open video in Safari

| 26 Comments

While Safari is supporting the <video> tag, they're doing it with patent-encumbered codecs brought to you through QuickTime. This means that Safari doesn't support the open video technology that Firefox does which today includes Theora video, Vorbis audio and the Ogg container (and will likely include Dirac and other codecs and containers in the future.)

The good news is that QuickTime is pretty extensible with plug-ins called QuickTime Components. The guys over at the Xiph Foundation (stewards of Ogg, Theora, and Vorbis) have built something appropriately named the Xiph QuickTime Components which when installed on your machine will bring Ogg, Theora, and Vorbis support to any of your QuickTime-enabled programs. This not only makes Safari compatible with Firefox's open video technology, it also brings Ogg/Theora/Vorbis encoding and decoding capabilities to iMovie, iTunes, QuickTime Player, Final Cut, and other QT-enabled programs.

I wrote a quick little howto doc over at Air Mozilla for installing the Mac decoders which will allow Air Mozilla playback in Safari. If you want to install the marginally larger full package for Mac to enable encoding (export from iMovie, Final Cut, QuickTime Pro, etc.) or you need the Windows package, you'll want to head over to the main Xiph QT Components page.

I'm also working on a short QuickTime movie/screencast that describes the simple process of adding the Xiph QT Components that can be used as an extra src= in your <video> element to help out your Safari users. Stay tuned for updates on that.

26 Comments

I think Mozilla should be proactive and install its Ogg components for system-wide use on these other platforms. On Mac, this means exposing the Ogg components as QuickTime codecs. On Windows, expose them as DirectShow codecs.

So, if Quicktime/MP4/H.264 are patent encumbered formats, what about JPEG, PNG, and GIF that Firefox (and everyone else) supports? Are they not encumbered by patents?

Looking forward to the screencast to include the Xiph QT Components.

Kenneth, shipping an H.264/MPEG-4 AVC implementation requires paying patent licensing royalties to a company called MPEG LA.

JPEG's patents were always in dispute and even so expired several years ago.

Unisys' LZW compression method (used in GIF) patents were never enforced against non-commercial use and those expired several years ago.

PNG was explicitly created to avoid the patent concerns about GIF.

So, to answer your question, no, those image formats are not patent encumbered. GIF and JPEG were somewhat concerning half a decade ago or more, but are no longer, and PNG was never patent encumbered.

- A

3 words: I despise quicktime.

Quicktime is the main reason why I don't install Safari on Windows. I used to install quicktime alternative but thankfully flash video and downloads in normal formats have kept me from installing quicktime for over 2 years now. I was very surprised when I installed iTunes for the first time and it didn't recognize over 30% of my music because it is in ogg or flac. iTunes was of course immediately removed and never installed again. Unholy relic of the Apple empire.

2 words: me too.

+1 to Matt's comments. And why limit it to Firefox? Why doesn't every piece of open-source software that uses Ogg push out the codecs in a format that every piece of software can use - QuickTime, Windows Media Player etc...

Asa, gotcha, thanks for clarifying.

Matt, I don't think it's Mozilla's mission, nor should it be, to spread these components to be system-wide. That could be viewed as intrusive as Microsoft installing Real Player on your computer or something. I agree with Mozilla's directions here, truly work to support Ogg in-browser as handily as any browser supports jpegs and pngs. The exciting part is that Firefox now has enough market share that it can influence the other big players (hopefully). I'm extremely hopeful that that, along with Thusnelda, will become a perfect storm for Ogg. In the meantime, mv_embed looks pretty promising.

It just sucks that a lot of 'big media' like YouTube probably won't support it because of its lack of DRM. Stupid DRM. Scourge of our time.

Boy this is confusing. I've been reading about when the H.264 patents expire and it might also be included in this open video implementation. I've seen some things that indicate that December 31, 2010 will be a game changer and it'll be able to be included in open implementations, and others that AFTER December 31, 2010 more patent license fees can be charged.

Even assuming that the former is true, I come across a whole new codec, H.265, due to be completed in the 2009-2010 time frame... just in time to start earning patent holders more license fees. Boy do we ever need patent reform.

There's so much conflicting data out there, it's difficult to know how to best future-proof stuff. I mean, we can guarantee that our family photos in JPEG format will always be pretty good, but with video it seems like there is a higher quality codec out there every day, and all of them with restrictions. I guess it's just going to take some time for Theora/smaller vidoes and Dirac/larger videos to settle in.

As someone set on implementing theora video, this is sad. Does this mean Chrome won't use Theora either? Or is this only Safari's implementation of Webkit?

Mozilla is in a very exciting position during the transition to HTML5. Whether anyone else likes it or not, Firefox is the second most popular web browser on the internet. The decisions Mozilla makes with future versions of Firefox are far more influential than what Apple, Google nor Opera choose to do. The only one with more power is Microsoft, and obviously they'll never officially support Ogg, Dirac or anything but their own proprietary codecs until the market absolutely forces them to do otherwise.

