec or container just limits content creators and end-users. Perceptual interframe compression sucks for the type of graphic work that often appears on this site and in our larger community. There isn't, and will likely never be a one-size-fits-all solution for video delivery. Going forward, an expansion of options seems like a better solution.

  • Chris Thorpe

    A couple more points:

    Why must I choose between H.264 and VP8 for my HTML5 video tags? If I'm perverse enough, why can't I have Real video (or whatever that shit was called)? Or, why not some groovy new not-invented-yet codec?

    Second – what do browsers have to do with video decoding anyway? Leave that to OS libraries, or better still, to plugins and let us decide what we want to install. 

    Anyway, this is going to be fun to watch. VP8 could kill H.264, or it could be Google's Plays For Sure. No doubt Facebook's own codec is on the way – it will be free but will over-expose your private parts. (Blatant attempt to start an 'If xxxxx was a video codec…' meme)

  • Peter Kirn

    @Chris: Actually, you're asking a very pertinent question. According to the HTML5 spec, you *could* use any codec you want. Various parties to the drafting of the spec weren't necessarily happy about that, but objections from Apple and Nokia (as I recall) sunk the idea of standardizing on a codec (which then I think was to be Theora, since VP8 wasn't yet around).

    @Andrew: No, good points … but that could be a good argument for plug-ins. For some sort of standard video tag, WebM + VP8 seems to me the best choice, for HTML. It's not going to work for all recording, encoding, etc.

    Google says explicitly that they're not throwing everything behind VP8 and Theora, that other codecs – better codecs – could be possible in the future. People largely ignored that part of the post. There does seem to be a recognition that H.264 is running away with the HTML5 momentum, which is counter to what many believe (outside Google, too) is constructive for HTML5.

    I don't see any reason WebM can't do what you need as a container, but for codecs, I agree. One size won't fit all.

    For this community, I wonder how much *video* is always our best distribution outlet, given we can now deliver our work as it looked originally, rendered in OpenGL. ;)