Mozilla and I have a long and not-so-storied history together. I first began running what was at the time called the Mozilla App Suite with the “Milestone 10” release in October 1999. I was (I believe) one of the first people outside of the core team to build (as in compile) the browser that would come to be known as Firefox. It was called Phoenix at the time and the team hadn’t released any binaries. I remember talking to the core team on IRC getting instructions as to what build flags to add to build Phoenix rather than the app suite. For a time, I released unofficial nightly versions of Phoenix (then Firebird) compiled with Xft support for anti-aliasing on Linux that were distributed through MozillaZine. Long story short: I’ve contributed only slightly to the project; but I’ve supported it as long as anyone and generally my ideals match up nicely with the Mozilla team.
Until now.
I’m a huge supporter of open formats; I always have been. One of the main reasons Microsoft was able to rise to its Monopoly-level dominance was the proliferation of the proprietary MS Office file formats. However, I have reason to believe that Mozilla’s decision not to support H.264 encoded video via the HTML5 video tag due to the “patent encumbrance” of the codec, is a wrong decision and one that, unless they change their mind, will kill any hope of ushering a new era of online video distribution that exists without plugins. Mozilla has always been an organization willing to take a stand for what they believe in; and they believe in the open web.
Three of Four
With Microsoft announcing support for H.264 video in IE9, three of the four “big name” browsers will be supporting H.264 video. Truthfully, this move surprised me when it was first announced; I had assumed, like others, that Microsoft would choose only to support their own proprietary format. The success of Office taught the team in Redmond that simply using proprietary formats isn’t good enough; you have to completely own the format. However, without support from Mozilla, H.264 can never be the “encode once, deploy anywhere” format that people and businesses need it to be.
Encoding is slow and expensive
Video encoding is a very processor-intensive process; it’s time consuming, expensive, and the resulting files are large which lead to bandwidth and storage costs. The current Flash implementations on the web are already H.264 capable, and much of the video content for the web is already encoded in H.264. It’s nothing but an expense to create and maintain additional encodings. And even if we as web developers can create a robust fallback system that prevents the user-experience of online video from regressing to the days of “choose your format,” the added costs are something most companies simply cannot ignore.
Theora Sucks
I say that harboring no ill-will towards the very talented engineers who have sunk countless hours of their time into Theora. But as numerous comparisons have shown, it simply can’t keep up with H.264 in terms of quality; especially at low bitrates. MPEG2 was once considered great; but in the face of the current top-dog codec, it (like Theora) sucks. People and businesses are willing to embrace free software when it provides an equal or better product than the proprietary alternatives (see the success of Linux on the server). However, when free software doesn’t keep up with the best non-free products, people stay away (see the lack of success of Linux on the desktop). Simply put, there just aren’t that many people who share the same moral imperative as the Free Software Foundation; most of just want it to work.
There’s a Precedent Too
Mozilla has a track record of not sharing the stance of “nothing proprietary,” which implies that as an organization, Mozilla is willing to pick it’s battles. If it wasn’t, it wouldn’t allow proprietary plugins like Flash. Imagine also if Mozilla had taken the same stance against GIF images in the early 2000′s when the format was patent-encumbered just as H.264 is today; the browser barely would have gotten off the ground. In that instance, pragmatism and the need to not break the web won out over a desire to only support free and open formats.
Don’t Break the Web
Team Mozilla: I understand your desire to show support for free and open formats and I empathize with your belief that the best way to advance the web is to truly embrace those formats. But I wish this fight was one in which you’d lay down your sword. I disagree that “being idealists” is your reason for being. I believe that your first responsibility as a browser vendor is to make the web a better place; it’s certainly what you’ve strived to do since your inception. But honestly, from the view of a realist, the only thing you’re going to accomplish is making the HTML5 <video> tag unusable.
Two Proprietary formats
What’s worse, you’ll force people to continue to use Flash as the preferred delivery platform for video on the web (which, as covered above, already uses H.264 for video). Instead of one proprietary format, people will now be forced to use two simply because of your refusal to accept that sometimes you have to take a few small steps to get to your goal. I hope you can at least agree that going from two proprietary formats to one is a marked improvement.
