Google's dropping H.264 from Chrome a step backward for openness
arstechnica.com
arstechnica.com
1) Google uses us as tools in its battles and wars, because it is us who will be inconvenienced for years to come if companies continue to fight over video codecs. Sure, the explanations are grandiose — "we fight for openness" and the like — but the fight is being done by our expense.
2) Google's hypocrisy is unbelievable. H.264, an ITU/ISO standard, documented, with reference implementations, with a free/oss encoder (x264) and hardware support in a bazillion devices is called "closed and proprietary". Meanwhile, Adobe Flash (which this whole fight will just strengthen) is just fine, thank you. Oh, MP3 and AAC are fine, too.
I'll skip over technical details in this post, suffice it to say that I know and understand why H.264 (especially Main and High profiles) is vastly superior to On2 VP8 (now renamed WebM).
The h264 removal in Chrome is probably an experiment more than anything else. It will make people remember the h264/WebM debate again, and it will show that Google takes a stance for WebM.
The long term goal is to fragment h264. First move is this. The next move is to play an alliance with Adobe: We can make your Flash survive for 5 more years if you include WebM playback in it. Then you play the piece where you force Apple to adopt WebM by virtue of Youtube: you let it hint that Youtube will in the future only play back WebM. At this point in time, all Android devices have WebM hardware decoding circuits (notice that mobile phones have an incredible short half-life time. You don't see a 3-4 year old mobile phone much these days).
At this point, all of the internet is WebM capable and h264 has been limited to every device which is non-internet. The MPEG-LA consortium will have been marginalized by this time and will have to bite the apple and add in WebM support as well.
/conspiracy
I think I know which one the world is behind and which one Google is behind. I vote: world.
First, Google is interested in boosting the web indirectly because it puts their technology (adwords) out there in the right place. If you watched most of your video over WebM over the net, it would indirectly benefit Google. They also know that a large part of their power stems from the fact that they do "nice" things for their users, while their customers pay Google to facilitate a link between googles users and googles customers through adwords. Using their market share for the "greater good" is highly valuable to them in the long term.
Second, if you own the WebM format, you could use it to bully the MPEG-LA members individually. MPEG-LA is an alliance with the sole purpose of keeping everyone else out of the alliance. The patent pool in MPEG-LA is a defensive measure - not an offensive one.
None of the things Google does seem to be short-term invested. It rather looks like they are shooting for 5-10 years ahead.
So perhaps Google should start saying openly it's about the money, not about "openness" (which seems to be defined every week as something different by one company or another).
If h.264 becomes the de facto standard for HTML5 video encoding, then a browser such as Firefox, or any free/open browser, will be financially unable to implement a completely standards compliant browser due to the cost of licensing.
Yes, one could supply other codecs to support the video tag since no codec is specified in the spec but if the majority of content out there is focused on h.264 then the effect is the same as not supporting the video tag at all.
As far as I am aware software patents only hold in the USA and a few other countries that the US has strong armed into accepting them. So I don't think you need to worry about the MPEGLA coming after you if you are somewhere on the other 90% of the planet.
Also there's the World Intellectual Property Organization (http://en.wikipedia.org/wiki/World_Intellectual_Property_Org...) of which the majority of counties are members.
Yes, specifically software patents may be considered a bit of an unconfirmed area, but I doubt it's something many people would want to risk going up against.
So anyone is free to tinker with x264 as much as they like. The people who have trouble are the free OS distribution folks, e.g. Debian, who ship enough to trip the license… and google chrome.
Maybe they just got tired of paying a quarter for each download.
Imagine a scenario where an ink/printer producer patents blending of the colors and every newspaper in the world has to pay $0.25 for every paper they print after 10,000 units.
What I hope that people understand, and I'm aiming high by including the well connected writers and media personnel, that tools that serve as backbones for information flow need to stay free. There are software engineers working on that tirelessly. To undo their effort or trip them up, is to hurt the very rights to freedom to information.
It's far worse than that! My city's newspaper is printed with patented ink[1]. They pay for every drop they use, even for papers they don't sell, even for ink they wash down the drain.
And your numbers are way off. Imagine if I had to pay a one time $0.25 fee to read my $22/month newspaper would be more like it.
I look forward to a time when I can code and not worry about patents[2], but hobbling people to make a strategic move is not the path to that end.
Cost Thing
################ Computer hardware: $400/year
####### Computer software: $200/year
## Power for computer: $50/year
############################ Internet connection: $720/year
. H.264 license: $0.25
(Note: The graph is probably off. I have 26 pixels in a #
so the '.' would need to be a single pixel, but is four.
So, mentally expand the '#' lines by 4 to get the
perspective right. Oh, and that is not annual, so
maybe another three times.
And while we're at it, notice that video is probably your
biggest bandwidth need and that by using a more efficient
encoding you can use the next tier down at your ISP or
pay for fewer bits and save 100 times the cost of the
H.264 license.)
