Blog Post

Did Google Just Kill Ogg Theora?

Ever since we broke the news earlier this week that Google is going to open source its VP8 video codec at its Google i/O event next month, speculations have been abounded as to what this means for Ogg Theora, the video codec of choice of open source advocates and free software developers alike.

Theora is currently supported by the Mozilla foundation, whose Firefox browser utilizes the format instead of H.246 for HTML5 video playback, and the Wikimedia foundation, which is planing to use the codec for its upcoming Wikipedia video roll-out. However, Google and others have been skeptical of Theora. So is Google going to kill Ogg Theora by open sourcing a superior video codec?

Talk about Theora and VP8, and there’s no way to avoid a little lesson in video codec genealogy: Ogg Theora is based on an erstwhile proprietary video codec called VP3.2, which was developed by a little company called On2 Technologies. On2 introduced VP3.2 in August of 2000, originally with the idea in mind to optimize TV quality video broadcasts for users with as little bandwidth as 200kbps. On2 released a successor dubbed VP4 less than a year later and announced in August of 2001 to open source VP3.2. It took a little more back and forth between open source advocates and the company, but eventually, VP3.2 became Ogg Theora. On2 meanwhile continued to develop new codecs, reaching its 8th generation with VP8, which was announced in September of 2008.

Long story short: VP8 came out eight years after VP3.2, eight years in which much happened in the online video world. Consumers got increasingly faster broadband connections, video hosting sites moved towards HD, and codec developers figured out a whole lotta tricks to improve things like HD streaming. That’s why some have been concerned that Theora isn’t up to competing with H.264 for online video. One of the most prominent skeptics is Google’s Open Source Programs Manager Chris DiBona, who said last year that it would need “substantive codec improvements” before Theora could power a site like YouTube.

Others have been more optimistic about Theora. Wikipedia has started to host Theora files, and Wikimedia Foundation head of Communication Jay Walsh told me in January that the site plans a wider roll-out of video based on the format in the near future. I caught up with him this week to see how these plans are affected by Google open sourcing VP8, and he said that his organization would be open to host multiple open video formats, just as it is now supporting a number of patent-free image formats. “Ultimately this isn’t so much about switching formats as it is about making more options available for more web users”, he added.

Ben Moskowitz from the Open Video Alliance echoed this sentiment, proclaiming: “Theora is here to stay.” He added that Firefox and Chrome would likely support VP8 as well as Theora, but was also enthusiastic about VP8’s potential. “A royalty-free codec that’s indisputably superior to H.264 will be very disruptive,” Moskowitz said.

The most revealing answer I received about Theora’s future came however from Christopher “Monty” Montgomery, the founder of the Foundation, which is the driving force behind Theora. Montgomery told me that he couldn’t specifically comment on our article, only to state: “I think it’s important to repeat that we think open sourcing VP8 is a great thing, a big deal,
and we’re all for it.” And asked by someone on a mailing list whether the news meant “an end for Theora,” Montgomery replied: “Maybe. Unlikely.”

Montgomery is right. It’s unlikely that open sourcing VP8 is going to kill Theora. There will still be a small but dedicated community supporting the format, and there are going to be cases when it actually makes sense to use Theora and not VP8. What it will kill however, is the notion that Theora could one day become the standard of the HTML5 video web. For that, it would need to be a codec that’s superior to existing commercial solutions, and Theora just never was up to that challenge.

Image courtesy of (CC-BY SA) Flickr user llimllib.

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

31 Responses to “Did Google Just Kill Ogg Theora?”

  1. Was XVID/DIVX killed when H.264 came out? no

    ogg theora may be positioned like XVID/DIVX and VP8 may be positioned like h.264….

    • We won’t know this for sure until VP8 actually comes out, but there have been speculations that Theora could in some instances be less resource-intensive. This might be an issue of trade-offs. If you don’t care as much about either the look or the bitrate, but need a smaller footprint, going with the older and simpler solution might actually be better.

      • Low resources usage is becoming less and less important, particular with the rate of development of the traditionally lowest-powered computing devices like smartphones. Also, considering that IE will probably never support Theora, the niche that Theora will occupy may be a very, very small one.

