You are reading a MIX Online Opinion. In which we speak our minds. Joshua Allen Meet Joshua Arrow


24Comment Retweet

HTML5 Video: You’ll Shoot Your Eye Out!

Jan 25, 2010 In Web Culture By Joshua Allen

You'll shoot your eye out! A few months ago, we shipped our cross-browser, cross-platform implementation of the HTML5 video tag, using the H.264 codec. In the past week, YouTube and Vimeo announced their support for HTML5 video tag, also using H.264. All good deeds, but unfortunately no good deed goes unpunished. Now we see certain bloggers clucking in disapproval, in essence saying, "H.264 video? You’ll shoot your eye out!"

The most detailed entry of this genus is Chris Blizzard’s post from yesterday. He starts out by raising the alarm (emphasis added):

The players from Google and Vimeo do present a pretty serious problem, though. Each of these require a proprietary H.264 codec to be able to view them. These codecs aren’t compatible with the royalty-free web standards that the rest of the web is built on. … most people don’t understand that something very dangerous is taking place behind the scenes.

He then goes on to assert that the Web grew because of royalty-free, and suggests that permitting HTML5 video to play H.264 will compromise this historical position and lead to a slippery slope.

Let me start by saying that I have nothing against royalty-free codecs; we’ve even published content on this site in Theora (unfortunately nobody played the content, but that’s a different story). But regardless of how idealistic you are about royalty-free codecs,the above line of reasoning has several problems.

Misleading Premises

For starters,it’s very misleading to say "These codecs aren’t compatible with the royalty-free Web standards that the rest of the Web is built on" or, as Blizzard says elsewhere, "The Web exploded on royalty-free". The implication is that the existence of H.264 codecs is somehow corrupting and endangering the earlier "royalty-free" purity that made the Web grow so quickly.

This assertion is simply untrue. The Web grew from nothing to nearly a billion people, and never used a royalty-free video codec. Never ever. In fact, the existence of all of that proprietary video content is largely to credit for the popularity of the Web.

Again, royalty-free is fantastic, and may even be important for the next wave of growth on the Web. But let’s not mislead people about the role that royalty-free video played in the growth of the Web.

Slippery Slippery Slopes

Next, I think the danger is overstated. It’s far too easy to invent "slippery slope" arguments to support any position, so such arguments should be used with caution. In the case of Ralphie’s "Red Ryder carbine-action, two hundred shot Range Model air rifle with a compass in the stock and a thing which tells time", the warning of "You’ll shoot your eye out!" may have been reasonable. But in the case of H.264 video, I think the "slippery slope" argument goes too far. Chris compares H.264 to the debacle with the GIF image format, but he appears to make the link largely for rhetorical purposes, and ignores the dissimilarities between the two situations.

The Web has depended on proprietary video formats for 15 years, and hasn’t yet been held hostage; it’s not in anyone’s best interest to try. Besides, the last time someone tried to hold the Web hostage, we routed around the situation pretty quickly. We all know how that movie ends, and nobody wants to watch it again.

Bottom line: The Web is now much bigger, with a far more competitive and diverse vendor ecosystem. The idea that someone would try holding it hostage now, let alone stand a chance of succeeding, seems ludicrous to me.

Lose the Battle to Lose the War

Finally, it’s difficult for me to see how refusing to support H.264 in the player will help spread royalty-free content on the Web. Content providers seem perfectly capable of publishing content that uses proprietary codecs, and getting users to install plugins to play that content. I don’t see how any browser vendor can stop this. So this discussion must be limited to codecs supported by the HTML5 video tag.

Apart from the fact that the ship has already sailed, it’s difficult for me to imagine how we would change publisher behavior simply by crippling the HTML5 video tag. In exchange for dumping the H.264 codec, what would the content provider get for switching to the video tag? An HTML tag that’s one character shorter than the object tag? A video player that requires users to install a new web browser instead of a plugin? There seems to be a bit of hubris in the idea that we can control content publisher behavior in this way.

Indeed, if we really believed in such rhetoric, we would boycott media players such as iPods, because they support formats like MP3 and AAC. We can do without music for a few years, while we wait for someone to build a royalty-free replacement codec. After all, it’s for the sake of the ecosystem!

There are many ways to encourage content providers to publish using royalty-free codecs. I’ll even go so far as to predict that content providers will publish more content using royalty-free codecs in the future. But it won’t be because the HTML5 video tag was deliberately crippled, or because people wrote "slippery slope" blog posts.