[1] Probably. They announced switching to the new eco friendly ink with great fanfare. I suspect if they went back to a traditional ink I would not hear.[2] I've personally cancelled a lucrative product after development was completed over patent fears (needlessly it turns out, the patent owner in question never elected to go hunting and we probably didn't infringe but were unwilling to endure the legal distraction), and just this week was discussing picking the bones of a local company that briefly stepped on a ridiculous software patent and had their plug pulled by the parent company instead of fighting the lawsuit.
Every company does. Or do you really believe that Apple is actively preventing Flash just because they really care about the consumer?
Google's hypocrisy is unbelievable
Google is a big company. The world is full of crazy and complicated choices, and there are many shades of gray.
I'll skip over technical details in this post...
Um...okay. Thanks.
Microsoft (Exchange), Cisco (VPN), Yahoo (Mail and Weather), and Google (Mail and Maps) all have code and/or data provided by the OS in iOS. In all of these cases, these third parties improve the iOS experience without any significant trade off.
Flash on the other hand has a huge trade off in application responsiveness and battery life. These are features Apple touts as being important to how their device is the best device for their customers. Flash would undermine those features, and Adobe has had a poor track record of delivering on promises regarding Flash (significant performance improvements are always one release away, for example), especially for non-Microsoft OSes.
Flash, on the other hand, is closed-source. Adobe could, at best, provide Apple with a binary plugin to drop in. Apple would not be able to fix any bugs or track down any problems, and judging from their performance on OS X, it would be the source of a lot of crashes Apple couldn't do anything about. Apple might announce iOS 5.0 in June, a few days before or after Flash 11 comes out. Now what does Apple do? Rush to incorporate a major version change into a minor version release? Hold off on releasing until Adobe can get them a stable version? What if it's buggy? Now users who bought a new iPhone with iOS5, or users who updated to iOS5, think that Apple's released a buggy release, when in reality they've released a (probably good) release with a crappy version of Flash. This is exactly what happened with Snow Leopard already, and Apple caught flak for distributing buggy software because Adobe had released security fixes after SL went gold.
It's not about control, but determination. Adobe for years hasn't been able to show off a stable, reliable, and performant release, and so in exchange for Apple accepting Adobe's crappy code into the OS, Apple would also lose the ability to speak for the quality of the OS and experience as a whole. It's lose-lose for Apple, so where's the incentive? They got screwed with Snow Leopard, they'd get screwed with iOS, and in the end the ones who suffer are the end users, who can either wonder why their browser is so slow/crashy/laggy/jittery when scrolling, or who can wonder why they don't have Flash, but then shrug it off because most sites don't need it.
> ISO and ITU do not require the members working on their various standards and specifications to give up any specific patent claims that may cover the technology that they define.
Quite the contrary -- the ITU specifically had language that every contributor agreed to provide their contributions royalty-free. Unfortunately the politics got complexated when it became a joint ITU/ISO thing. Some of the biggest players made noises that they would agree to a "royalty-free baseline profile" (ie without some of the more patent-rich features, such as interlace and B-frames).
At the last minute, like Lucy from Peanuts, they took their ball and went home, leaving all the smaller players who had an interest in a truly free standard high and dry. This event IMSHO left a vacuum that resulted in pent-up demand for a truly free codec, which created room for Theora and WebM.
[disclaimer: I was involved]
MPEG-LA are really just a bunch of thugs. They remind me of Tony Soprano: "Satisfaction guaranteed -- or double your garbage back!"
Many sites that serve mobile devices html5 video still serve flash to html5 capable desktop browsers.
And I would argue that the majority of traffic for websites like YouTube comes from desktops / laptops.
Also, Chrome is heavily used by early adopters and tech crowds, so it will influence their perception of things and this might trickle down to more average users over time.
VP8 and Flash are not open standards, but they are royalty free. AVC was openly developed and is an ISO standard but is not royalty free.
The problem of FMS parity is a wholly different ball of wax, aside from (although perhaps relevant to) the behavior of the Flash VM itself.
Although the Ars article mentions that MPEG-LA won't charge per-stream royalties for free youtube videos (for 5 more years, anyway), H.264 content from the Youtube Store are already subject to per-stream fees, starting at the beginning of this year. Also, Google TV streams may not fall under MPEG-LA's definition of royality-free "Internet Broadcast AVC Video". Furthermore, MPEG-LA hasn't yet made it clear exactly how much these fees will be!
WebM (vp8), along with lower-quality FLV (vp6), is a hedge against both current and upcoming per-stream h.264 royalties.
http://www.mpegla.com/Lists/MPEG%20LA%20News%20List/Attachme...
