263 Comments

Summary:

Google will soon make its VP8 video codec open source, we’ve learned from multiple sources. The company is scheduled to officially announce the release at its Google I/O developers conference next month, a source with knowledge of the announcement said. And with that release, Mozilla — […]

on2videothumb

Google will soon make its VP8 video codec open source, we’ve learned from multiple sources. The company is scheduled to officially announce the release at its Google I/O developers conference next month, a source with knowledge of the announcement said. And with that release, Mozilla — maker of the Firefox browser — and Google Chrome are expected to also announce support for HTML5 video playback using the new open codec.

Google has controlled the VP8 codec ever since it finalized the acquisition of video codec maker On2 Technologies in February. When reached for comment as to its plans, a Google spokesperson told us the company had “nothing to announce at this time.”

The move comes as online video publishers are gravitating toward standards-based HTML5 video delivery, bolstered in part by the release of the iPad. However, that acceptance has been slowed by the fact that the industry has yet to agree on a single codec for video playback, with some companies throwing support behind Ogg Theora and others hailing H.264 as the future of web video.

Google’s YouTube, Microsoft’s Internet Explorer 9 and Apple — through its iPad, iPhone and Safari browser — have all thrown their weight behind H.264, which many believe provides superior picture quality and playback to the Ogg Theora codec. However, a few organizations — including Mozilla — refuse to support H.264 due to potential licensing issues. Whereas Ogg Theora is completely open source, the H.264 codec is managed by licensing body MPEG LA. Even though MPEG LA announced in February that it was extending its royalty-free licensing for web video using H.264 through 2016, that was little consolation for Mozilla and others that are committed to supporting open standards.

The result is a divide between which video format can be viewed in which browser. H.264-encoded HTML5 video can be viewed in Apple’s Safari, Google’s Chrome and in the upcoming Internet Explorer 9 browser from Microsoft. Meanwhile, Ogg Theora playback for HTML5 video is supported by Firefox, Chrome and Opera.

Google hopes to stem that divide by making VP8 open source, providing a high-quality and open alternative to existing codecs. On2 first announced VP8 in late 2008, promising more efficient video compression than other available codecs. At launch, On2 went so far as to claim that it could provide “50 percent bandwidth savings compared to leading H.264 implementations.”

Google’s plans to open-source the codec have been widely expected ever since it announced plans to acquire On2 in August 2009, and speculation intensified after the deal closed. The acquisition even led the Free Software Foundation (FSF) to urge Google to kill Flash by open-sourcing the VP8 codec.

While an open-source VP8 could end concerns about H.264’s licensing issues and Theora’s quality, questions still remain about whether Google can provide a video standard on which everyone can agree. Microsoft only recently announced support for H.264 for HTML5 playback, and has never been quick to adopt open standards. And Apple, which has been the driving force behind HTML5 video and H.264 playback on the iPhone and iPad, might not be keen on the idea of switching up its codec support on those devices anytime soon.

Related content on NewTeeVee: Google TV: Another Reason Open Sourcing VP8 Matters

Related content on GigaOM Pro: What Does the Future Hold For Browsers? (subscription required)

  1. Wow, that is pretty major. I did not think Google would actually go ahead with making VP8 open source, but it will surely have a big impact.

    Nonetheless, major questions still remain. How good will VP8 actually prove to be in widespread application? Will MPEG-LA make good on their threats to enforce H.264 patent claims against Ogg Theora now that it could be a real contender? Can an improved Ogg Theora really do anything at this point to slow down the momentum of H.264 adoption in hardware?

    It will be very interesting to see.

    Share
  2. This seems like it’ll only be really great news if hardware acceleration, specifically for mobile, also comes with it. That seems to be where the biggest news in video is these days, and that’s where there’s a HUGE uphill struggle against h.264. Apple’s got to get on board behind this codec as much as Google may be.

    Share
    1. If it came down to a scrap between Apple and Google, on past form I really wouldn’t fancy Apple’s chances.

      Share
    2. Hardware acceleration is overrated, and the whole discussion about low end devices is a bit pointless, considering that each handset nowadays is a lot faster than bleeding edge PCs 10 years ago. As long as Moore’s law remains intact and batteries are advancing at same speed so far, there is no real need for hardware accelerated VP8. The Apple fanboys always tried to spread this hw accel fud in order to have an excuse for h.264…

      Share
      1. All video decoding on consumer electronics and smart phones has got to be hardware accelerated. It’s not just a processing issue, it’s also a power consumption issue. Dedicated hardware acceleration for video is the only way to decode and encode 1080p with under 1W of power consumption, perhaps even lower than 0.5W power consumption. Which is amazing considering an Intel powered Desktop PC consumes often over 100W in power consumption.

        I think Google must be planning the hardware acceleration accordingly and provide free codecs for that to all manufacturers of consumer electronics.