Follow the Conversation

24 comments so far. You should leave one, too.

Eeyore Eeyore said on Jan 25, 2010

Did you read Mike Shaver?

Mike didn''t talk about saving the world. Firefox saves 5 million by skipping the codec. Smart move I say.

Joshua Allen Joshua Allen said on Jan 25, 2010

@Eeyore - Yes, I think Mike''s post was pretty pragmatic and diplomatic.

The whole situation seems odd. Google spent the cash to license H.264 to ship with Chrome, but doesn''t make that available to Firefox, even though both browsers are open source and both funded by Google.

Then YouTube (owned by Google) and Vimeo both launch sites which specifically exclude Firefox, and which support Chrome and IE. In other words, despite Firefox making a lot of noise about HTML5 support, you''ll need to switch to Chrome or IE if you want to watch YouTube videos in HTML5.

It''s just weird, IMO. Both sites recommend that people use Chrome Frame to leverage Google''s H.264 license inside IE, even though Microsoft is already an H.264 licensee, and there are already two other H.264 implementations on IE which have wider deployment than Chrome Frame. And Chrome Frame is explicitly *not* available for Firefox. At a minimum, this would appear to be a move to drive Chrome adoption at the expense of Firefox, which cannot be all that morale-boosting for Firefox.

Of course, Mike Shaver has argued in the past that Chrome Frame''s model of swapping out the entire rendering engine is a bad idea. In any case, it would certainly be better to hijack an existing deployed decoder like Flash or Silverlight, IMO, than to overlay an entire rendering engine. And it should be trivial to write a GreaseMonkey script that uses our implementation to enable YouTube and Vimeo HTML5 to run inside Firefox (assuming they don''t do UA detection).

Bingo Bingo said on Jan 26, 2010

I think you''ve subtly misrepresented Mr. Blizzard''s point. He''s saying that the web (in it''s entirety) grew because of open innovation, not that it grew because of open video (which would be a ridiculous claim). In his version Flash, Quicktime and various other proprietary blobs that do video and other things rose on that same tide.

I also think that you''re not really giving the whole story if you don''t mention that a) Google Chrome plays Theora content, which it didn''t have to if they didn''t want to as no-one but the usual suspects is really bitching at Apple for not supporting it, b) Google Chromium, the much trumpeted open source foundation of Google Chrome (and Chrome OS) doesn''t play H.264 or any other patent royalty format, and c) that Google just paid $100 million to buy a codec company which, almost uniquely in the industry, happens to produce codecs that avoid the MPEG-LA patents.

You''re painting a picture of Google intentionally shafting Mozilla on this but the reality is much murkier at the moment and potentially much more interesting.

Bingo Bingo said on Jan 26, 2010

Also, have you considered a silverlight backed HTML5 video player that supports Theora? I believe the codecs are pluggable and some C# games already use a suitable version of Theora.

It would be technically neat at the very least.

andrej said on Jan 26, 2010

I agree with Bingo, but I don''t think the misrepresentation is that subtle. Mr. Blizzard said that the *technical foundations* for the web have always been royalty-free, not video codecs:

"The truth is in the tests: you can still build a web browser, spider, client, web server, image editor, a JS library, a CSS library, an HTML editor, a web publishing system, commerce system – anything that is based on fundamental web technologies – without asking anyone for permission. "

There''s no mention of video codecs anywhere in that section. Instead, there is an attempt in the essay to compare the historical royalty-free development of the web with the licensing situation of H.264.

I won''t pretend to have an idea of how this situation will play out, but I do believe the concerns voiced by the Mozilla folks are valid, especially with respect to costs for content creators and would-be hosts who are not as big as Google, or Microsoft.

Joshua Allen Joshua Allen said on Jan 26, 2010

@Bingo, @Andre - I don''t think I misrepresented his position at all. His entire premise is misleading, and deliberately so.

His post is specifically about video on the Web, so saying that "these paragraphs about royalty-free applies to everything except video, and thus are irrelevant to the article" would be no excuse. If they don''t apply to royalty free video, why put them in the article at all?

The test is simple: after reading his article, would the reader know: A) something between 50-70% of the traffic on the Internet is encumbered video, and B) video in these encumbered formats is responsible for the Web "exploding"?