Apple, Nokia and others can try to fight Ogg's progression on the web, but hopefully between Mozilla and the rest of the open source community we can make Ogg and other open source codecs just as ubiquitous as Mpeg/Quicktime/WindowsMedia/etc... honestly, if we can get websites using Ogg formats even half as much as flash is used currently it can easily be considered a success. Unfortunately, in this stage of the battle, making 3rd party plugins for Quicktime and Windows Media Player are some of our best bets to help the transitions along.

Aza and the rest of the Mozilla team: thank you very much for what you're doing here. I'm thrilled to see audio and video functionality becoming a standard, built-in part of my favorite browser ;)

"It just sucks that a lot of 'big media' like YouTube probably won't support it because of its lack of DRM. Stupid DRM."

YouTube doesn't use DRM.

the problem about Theora is that it's out-dated, inefficient, and basically sucks compared to other modern codecs. If Theora truly replaced H.264 implementations in online video services like YouTube, it will actually be a step backwards, in terms of quality at least.

The free software community should aim to something free and superior, like Firefox and Vorbis, not something free but inferior, like Theora.

Yeah but why trying to make something new from scratch ?
Instead, the Mozilla Foundation gave $100.000 to the Wikimedia foundation "to help coordinate improvements to the development of Ogg Theora and related open video technologies".

http://blog.wikimedia.org/2009/01/26/mozilla-and-wikimedia-join-forces-to-support-open-video/

So maybe one day, Theora will be a really good and modern codec. I hope so.

the scourge of our time: Siverlight

And Silverlight ?

"YouTube doesn't use DRM."

Oh, really? YouTube thinks that it does. Here is how it is spelled out in the Terms of Service: "You agree not to circumvent, disable or otherwise interfere with security-related features of the YouTube Website or features that prevent or restrict use or copying of any Content or enforce limitations on use of the YouTube Website or the Content therein."

"The problem about Theora is that it's out-dated, inefficient, and basically sucks compared to other modern codecs. If Theora truly replaced H.264 implementations in online video services like YouTube, it will actually be a step backwards, in terms of quality at least."

You definitely have a point there. Theora is definitely inferior to H.264....assuming your machine has enough processing prowess to decode it properly. But...even so, is that saved bandwidth worth the dropping of your freedom? To me the only other viable alternative to Theora, at this point, is Xvid...but I'm not sure how legally sound they are with Divx and such potentially having patents and whatnot to use against them. Also don't forget Theora started out as On2's VP3, and VP6 is much higher quality...so in theory the open source community should be able to do the same with Theora. In the end, the new breed of wavelet based codecs...which I believe Dirac is IIRC, will even surpass H.264 at some point. At least we have something better than Mpeg2 or Sorenson ;)

"Silverlight?"
shmilverlight :P ...with the progression of HTML5, DHTML/Javascript, and SVG technologies like Silverlight and Flash will no longer be necessary.

....on DRM:
It's funny, the number one reason Flash video has exploded into the most popular video distribution format is because so many content providers "think" it has some sort of DRM in it, but if you have the right plugins it's just as easy to download and save to your hard drive as any other format....but since in the majority of browsers you can't Right Click and choose Save File, the content providers seem content...for now.

As long as Firefox's built in Ogg Vorbis & Theora playback goes off without a hitch, and if just a handful of other browsers....like say Opera and Chrome? follow suit, Ogg will finally go mainstream I hope...

"You definitely have a point there. Theora is definitely inferior to H.264....assuming your machine has enough processing prowess to decode it properly. But...even so, is that saved bandwidth worth the dropping of your freedom? To me the only other viable alternative to Theora, at this point, is Xvid...but I'm not sure how legally sound they are with Divx and such potentially having patents and whatnot to use against them. Also don't forget Theora started out as On2's VP3, and VP6 is much higher quality...so in theory the open source community should be able to do the same with Theora. In the end, the new breed of wavelet based codecs...which I believe Dirac is IIRC, will even surpass H.264 at some point. At least we have something better than Mpeg2 or Sorenson ;)"

well, my point is exactly that, we should be able to keep our freedom AND quality of life (and video :P ), instead of dropping either one of them. That's why I said the FOSS community should aim for something "free and superior", not "free but inferior". The FOSS movement will only truly succeed when they provide better quality products to their patented commercial counterparts, like Firefox, not by trying to convince people to use a lower quality product just because it's free and not patented. I'd say Theora will only go mainstream after it can develop into a superior codec.

About VP3 vs. VP6, since VP6 is a completely new codec from VP3, and VP6 is not exactly developed from VP3, so the existence of VP6 can't really say anything about Theora's possible future if the open source community just want to build and improve upon the VP3 codebase instead of developing something new.

And Opera and Firefox started to work on built-in Ogg Vorbis & Theora support at around the same time :
http://dev.opera.com/articles/view/a-call-for-video-on-the-web-opera-vid/

So you can count on Opera's support of Theora.