        Share
      2. It actually is a big deal, because video at the quantity of Youtube can’t just be turned around and scrapped. Youtube has contractual deals to uphold H.264 videos with not only Apple and the iPhone, but also Tivo and a whole class of set top boxes. I’m sure such deals are for a reasonable lifetime of the product.

        I don’t think anyone is trying to make an excuse for h.264. It’s currently the best of line codec. Outside of Apple, it’s pretty much the standard that most all professional videography toolchains are based on. It’s built into Windows 7. A vast bulk of consumer camcorders support it natively. It’s one of the premiere codecs supported in Bluray. It’s supported in Flash. And many of those devices do have hardware acceleration as a major benefit. Like it or not, the industry has settled on H.264, even if the web hasn’t.

        I’ll be first in line to start supporting a royalty-free and license-free codec that’s not h.264, given that it meets or exceeds the quality bar set by h.264. I’m looking forward to it. I think that’s a powerful and wonderful incentive for an open web, and I’m thrilled at the possibility. But everyone’s got to be on board with it, including Microsoft and Apple. Otherwise, it’s not an open standard. It’s an open political grandstanding game, which is where we find ourselves now with Theora vs. H.264. Right now, H.264 happens to meet more of those needs than anything else. Time will tell what VP8 will become, and I’m excited to see it become something important.

        Share
      3. Sherman, the Ramni Tuesday, April 13, 2010

        @garbeam: You obviously don’t understand hardware acceleration and its importance for the new generation of low-end devices (netbooks, smartphones, etc.).
        @Charbax: CUDA and OpenCL will fill that gap real quick, trust me.
        @Kenneth Pardue: No one cares about Microsoft’s opinion. They only follow the current trend. Apple and MPEG-LA are the only ones that can pose a serious barrier, but 2 against the great majority of video-on-the-Web players? No royalties until 2016 vs. no royalties FOREVER? A total curbstomp.

        Share
      4. @Kenneth, your definition of an “open standard” requires Microsoft and Apple to be in on it?? I don’t think that would fly with the ISO…

        Share
      5. I’m not talking about ISO sanctioned standards, I’m talking about practical standards. If Apple and Microsoft aren’t on it, then it’s not standard for the web. I haven’t used Internet Explorer for nearly 6 years and rahter loathe it, but I do still recognize that they’re 60% of the market. Firefox+Chrome is what, 28%? Like it or not, that’s significant.

        Share
      6. Ah. Then the sentence should’ve been “otherwise it’s not a de facto standard”, not “otherwise it’s not an open standard”.

        It’s an important difference.

        Share
    3. I spoke to a few of the Theora people at Libre Planet and asked them about hardware acceleration since apple/h.264 fanboys always seem to bring it up. They indicated that there’s nothing stopping someone from hardware accelerating Theora, companies just aren’t doing it.

      Share
    4. Considering that many of the mobile devices aren’t using specific hardware and using advanced DSP’s to do the decoding (Each and every one of those iPhones and Android phones don’t HAVE dedicated h.264 hardware, they’ve got a TI DaVinci class DSP in them (Anything with an OMAP2/3 that does media has one…)) this is a specious argument. Much of the space doesn’t do dedicated h.264 or MPEG2 hardware except at the low-end of the spectrum- mainly because they’ve got so many differing tasks that it’s actually cheaper to just use a DSP on the SoC.

      Share
  3. [...] http://newteevee.com/2010/04/12/google-to-open-source-vp8-for-html5-video/ April 12th, 2010 | Tags: Google, On2, opencu, opensource, VP8 | Category: streaming [...]

    Share
  4. This is big. The release and success of this open-source codec could potentially kill Ogg Theora.

    Share
    1. Was Ogg Theora ever alive?

      It never got any major attention, and if it weren’t for Mozilla backing it, it would have remained in the dark.

      Share
    2. This is Ogg Theora 2.0 basically. Ogg Theora is based on On2 VP5 codec, Ogg Theora 2.0 would be based on the On2 VP8 codec.

      Share
      1. Just wanted to correct a slight error there, Charbax. Ogg Theora is based on On2 VP3 codec, not the VP5. VP3 was released to the public domain back in 2002.

        I totally agree with you that the open-sourcing of VP8 won’t be the end of Theora since I’m sure they’ll take the code and build on it like the built on VP3.

        Share
      2. Yes, implementing VP8 as Ogg Theora 2.0 would be an excellent idea.

        One of the websites that use Ogg Theora the most is Wikimedia Commons. It is the only video codec allowed there.

        Share
    3. Uh… You DO realize that this is merely the latest generation in the class of codecs that Theora came from, namely VP3?

      Share
  5. Good. To me, this has two potential, positive outcomes for the majority of the “netizens”: killing Flash on the web-based video/audio playback scene, and making Theora useless in favor of something better (and everything is better than Theora).

    Share
    1. Are you sure? Don’t come with your lame affirmations if you do not prove. You must be a very biased person or very frustrated. So sorry about you thinking.