I think not. I think readers will draw the opposite conclusion, which is completely wrong and not supported by any evidence, historical or quantitative.

Eeyore Eeyore said on Jan 26, 2010

Bingo is it true what you said? "Google Chromium, the much trumpeted open source foundation of Google Chrome (and Chrome OS) doesn’t play H.264"? Can you tell me where it says?

I make browse on Chromium. Why I cannot use same codec as Chrome. Microsoft gave codec to MonoProject, so I don''t think Google doesn''t support Chromium? If thats true nobody will use my browse, they just use Chrome instead. I don''t think you are correct.

l.m.orchard l.m.orchard said on Jan 26, 2010

Blizzard''s position (as far as I can read) is that for the most part, the web grew up on royalty-free tech. You don''t have to pay anyone or ask permission to build a new browser or author content for the web. If the rest of those royalty-free technologies hadn''t taken off, there would have never been a spot at the top of the stack for proprietary video codecs. What''s misleading about that position?

Video and audio have been the odd men out, and there''s a chance now to try to correct that. That''s what he''s talking about. It''s not that video up until now has been royalty-free, and suddenly things have changed—it''s that things are changing with HTML5 and now''s the time to try to fix this odd case of proprietary crap in the otherwise open web stack.

l.m.orchard l.m.orchard said on Jan 26, 2010

@Eeyore: Chromium can''t play H.264 because the license doesn''t allow releasing the codec source code. If they did that, then legally *you* would have to pay the license fee to build and distribute the codec.

Same goes for Firefox. Mozilla could pay the license fee and Firefox could ship with H.264, but it would never show up in the public source code. And so the Mozilla-distributed binaries would come with it, but no one building Firefox from source would get it.

Joshua Allen Joshua Allen said on Jan 26, 2010

@Eeyore - Yes, Google could pay the H.264 fees on behalf of Chromium or Firefox, just as we do on behalf of Moonlight on Linux. But why would they? These other browsers would then be on equal footing with Chrome.

@l.m. - If he were arguing that this is a good time to try to opportunistically change history, then I''m totally with you. If he were arguing mainly about why this change would be good for the future (he spends all of two sentences on emerging markets), then I''m also with you. I might disagree that taking a "principled stance" and crippling your own browser''s video tag will actually have any positive impact, but I would agree with the sentiment.

It''s just the "you don’t have to pay anyone or ask permission to build a new browser" part that I''m objecting to. The fact is, your Web browser has always needed to allow encumbered codecs if you wanted anybody to use it (as Eeore seems to be finding out).

Chris titled his article, "what history tells us", but history completely refutes his thesis -- video has *always* been proprietary, nobody has *ever* tried to hold the Web hostage over video, and the Web rode to success primarily on the back of encumbered codecs and proprietary content.

Again, I''m not opposed to people calling for a revolution and establishing a new order, but we should be doing it based on a vision for the future, not by distorting the historical record.

John Dowdell John Dowdell said on Jan 26, 2010

fwiw, I''ve never understood the "open codecs" argument well enough to relay it to others, but I do recognize that many folks feel it a valid need, and so I respect that desire.

(Best argument I''ve heard for "open codec" ecologies has been "because then anyone can write a video-editing tool without paying license costs"... seems reasonable, although I''m not sure how much impact it would have, considering state of Inkscape, Gimp, etc.)

If some would like "open codecs", that''s great, and I respect the desire even if I don''t share it. But it''s too bad that the last three years'' worth of hubbub about VIDEO tag and "Kill Flash & Silverlight!" didn''t focus on their actual core desire, though. The situation should never have gotten this messy.


mtz said on Jan 26, 2010

@Joshua Allen,

firefox is a FOSS project and can not ship with any proprietary codec/code and maintain its FOSS status.

chromium is also a FOSS project and also can not ship with any proprietary code/codec build in and maintain that status. Chrome is chromium with google branding and is not a FOSS project and that is why it ships with this codec.

You will have to understand how FOSS projects works and limitations of their licenses to know why firefox can not ship with this codec build in.

Even moonlight doesnt ship with proprietary codec, it offers to download a binary blob from microsoft when it encounters a page for the first time that require wmv/wma codecs.

The best firefox can do is doing what moonlight is doing or coming up with a codec plug in system and use codecs from the system and live it up to the user to install the codec through other means.

@John Dowdell,
"open codecs" means "you dont have to pay or ask for permission to use them". Its kind of hard to comprehend how problematic proprietary technologies are if you are using windows because everything starts there.