Please…
…don’t do this to one of the most exciting and promising technologies being delivered in HTML5. You’re voluntarily creating another format war and the only result of format wars is people avoid both alternatives because it’s too much work to support both. Help us continue to make the Web a better place and move it forward one step at a time; even if those steps aren’t quite as long as you’d hoped they’d be.
Update (8:20 am EDT)
I’ve created a response post that addresses some of the points in the comments titled Mozilla’s Not Non-Profit (and Other Thoughts). Thanks to everyone who read this and considered its content and has an opinion to share. Disagreement builds great conversation!
Update (March 22)
I’ve written a follow-up piece about browsers supporting Theora and my belief that despite the its shortcomings, it has a real place in the future of the Web. Please check out Dear MSFT and AAPL, Embrace Theora!
Sorry buddy, but this was a *SAD* post. Didn't you learn anything form GIF man …?
H.246 == potential lock-in…!
Oh yeah…we all got totally screwed by supporting gif…
*eyeroll*
Pingback: The fight over HTML5 « mike.williamson
Pingback: LuisRei.com » Blog Archive » Dear Mozilla, Please Don’t Kill HTML5 Video
You're a dumbass for proposing pro the heavily draconian licensed H264 codec. Firefox is helping both our asses by doing what they do, except you don't realize it.
Pingback: Dear MSFT and AAPL, Embrace Theora! | Brian Crescimanno
Mozilla's decision not to support H264 due to patents is the right decision. Being idealists IS their reason for being, their mission is to support an open and accessible internet, not an internet where you can only post content if you have paid the MPEG-LA.
GIF is not a precedent. The GIF format was used in Netscape, Mosaic and other early browsers in the early 90s before anyone realised Unisys' LZW patent applied to it. In '99 Unisys announced that you had to pay a royalty if you used unlicensed software to generate (not view) GIFs. By the time Mozilla was open source everyone was using JPEGs intsead so the issue had all but dissapeared and in 2003 the patents expired.
Pingback: ★ GIF, H.264, and Patents | reabilita.me
Pingback: ★ GIF, H.264, and Patents «
Operas' market share is larger than Safari on every region except North America & Oceania (meaning Antarctica
, Europe, Asia, South America and South Africa).
For an example, in Russia Opera has the largest market share of all browsers.
I'm not saying you're wrong; only that I'd love to see some data that backs up those claims.
That said, I'm less concerned with regional share and more concerned with global share. For my blog at least (which theoretically should have the highest possible share of opera users given the demographics of readers here) Opera is still exceptionally small.
http://www.flickr.com/photos/bcrescimanno/44477...
Hi Brian.
Webkit supports h264, only if the h264 codec is available on the system, and it also does depend on how the developers implement webkit on their ports. Due to the patents that h264 has, for example, Qt, which has Qtwebkit integrated into it since 4.4, is supported through the phonon framework (I've made windows builds for arora, and I notice that whenever it encounters the <video> tag, it opens up the ffmpeg codecs I installed on my windows system) – funny thing though, that makes arora crash frequently, forcing me to use flash instead. Chromium, I believe, the open source version of google chrome doesn't support h264, although it has support for the video tag (allowing use of theora), through the use of ffmpeg. Again, this is because of the h264 licensing issues. I'm pretty sure that google chrome has h264 precisely because they pay for it.
Windows (at least for the starter, basic editions) for example doesn't have h264 codec support by default, you have to download it. I use ffmpeg compiled for h264 support.
There you have it. What initially went with good intentions of standardizing video through the web, became one that splits it instead – which was why flash was invented in the first place… because it provided a consistent venue for playing video anywhere, as well as animation (write once, play anywhere, like what java was supposed to do).
“With Microsoft announcing support for H.264 video in IE9, three of the four 'big name' browsers will be supporting H.264 video.”