      Share
      1. Grover Saunders Tuesday, April 13, 2010

        Wow, you really showed him how to post a reasoned and fact-based past. It could have shortened to “So’s your face.”

        Share
    2. MPEG2 isn’t.

      Share
  6. “Microsoft only recently announced support for H.264 for HTML5 playback, and has never been quick to adopt open standards.”

    Adopting an open standard like the HTML5 video tag is the hard part. Plugging in additional codecs is probably a breeze.

    This likely doesn’t have to be an either or for any browser maker.

    Flash didn’t drop Sorenson Spark (h.263) when they added VP6 and they didn’t drop VP6 when they added H.264. Flash also contains a Screenshare codec that is much lower quality than any of those others but is still extremely efficient for screencasting.

    The same holds true for audio. Speex isn’t nearly as high quality as AAC, but it’s got size and latency advantages that make it kick AAC’s ass for VoIP. And guess what? Flash includes both.

    Different codecs have different quality, size, and cpu features. No one codec will solve every problem and the overhead of carrying an extra “free” codec is minimal.

    Microsoft should just add Theora and VP8 to their H.264 plans. Apple should follow suit.

    Share
    1. “Plugging in additional codecs is probably a breeze.”

      Technically yes. Politically ? That too a google OWNED standard ? Will Big Apple Brother ever look above the spite ?

      Share
  7. Just because something is open source doesn’t mean it’s free from patent issues. What does the mpeg-la say about VP8?

    Plus, if Apple doesn’t support it then it doesn’t really matter. Seems like everything will be h.264 no matter what anyone else tries.

    Joe.

    Share
    1. This is the key point. h.264 can be “open source” too — there are reference implementations, and of course x264. It’s the patent licensing that hinders adoption, not the unavailability of source code.
      With h.264, the patent costs are high but knowable (at least for the next few years). Ogg Theora may be unencumbered by patents, but until it’s adopted by someone with deep pockets who would be a good target for a patent troll or MPEG-LA, its patent status is uncertain.
      As far as VP8 goes, we’ll have to see how things unfold in terms of patents.

      Share
    2. That, my friend depends on a handful of things.

      VP3’s been about for a while.
      VP8’s not been so much so, but enough for what I’m going to get to here.

      If you delay on dealing with an “obvious” infringement, you risk not being able to enforce against that specific infringer. The term in legalese is “Laches” which is latin for delay. Just because it might be covered, doesn’t mean that MPEG-LA, or the rights holders they represent, can enforce against the implementation at this time. It’s definitely too late for them on Theora’s score- years have passed and no court will buy they didn’t know about it. VP8 might have issues, but equally unlikely as the key pieces are under their OWN patents, just like VP3 was.

      Out in the open and nobody’s spoke up. If it’s long enough (and it has been for Theora and most likely with VP8 as well)- that spectre has no force whatsoever.

      Now, Apple can do what they so choose- but if YouTube doesn’t go with h.264 and goes with VP8/Theora, then there’s LITTLE that Apple can do about that other than cut their nose off to spite their face.

      Share
  8. [...] Google to Open-source VP8 for HTML5 Video Google will soon make its VP8 video codec open source, we’ve learned from multiple sources. The company is scheduled to officially announce the release at its Google I/O developers conference next month, a source with knowledge of the announcement said. And with that release, Mozilla — maker of the Firefox browser — and Google Chrome are expected to also announce support for HTML5 video playback using the new open codec. [...]

    Share
  9. Very cool. However, it’s my understanding that it’s going to take quite a few years before hardware will provide efficient decoding for VP8 as found in H.264. Mobile devices with slow CPUs depend on using hardware accelerated video. Also HTML5 video on Safari on a Mac, HTML5 video on IE9 and Flash Player 10.1 (only on Windows right now) all get smooth HD video playback thanks to using the GPU. In all 3 cases they can use the GPU because there’s a H.264 decoder in the hardware. It could take a few years for VP8 to be offered in hardware to get that same GPU performance found with H.264 and then the wait for people to upgrade their hardware.

    Share
    1. Much of the “hardware” decoders are DSP cores programmed to do the task. Each and every one of the “media” or “smart” phones utilize a DSP on their SoC to do the video and audio decode tasks. Ditto much of the media players. It’s cheaper and easier to support multiple codec formats (MP3, WMA, MPEG1/2, MPEG4, h.264, etc- they’re all differing in what “hardware” really means- unless you use a DSP and do the “hardware” in a specialized stream processor.) with a DSP than doing it with discrete parts. There’s component aspects that might lie within the GPUs of desktop machines, but in truth, you can just as easily (and they’re beginning to do it in the Linux space…) dedicate a GPGPU thread or two to doing DSP work- and do it there as well.

      “Lack of hardware support” is more of an excuse than much of anything else- it’s not valid for a large range of devices that this would be relevant for, including all those iPhones, Droids, etc.

      Share
  10. [...] full post on Hacker News If you enjoyed this article, please consider sharing it! Tagged with: Google • [...]

    Share

Comments have been disabled for this post