Take an example of activeX. It just plainly doesnt work on any other system other than windows and IE and any web page that uses it locks a user to windows and IE.

look at flash, linux users pleaded and begged adobe to release version 7 update for ages. At least now they release update at the same time on all platforms.

Microsoft, Apple, Google have enough money to license this codec so these browsers are safe, what about the rest? like konqueror, arora, rekonq, flock and the rest?

Firefox resisted implementing active X and it survived, i dont see why it shouldnt survive this time.

The web should be build on open and license free technologies. I am all for firefox sticking this one out

Joshua Allen Joshua Allen said on Jan 27, 2010

@mtz - If you knew who JD was, I doubt you''d accuse him of seeing the world through a "Windows" lens :-) Likewise, I suspect that Miguel would object to your insinuation (insinuation, since you''re not being dishonest) that linkage to an H.264 download automatically disqualifies a project from being FOSS -- particularly since Firefox and Chromium both send users to downloads of proprietary video codecs, and would both be disqualified from FOSS under such criteria.

I''m not sure where you''re going with "Firefox resisted implementing ActiveX and survived"? Firefox has always supported Flash and Silverlight, both of which are H.264 (either of which could be leveraged as a stop-gap HTML5 video implementation). That''s no different from the Moonlight situation, and was necessary for Firefox''s survival. Unless Firefox also killbits Flash and Silverlight, it''s hard to see how the deliberate lack of H.264 in their HTML5 video constitutes a principled stand.

In fact, I was on a panel with Mike Shaver about a year ago where he made the point that plugins are a suitable vector for innovation; and presumably a more suitable vector than something like CF. NPAPI is, for purposes of this discussion, equivalent to ActiveX, so the suggestion that non-support for ActiveX is a triumph for open standards seems a red herring to me, and a tangential discussion with little relevance to this thread.

And of course "survival" is a very dramatic word. Nobody is arguing that Chrome''s refusal to provide a link library implementing H.264 for Firefox is a death knell for Mozilla. But it''s certain to spread Chrome at Firefox (and Chromium) expense.

Jeff Putz Jeff Putz said on Jan 27, 2010

What amuses me about this "debate" is that people quote the standards as being the awesome enabler that made the Internet what it is today. Are you kidding me? HTML 5 is still not on track to be a ratified standard until 2012. How long has HTML 4 been around? And how is CSS 3 coming along? If it takes this long for standards bodies to adopt markup, what would it take to roll a video codec?

The scary awful "proprietary" stuff like Flash, and now Silverlight, has gone a long way toward advancing the Web and how we use it. The ideological nonsense does not serve users (we call them customers), they just want stuff to work.

mtz said on Jan 27, 2010

@Joshua Allen

i am not talking about linkage, i am talking about bundling and shipping proprietary code with FOSS code. This is not allowed and moonlight doesnt do it too.

Firefox can not ship h.264 implementation in its code, moonlight can not ship wmv/wma codecs because both will be a violation of licenses they are shipped with.

There are workaround to this. moonlight workaround this by shipping free of proprietary code and instruct a user to download the codec from microsoft site. This makes microsoft the distributor of the codec and not moonlight. moonlight license prevents it from distributing proprietary code

Firefox could work around the limitation of the license by building a plug in system that uses system codecs and there are proposals to let firefox use gstreamer(in linux at least) and let gstreamer and the user figure out where to get the codec. Opera said they will not license h.264 and will instread use gstreamer to handle video tags too on all systems.

Again, i am talking about bundling and shipping, not loading binary proprietary blobs that already exist on the user system at run time.

flash is a proprietary plug in that is loaded at runtime. There is a difference btw loading a proprietary code at run time(allowed) and bundling it and distributing it(not allowed).

flash,h.264 and other proprietary technologies are bad for the open web. Flash is a bad that will have to be tolerated because it already exists. h.264 is a technology on its way to dominance and firefox can play a part in discouraging its use and favor an open format instead.

ActiveX as a proprietary technology is relevant here. It only works on windows, it only works on IE, it fragments the web. Plug ins that are tied to any specific platform helps in fragmenting the web and that is bad

h.264 is a proprietary technology, it fragments the web. Only browsers with big money backing can support it legally, FOSS browsers can not support it natively and maintain their FOSS status, only websites with big money backings can support it.

mtz said on Jan 27, 2010

@Jeff Putz,

if you remember, back in the days, media online were handled with proprietary technologies from microsoft, apple and real networks. Microsoft and apple did not make any plugins for linux, real network player for linux was an insult more than anything else. The problem back in the days didnt seem that big on windows but it is an experience we wish to avoid in the future.