This makes the rather big assumption that they don't support any other codecs. IE9 doubtless uses the Media Foundation framework. So IE9, like Safari, would support any installed codec. Therefore, just as you can reach 100% of HTML5 browsers with Ogg Theora today, you will almost certainly be able to reach 100% of HTML5 browsers with Ogg Theora tomorrow.
The bottom line is that major HTML5 video implementations (like Wikipedia's) will never use closed video formats. Any HTML5 browser worth its salt will support open video.
“But as numerous comparisons have shown, it simply can't keep up with H.264 in terms of quality; especially at low bitrates. MPEG2 was once considered great; but in the face of the current top-dog codec, it (like Theora) sucks.”
Those comparisons are deeply flawed. In fact, the one linked to in the Ars Technica article didn't even get the *colorspace* conversion right. No one's saying that H.264 with all encoding options turned on isn't good, but it doesn't matter how good H.264 *can* be when in practice that's not how it's deployed. The only meaningful comparison is to compare what a major site like YouTube does with H.264 in the real world versus what they could do with Theora. Here are two such comparisons:
http://people.xiph.org/~maikmerten/youtube/
http://people.xiph.org/~greg/video/ytcompare/co...
If you waved a magic wand and replaced H.264 with Ogg Theora on YouTube, no one would notice the difference on the basis of quality or bitrate.
“Imagine also if Mozilla had taken the same stance against GIF images in the early 2000's when the format was patent-encumbered just as H.264 is today; the browser barely would have gotten off the ground. In that instance, pragmatism and the need to not break the web won out over a desire to only support free and open formats.”
Well, I'm going to have to give up on you at this point. To use the clearest demonstration in the history of the web on the problems with embracing closed formats as a justification for why the web should do it all over again is frankly mind boggling.
Really, the core idea of your pro-H.264 stance is that H.264 usage is ubiquitous. However, this is the wrong way to see it. H.264 is in fact highly centralised and concentrated. Consider, if the top five video sites didn't have Flash or H.264 as a requirement, what use would the vast majority of web users have for either?
Change is easy to come by here. Google is likely to move to open video for YouTube (with VP8) and when they do, all other major video sites will follow.
Pingback: Mozilla, los formatos de vídeo y los dispositivos móviles | Daring Fireball en español
Pingback: Mozilla, le H264 ne doit pas passer par toi ! | Dico Micro
Pingback: Mozilla, le H264 ne doit pas passer par toi ! « Injazz Consulting's blog
Pingback: GIF, H.264 y las patentes | Daring Fireball en español
Hi Brian. I don't know if you're gonna read this, but I hope you will, if only in the hope that you might reply and tell me something I don't know or am wrong about (or that you see my point, at least).
One thing that I'd like to point out, first and foremost, is that H264 can never be the HTML5 video standard, because it doesn't comply with the W3C requirements, since it's not open or royalty free. Even if it'd be free, it'd need to be compatible with the open source development process. Which leads me to my second point.
Firefox cannot support H264 HTML5 video rendering because Firefox is open source, unlike all the other major browsers out there. As a Firefox supporter, you surely understand how crucial it is that Firefox keeps that status and remains open source. Because, in order for it to support H264 HTML5 video render natively, it would need to stop being open source. Because the decoder isn't open source.
Thirdly, you'll have to agree with me that Theora HTML5 video is the most widely supported HTML5 video codec. Indeed, it's three of four that support H264, but it's also three of five if you count Opera. And if you count Opera, it's also three of five that support Theora. To put it simply, right now, H264 is supported by about 10% of the browser market, if you put Chrome and Safari together. On the other hand, Theora is supported by 35% to 45% of the market, if you put Chrome, Firefox and Opera together. IE9, the “third” H264 browser, will take a long time to bridge that game, and, in the mean time, Theora will still be the most widely supported format.
My fourth point is about the cost. Encoding is expensive. Yes. And H264 is encoded a lot. But I should say that encoding H264 NOW is much more expensive than encoding Theora NOW, because you have to pay royalties for H264. And I think that the price those royalties take from web publishers will at least soften the blow that re enconding their libraries will be. In fact, in some cases, reencoding the libraries will be even CHEAPER than keeping encoding in H264. This is, of course, considering web publishers start using their enconding tools, because you have to pay royalties for that. Which, by itself, it's an enormous threat to the free web.
