WebM? VP8? What the hell is going on!?

Chances are that if you have not heard of WebM by now, you likely know that Google had plans to buy On2, or that speculators hoped Google would release Theora’s younger brother, On2’s VP8, to the public. And, what you know, that’s exactly what happened.

Google united most of the industry around this plan and the result is WebM, a royalty-free, BSD-licensed mixture of the Vorbis audio and VP8 video codecs on a modified Matroska container. For the discernible user, it tastes raw with a hint of spiciness, and you soon realize you haven’t tasted anything like it before.

In a surprising turn of events, the whole industry (barring Apple, but they don’t count anyway) embraced the newcomer, apparently solving the whole video on the web debacle that’s been going on in the last years. So, expect browser and even hardware support coming really soon.

Of course this is all new and the paint’s still not exactly dry. Documentation is a mess, encoder quality is terribly optimized, and there’s a whole lot of work ahead for those involved, but all things considered? This is likely the best outcome possible. Even Adobe, in an (what I assume) attempt to screw Apple for the whole “die Flash, die” chapter, sided with WebM and will support it in Flash.

By the way, while we are ahead, shouldn’t we try addressing the FUD certain people of the h264 faction are already spreading? Certainly. Less glamorous comments by those who should know better must always be addressed before they spread like an infection. Did you know some people still take that silly claim of “Youtube in Theora would take all of the Internet’s bandwidth” seriously? Why yes, some people not only believe that but will always take a moment of their monotonous day to quote it during Internet discussion.

Some people also believe in a Flying Spaghetti Monster.

But where was I? Oh, right. I was debunking the commentary from a certain developer without any knowledge of software patents who decided he knew better than On2 when it came to waltz around MPEG-LA’s patent pool and now claims VP8 must surely be infringing on something or another. Or rather, in a Glenn Beck logical fallacy: “I’m not saying it does, but it didn’t prove me wrong either”.

He then went on to compare years of h264 optimizations against the upstart and surprisingly, considering how unbiased someone whose livelihood depends on nothing replacing h264, came out in favor of the establishment. He forgets, however, that Google has enough money to buy whole countries hire all the engineers they could possibly want to improve VP8’s encoder quality in record time.

Anyhow, WebM’s adoption of Vorbis, just as this one introduced drastic surround optimizations in its code base a couple of months ago, is a great sign of trust in Xiph’s work. But, what does it mean for Theora now that it has apparently been left out of the Web Video Wars? A good question, but with a simple answer: the Theora community will likely find more niches that Theora can tap. Considering Theora is the most lightweight video format of all the modern offerings that’s a no-brainer.

And the Ogg container? Now with the recent addition of an index to improve seeking in corner cases, nothing much is stopping it from being employed in more and more situations where a container as malleable as Ogg is needed.

5 Responses to “WebM? VP8? What the hell is going on!?”


  1. 1 Robin

    Hey, the Flying Spaghetti Monster is real and it’s very happy about WebM, so don’t anger it now.

  2. 2 Spudd86

    Eh the comments that really need addressing from the x264 developers post are the ones about the spec (and the claim that it won’t ever be able to match h264’s high end profiles) and the stuff about the way bugs in On2’s implementation are now effectively part of the spec…

    He’s probably right that it would have resulted in a better codec if they had spend a year or two letting people at Xiph, Google, etc. revise it and shake out issues (like the kind of things you get from having only the one implementation…)

    Look at what happend with Theora On2’s original code and spec sucked Xiph made it suck much, much less… VP8 could probably benifit from the same treatment… but it is too late for that to be part of WebM…

    The most obviuos problem with his post is that he states that the assembly code was ‘writen by retarded monkeys’ then he uses the speed of the implementation against it and says it probably won’t get much faster…

  3. 3 Tim

    Yeah I’d like to see a bit more *actual* debunking!

    And you have to admit, “VP8 is brand new” can’t possibly be used as an excuse for anything - VP8 is based on years of VP7, VP6, and so on.

  4. 4 Michael

    As it’s Ogg and Matroska, can we re-title a WebM file as ‘.mkv’ and get people used to it as they probably -have- seen an MKV video before?

  5. 5 Ivo

    In theory renaming .webm to .mkv shouldn’t cause much of an issue. Some players hardcoded to assume Matroska using .mkv as file extension only carries h.264 video may have issues, but those aren’t the majority. File extensions are an old relic. WebM is a clever way of saying Matroska with VP8 and Vorbis.

Leave a Reply