Flash won that round and us linux users were stuck with flash7 for what seem like eternity because it was a proprietary technology we couldnt do anything about and adode just plainly ignored us.

Flash won that round and most online media are in flash. Flash player in linux is a pig but atleast its up to date.

The battle tomorrow for online meda will be btw plugin based solutions(flash/moonlight/silverlight) and native solutions(html5)

With native solutions, issue with be btw using proprietary codecs (h.264)or non proprietary codec(ogg theora). These struggles exist for practical and monetary reasons and are not "ideological nonsense"

the debate is about "open web" vs "proprietary web"
..users will be best served if the web is free of proprietary extensions that requires permissions and money to be implemented and those that are tied to specific browsers/platforms

rezlam rezlam said on Jan 27, 2010

Apart from the ideological side, my point of view as both a developer and user is:

10 years ago I wanted to launch a video site, I simply gave up the idea after realising how cumbersome it would be to have to deal with Windows Media, QuickTime and Real Video, in order to provision for the os/plugin combination a user might have. H.264 solves the problem of encoding the content, one format. I know Theora does too, but the industry is moving towards H.264 (Blu Ray, ipods, etc... not even Microsoft could stop it), so case closed.

But it''s not the ideal word yet. To display H.264 video on a web page, users still need a plugin and the developer has to opt for Flash or Silverlight (Flash being the safe bet) and hope the user has it or is willing to install it.

The ideal world would be the one where I only need to encode my video assets in a single format and the user wouldn''t need any plugin to see it.

If the video tag adopt H.264 or Theora (without the Flash nor Silverlight hack) I don''t care, as long as it makes my ideal world comes true.

Joshua Allen Joshua Allen said on Jan 27, 2010

@Jeff, @Rezlam - Correct. I guess people will still be pragmatic as before, and continue sending the H.264 to Firefox via Flash/Silverlight. So in addition to slowing Firefox/Chromium adoption, crippling the video tag like this will also slow down HTML5 adoption and ensure that Silverlight and Flash hang around, since that''s the only way to get H.264 to Firefox.

tanner tanner said on Jan 28, 2010

Youtube had to turn on HTML5 quick because its the only way to show H.264 video on iPad.

Anthony A said on Feb 3, 2010

The so-called Flash vs. HTML5 war is a pointless one because those of us that really care about innovation look to Flash and Silverlight. It will be almost 15 years from the inception of HTML4 to recommendation of HTML5 (hahahaha).

With the platform to create compelling user interfaces and interactions developers can concentrate on giving end users better experiences so that they can do what they really care about--reaching their goals.

I want to see good competition between Flash and Silverlight because the ideas we''ll see in them in 5 years won''t make it into the "open web" till about 2040 (if things continue at the current rate).

John D said on Feb 3, 2010

In the end the most influential factor will be the porn industry. VHS vs. Betamax, HD DVD vs. Blu-ray, remember? I have a strong feeling that the video aspect will come down to this. Flash will still live on though.

amit mendelsohn amit mendelsohn said on Feb 8, 2010

Funny but i have a feeling that its all a conspiracy to remove Mozilla from the market. Mozilla dependency on Google''s funding make it fragile. Fragile company that control a massive part of the market. While Google that support it is also a competitor. Google assume and quite logically that Firefox users will move to Chrome if they won''t have an alternative. I believe that when the time will come Google will close the financing tap for Mozilla and will watch it die over the licencing of the H.264 issue . Hoping to grab its users.
Microsoft will be there waiting to eat its share and Apple... well they will save the money of developing proper support for Flash.
You see, when 3 wolves are playing together a sheep is going to be eaten.

joe said on Feb 17, 2010

H.264 is patented, closed format. The MPEG LA asserts "decoder", "encoder", and "broadcast" rights over H.264.

You even have to get a license (a "broadcast" license) to use H.264 on your websites.

joe said on Feb 17, 2010

Please note: Using Silverlight or Flash does not exempt you from this "broadcast" license. Read the fine print. You still must acquire a H.264 license if you use Silverlight or Flash to stream H.264 video.