As for Theora sucks, that's right, but this war isn't really about H264 versus Theora, but rather about patented technology against open and free technology. It could be Theora, it could be Dirac, it could be VP8, I don't know. As long as it's free and open, and as long as it's a W3C standard, the web is ok with it, and, most importantly, the end user and the web publishers are ok with it.
In the end, no matter how much you or anyone wants it, Firefox is open source, and it can't have H264 support. And H264 isn't open or free, and it can't be the HTML5 standard. You can't really think that we're going to kill HTML5 by NOT supporting something that's not HTML5… That's like saying we're going to kill Windows by not supporting Mac OS X…
nice blog.. have a view of my blog when free.. http://www.lonelyreload.blogspot.com .. do leave me some comment / guide if can.. if interested can follow my blog…
Any further comments now WebM has arrived? Do you still think Mozilla should have given in and supported H.264?
Gerv
This actually sounds like trolling in my ears.
The license violate the spirit of free software. The license actually cost money (what I've heard, 50 SEK or ~$5 per user) that the end user will have to pay. Even though it may seem little, it adds up when Mozilla would add it. They would have to pay $5 000 000 (actual number, this is confirmed) per year to distribute Firefox with it.
And here's the biggest catch: They wouldn't be able to ship the code with the codec (just like with Chrome and Chromium), and whoever else that want to distribute Firefox with the codec would have to pay.
Too bad for the developers. But who cares, I guess they don't need any freedom, right? Isn't that what you said? They'll just have to pay up if if they want to help Firefox with adding cool features to the video playback functionality, unpaid too. That's what you wanted, right?
Everybody has to pay at some point. And depending on what they do, they might end up having to pay thousands of dollars some day in the future. You should know that promises are not legally binding.
Who cares about “how likely” it is that MPEG LA will sue people. People have thought all kinds of things were unlikely, and yet they got sued. Just think of GIF.
Remember that the the reason for it being implemented today is that the GIF submarine patent attack failed (they didn't get much money, probably nothing) and that the patents also has expired, and that some people still use it for animations.
I hope that WebM really can replace h264 soon.
In my eyes, Mozilla ARE the cool guys, and if they go to ANOTHER party, I'll follow them there. I'll rather be a little less cool in the eyes of strangers then to accept terms that can change at any time.
If they don't pay about $5 million/year, they can be sued by MPEG LA for that.
And what OS does it play H264 on/what extra software do you have? It might just be using something like ffmpeg or gstreamer to do it.
“At the codec level, the quality and low bandwidth of H.264 is exponentially better than Ogg, not just a little better. Google did a study of moving YouTube to Ogg and found that today's Internet does not have enough bandwidth for it, even if YouTube were the one and only Web site.”
I want to see numbers. As far as I can tell, that is all lies. Theora beats H264 in all the low-bandwidth tests that I have seen.
“The discussion of HTML5 video markup standards started in 2005 or so and ended in 2008 or so. The fact that we are talking about both of these settled issues right now is totally Mozilla's fault. “
No, it's Apple's fault for refusing to touch Theora. If they would have accepted it, then we would have had a free codec to use everywhere. Even MS would probably have accepted it then. Anyway, WebM is out and free now. Let's hope it can do the job.
Totally free, yes.
They have chosen to support video-capable devices. Not H264, they don't care about the technology.
Same thing with iPhone, unfortunately, it's the only thing that could provide what a lot of people wanted in a phone. Now the only thing that's unique about it is the hype. And the hype seems to be enough for it to stay big, at least for now and a while into the feature.
The reason for we didn't get screwed was because the submarine attack failed, ok? What if it wouldn't have failed?
The last patent covering GIF expired in 2006.
In my eyes, Mozilla ARE the cool guys, and if they go to ANOTHER party, I’ll follow them there. I’ll rather be a little less cool in the eyes of strangers then to accept terms that can change at any time.