  2. @ Emmanuel – rather ‘killing Ogg Theora” is meant not as “will the Ogg Theora project disappear” but more “will anyone care about it if VP8 goes open source?” I agree that the “Will A kill B” formulation is annoying, but the underlying question is valid and I can’t see any reason for Theora to continue to be relevant (or H.264) if VP8 is all that and open.

  3. Emmanuel Mwangi

    It’s clear that the previous article about Google open sourcing VP8 is making the rounds and is getting a bit of press. I’m sure you guys have been pressed to make followup articles but this headline is just mis-informed linkbait.

    Google can’t “kill” Ogg Theora any more than Microsoft could kill Linux or Sun/Oracle kill MySQL.

    With Google’s might behind it–Youtube’s influence and the fleet of lawyers to defend IP, an open source VP8 would be the next great open video format.

    And you hint to as much but, ever eager to find conflict where this is none, you ask leading questions like will one open source project kill another? The Ogg Theora people are probably thrilled at the prospect of an open source VP8, doubly so from a company like Google for reasons I’ve already said.

    Google should be commended for open sourcing VP8 yet, this article is trying find some kind of dark cloud to put over this event despite the positive response to everyone interviewed and the debunking of the headline buried in the last paragraph. Newteevee and GigaOM should know better and write to a higher standard than this linkbait.

  4. ASide from ideology would there be any reason to continue to use Theora? Even if it forked VP8, what’s the real benefit to that? So yes, this kills Theora unless Google basically turns VP8 over to Xiph and Ogg Theora basically becomes VP8.

    More to the point… doesn’t this kill H.264 as an entity in web video? Sure, there’s legacy content etc, but assuming VP8 gives the same or better quality at the same or smaller file size with similar bandwidth needs…. why would anyone use H.264 to distribute web video streams?

    • there are some things missing from V8 and Ogg Theora (at the moment) that will need to be addressed before H.264 goes away
      – DRM. Like it or not studios are going to require content protection for TV and Movie content. Today that really leaves you with either a Microsoft or (less popular) Flash based solution
      – Adaptive streaming, using fragmented MPEG4 (as in the IIS Smooth Streaming) to allow heuristics and code based variation in bitrate for each client based on their environment

  5. Shai Machnes

    If VP8 is free & better, why is it important to keep Ogg Theora ?

    Of course, people are free to keep working on Ogg, but why ?

    As long as it’s free – let’s just go with the best, whatever it is.

  6. “indisputably superior to H.264”

    Whoo, whoo there, Ben Moskowitz, jump off your stampeding horse. Stop counting beans and pennies and go get some “hands-on” with these things before you talk.

    • Ben didn’t actually say VP8 is indisputably superior to H.264. He says, “A royalty-free codec that’s indisputably superior to H.264 will be very disruptive,” which is correct.

      To be sure, he probably should have used the term ‘would’ instead of ‘will’.

  7. Theora stands a chance as a bread-and-butter codec for low- to mid-quality internet streaming. Honestly, who cares whether a quickly consumed and instantly forgotten video feed is top-notch quality.

    For archieval puposes Theora is simply so much outperformed by other codecs that nobody should really be using it. H.264 plays in a different league. Honestly, would you use Linux if it ran slow, crashed, didn’t have multitasking? No, me neither.

    If we are looking for a competitive codec we shouldn’t be looking elsewhere – VP8 might be good, the BBC’s open-source ‘Dirac’-codec might be better.

    • I’ve been wondering why everyone’s ignoring Dirac. It’s free and open source, comparable to H.264, and there have actually been hardware decoders built for it. I’d love to see VP8 get opened up too, but really, hasn’t Dirac already solved this problem?

  8. Janko,

    Did you know Monty from back in the day, when he was working with us at the Green Witch Internet Radio?

    As long as Monty is on the case, there is no way to count Theora out.

    Do not ever take Monty saying that he cannot comment on something to mean that there is not a plan. ;-)

    Clearly what would make the most sense would be for Google to hire Monty to take the lead on their video codec work.

    All the best,

    Brian Zisk

    • Hey Brian, Green Witch – that brings back memories. When was that, 2000?

      You’re right, it might be smart for Google to use the talent behind Theora. Then again, they just spent lots of money to get some talent through their On2 acquisition.