Chrome... I'm not so sure as it uses WebKit... but then even if Opera and Chrome both support Theora, considering their miniscule marketshare, it won't really help Ogg's adoption much anyway.

"GIF and JPEG were somewhat concerning half a decade ago or more, but are no longer"

But Asa, Firefox/Firebird/Phoenix DID support GIF and JPEG since more than half a decade ago, right? So your point in saying "they are no longer patented since half a decade ago" is moot here.

Oh, really? YouTube thinks that it does. Here is how it is spelled out in the Terms of Service: "You agree not to circumvent, disable or otherwise interfere with security-related features of the YouTube Website or features that prevent or restrict use or copying of any Content or enforce limitations on use of the YouTube Website or the Content therein."
That's called a "login page", not DRM.

Mozilla is in a very exciting position during the transition to HTML5. Whether anyone else likes it or not, Firefox is the second most popular web browser on the internet. The decisions Mozilla makes with future versions of Firefox are far more influential than what Apple, Google nor Opera choose to do. The only one with more power is Microsoft, and obviously they'll never officially support Ogg, Dirac or anything but their own proprietary codecs until the market absolutely forces them to do otherwise.

Apple, Nokia and others can try to fight Ogg's progression on the web, but hopefully between Mozilla and the rest of the open source community we can make Ogg and other open source codecs just as ubiquitous as Mpeg/Quicktime/WindowsMedia/etc... honestly, if we can get websites using Ogg formats even half as much as flash is used currently it can easily be considered a success. sıcak videolar Unfortunately, in this stage of the battle, making 3rd party plugins for Quicktime and Windows Media Player are some of our best bets to help the transitions along.

Aza and the rest of the Mozilla team: thank you very much for what you're doing here. I'm thrilled to see audio and video functionality becoming a standard, built-in part of my favorite browser ;)

Well, Firefox 3.5 will have it. Chrome 3 will have it.

WebKit is modular. It can use FFmpeg, QuickTime, or something else altogether for video engines.

Internet Explorer.... Hmm, you better believe it won't have it! And if it did, it would almost certainly integrate with DirectShow...

Opera 10 I believe will have it.

The QuickTime Components from Xiph are very outdated. Somebody experienced in Mac development needs to go and do something about it, because those codecs are OLD.

The same is true for the DirectShow codecs. Shortly after Xiph absorbed the project, those codecs haven't been updated in a YEAR. How are the thusnelda branches supposed to be tested if nobody is updating the codecs to use them?

what about Sun's OMS Video? http://www.robglidden.com/2008/12/oms-video-draft/

It aims to be a better royalty free video codec than Theora, and the audio part of Sun's Open Media Commons (http://www.openmediacommons.org./) container uses Vorbis, so overall that would be a complete and pretty capable solution.

What do you think?

what about Sun's OMS Video? http://www.robglidden.com/2008/12/oms-video-draft/

It aims to be a better royalty free video codec than Theora, and the audio part of Sun's Open Media Commons (http://www.openmediacommons.org./) container uses Vorbis, so overall that would be a complete and pretty capable solution.

What do you think?

Here's another link with OMS information: http://blogs.sun.com/openmediacommons/entry/oms_video_a_project_of

So you say h264 is superior. Please define superior in REAL Life terms.

From my experience with the latest encoders theora looks good and the colour was more accurate to real life footage taken by camera. (not darkened).

If you can't see anything wrong with it, then how suited is it.

You see a mobile mhz is not equivelent to a non reduced instruction set mhz (pc) and you should notice that some sites download, but do not render immediately, and that's just web code.

It uses more bandwidth but less CPU, which means it's more suited to mobile devices on >3G then and is perfectly acceptable everywhere else, except maybe super high resolution, which it's not meant for and which the web can't deliver. (e.g. Xbox say you get HD movies, but how come they are less than a quarter of the size??? and have to be downloaded, before played)

The increase in bandwidth is the dead giveaway. Video codecs are all about efficiency, and Ogg is simply not as efficient as H.264. It's arguably a tiny bit better than MPEG-2, but still falls far behind modern codecs because it lacks support for modern encoding techniques. The simple fact is that you can encode an H.264 video with substantially less bitrate than Ogg.

The argument about CPU usage is silly. Would you argue that DOS is better than current releases of Linux, OS X, or Windows because it utilizes less CPU? New operating systems are factors of magnitude more complex than old OS's, and the CPU usage reflects that. The new codecs require much more CPU because they do so much more. They also scale easily, meaning that H.264 profile can dumb itself down for low utilization requirements if you encode it that way.

For distribution sites like YouTube, bandwidth is king. The client does the decoding work, and any additional work they can offload to the client makes good business sense, and reduces internet traffic. It's a win/win for business, and deluding yourself that a 10 year old codec has a chance against today's codecs is silly.

I agree with the poster above. Their efforts would be better spent supporting a newer codec with real teeth, not something a decade past it's prime.

Monthly Archives