But there is a concrete issue involved: under MPEG-LA's licensing terms, Google should be paying $6.5 million a year to distribute a H.264 decoder. Under that license, the end users only get the rights to use the software "for personal and consumer purposes": they couldn't fork Chrome, redistribute it, etc. The Ars article mentions this, but immediately moves on.
It's possible Google expects MPEG-LA to try to enforce the patents, and that's why they're dropping H.264 support. After all, Google is the only browser vendor not in MPEG-LA worth trying to get money from. The yearly licensing amount just went up by $1.5 million, which might have influenced the timing of the announcement.
Comparisons with Flash are irrelevant: it may be less open than H.264, but not in the way that matters to Google.
if you weren't planning on doing so, you were either going to have to serve video via flash or not at all anyway. google's decision only tips the scales a little more.
+edit: note that I'm not prognosticating a new firefox monopoly or anything, just that firefox 4 will rapidly replace most of the current firefox 3.*s out there, which currently make up the referenced marketshare.
Here's the sticking point -- as the article says, no one's going to go through the pain of dealing with VP8 transcoding and storage. Flash will be for everyone except iOS, where HTML5 H264 acts as a backdrop.
however, there are increasing benefits to being first-class content in a browser, and there will be at least a few major providers (youtube, maybe (someday) people encoding to theora today) willing to transcode and distribute for you.
I believe (and I recognize that this is just my opinion) that web video should first be in an encoding that is free to use in any way. The only way this is going to happen is if Mozilla (and to a lesser extent, Google) make a stand now instead of waiting for one of the many long years between now and 2028 when the last h264 patent expires.
And Google can laugh all the way to the bank while they build and promote their own competing format.
The details of the implementation are interesting and concerning; however, your judgement of "people like me" is not germane, nor is it based upon any information in this context.
My extrapolation is a reach, but supporting h264 on Firefox doesn't make sense except if MS is seeing cash from h264 ubiquity.
Supporting H.264 makes sense for Microsoft as a licensor of the patent pool.
http://blogs.msdn.com/b/ie/archive/2010/05/19/another-follow...
No, that doesn't particularly make sense. Choose your battles. They are endless in number. Nobody has time to fight them all. Fight the ones that matter most.
All else aside, there is nothing hypocritical per se about taking a stand on one issue and not another.
If I understand things correctly, for people shipping decoders the royalty amount is capped at $6.5 million a year. So they're making $6.5 million from Microsoft, Apple, Sony, Toshiba, Samsung, etc. All in all, probably a few hundred million over everyone?
Would it not be more advantageous to leave that money on the table, completely entrench your format EVERYWHERE, and then pick it up on the professional/encoder side of things?
CNN, CNET, Fox, whoever, probably wouldn't mind paying a little more and getting a single universal format for everything everywhere and home users/hobbyists can play for free since they'll never hit the thresholds. Startups can get going by working in the free zones (< 100k paid subs, free un-paid use).
This is simply not true, MPEG-LA would go after any commercially successful project not paying royalties.
If someone comes up with a catchy slogan for their concept whether it's "Free Market" or "Fair Trade" it's simple politeness to engage with that and make a substantive point rather than making up your own definitions to suit and play semantic games.
If you read the ITU reaction to the attempt to define "Open Standards" this is obvious.
http://en.wikipedia.org/wiki/Open_standard#ITU-T_definition
Any limited use of "open standard" before that point was as meaningful as someone referring to "free software" or "open source" before those terms were defined.
A term does not need to have a formal definition in order to have a meaning.
For the love of God, stop saying 'Openness'
Edit: I think it's harder to argue against those points than it is to point to their hypocrisy in using both open and Flash (although I think that argument ignores the fact that they have to pick their battles, and Flash would not be one they could win). The "patent unencumbered" part seems to only be answered with vague accusations about possible patent infringement. The only real analysis on that is one blog post by a codec engineer, and as people here have pointed out, it makes more sense to trust that Google's patent lawyers looked at a multimillion dollar acquisition a lot harder than that analysis. The fee part was answered pretty weakly in the Ars article. Fees would cripple open source projects, and they just tried to gloss over that by pointing to Google.
Even the title doesn't make sense. They could say "step backwards for easy of use" or "step backwards for MPEG LA", but saying "step backwards for openness" is ridiculous.
If it's encumbered with patents, it's not open, no matter how many free software implementations it has. Open means, open to anyone, not only open to those who pay for it.
> This explanation is lacking, to say the least. It appears to be a conflation of several issues: openness, royalty-freedom, and source code availability, among others. In the traditional sense, H.264 is an open standard. That is to say, it was a standard designed by a range of domain experts from across the industry, working to the remit of a standards organization. In fact, two standards organizations were involved: ISO and ITU. The specification was devised collaboratively, with its final ratification dependent on the agreement of the individuals, corporations, and national standards bodies that variously make up ISO and ITU. This makes H.264 an open standard in the same way as, for example, JPEG still images, or the C++ programming language, or the ISO 9660 filesystem used on CD-ROMs. H.264 is unambiguously open.
> In contrast, neither WebM's VP8 nor Theora were assembled by a standards body such as ISO. VP8 was developed independently and entirely in secret by the company On2, prior to the company's purchase last year by Google. Theora was created by a group of open-source developers…
Let's just make forks of Chromium and Firefox that include built-in support for as many codecs and file formats as possible. Someone wants to use a psd in an img tag? Sure! Someone wants to use a flac in an audio tag? Go for it! Want to use a bink video in a video tag? Be my guest.
Even if it's illegal, who cares. You can do it the same way that Windows Media Player and Ubuntu handle it. Every time the user comes across a video they can't play, a popup asks them to click a few times, and then voila! They have the codec.
Web developer are always going to have a crappy time of it. We will never ever be able to simply make one site one way and have it work nicely on all devices with web browsers. Give up on that dream. But the dream of making things good for the users, is achievable. Aim for that first.
A prime example is embedding custom fonts with CSS. The code that handles them wasn't built with arbitrary fonts downloaded from untrusted sources in mind (I think the guy that wrote the code said that's insecure and a bad idea).
Having a company like Mozilla pay a royalty on each user would cripple Firefox and any new browsers that may come along.
I keep hearing the Flash analogy, which is nonsense, flash doesn't just do video playback, the technology is ubiquitously used in a variety of ways and is not comparable to a video codec, and does not have any of the patent licencing issues h264 has.
This move is a step forward towards openness as it will force publishers and users to consider the non-proprietary alternative.
I mean, isn't it strange that browsers that only support certain vendor-chosen codecs are described as more open than the browsers that allow the user to install any codec they like?
What on earth do you mean by "hijacked"? Web video has been dominated by h.264 since it was added to Flash, and Google itself is among the most prominent pushers. It didn't sneak in through the back door, it reached it's current status the same way anything else on the web has: by being useful.
Mozilla and Google are fighting for your and my right to produce fully compliant browsers without having to pay somebody else.
This is what gets me scratching my head, because it's Mozilla and Google who are the ones trying to make browsers, not me. Mozilla and Google are acting in their own interests (which is fine, don't get me wrong) but they keep telling me that it's for my sake even though, if asked, I would tell them not to do what they're doing. It feels more like an ideological bludgeoning than a favor.
Also, I take issue with your saying "fully compliant". Codec support hasn't been a standards compliance issue since Ogg was (rightfully) pulled from the spec.
If anything, the fact that the spec is codec-agnostic suggests that the most "spiritually" compliant position would be for the implementation to be, as well. But that's not what we get from Mozilla or Google. For that kind of openness we have to turn to...Microsoft and Apple? Weird.
You're right, that was a poor choice of words. I meant dominated.
>Mozilla and Google who are the ones trying to make browsers, not me.
The point is that you shouldn't have to pay royalties to someone to ship a competitive browser which has support for the most widely used codec. Think of Rockmelt or Flock. If they had to pay millions to stay competitive, isn't that a net loss for the internet?
>Also, I take issue with your saying "fully compliant".
You're right again, I meant a competitive browser.
>If anything, the fact that the spec is codec-agnostic suggests that the most "spiritually" compliant position would be for the implementation to be, as well
I don't understand exactly what you mean here. I thought the whole point of HTML5 was that video becomes a first class citizen, so if I had to install a plugin for each codec, wouldn't that defeat the whole purpose? Or are you saying that every browser should include support for the widest range of codecs that it possibly can? I think its unfair to ask a non-profit organisation like Mozilla to license a codec for many millions of dollars. Could you elaborate?
No. Quite the opposite, actually.
I am saying that a reasonable approach would have been for the browsers to defer support to the operating system. Apple does this. Microsoft does this. Mozilla, Google, and Opera have instead chosen to limit support to only the codecs they implement in-browser. (I say "would have been" because there's practically zero chance of this happening now.)
One effect of this is that even on a system with h.264 already present you can't view it in those browsers. This is not a requirement of the HTML5 spec, an h.264 licensing issue or a technically intractable problem, but a choice made by those browser vendors to favor certain codecs.
The point is that while the debate is frequently framed as a standards issue it's really more about the ideology of implementing things on top of those standards.
As a Chrome use, what changes? All three codecs can be watched, H.264 however is through a sandboxed Flash player.
Stating that Webm might turn into a patent minefield is FUD now? What happened to journalistic freedom and rational discourse?
1) WebM infringes on h264 patents AND h264 infringes on VP3 patents. Remember VP3 is the predecessor to WebM and came out before h264.
or
2) neither infringes on the patents.
So the 3rd unlikely possibility is FUD, yes.
The HN crowd don't really respect a conjecture without evidence. That's kinda why I come here.