Solving the codec problem

This is a response to John Dowdell’s Put down the Flavorade and slowly back away…. post, which is itself a reply to Tristan Nitot’s Making video a first class citizen of the Web. Here it is, starting with a quote from John Dowdell’s post:

I don’t appreciate how Tristan avers menacingly to “proprietary plugins” and “patented codecs” without ever making a strong, sensible case. It’s fear-mongering. It’s unnecessary. Beneath his rap is the assumption “Ogg Theora (aka On2 VP3) will solve all needs”, and that blanket assertion still needs a strong defense. You can’t say “patented codecs are bad” without addressing why the world prefers modern codec technology. It’s religious gaming, not technical discussion.

It’s unfortunate that Tristan didn’t describe the problems with proprietary codecs. Fortunately, I’ve already compiled such a list. It’s the “Problems with patented codecs” section of The codec dilemma. For reference, I will reprint the list of problems with patented codecs here:

  • Consumers and distributors have to pay for patented codecs in order to use them.
  • It is difficult to know whether a particular audio or video file you have will work with someone else’s computer or device. They may not have the same codecs as you or they may not have any patented codecs at all.
  • If a software developer chooses to write his or her own software that can read a codec such as MP3, that person could not give that software away for free even if they wanted to because they have to pay the patent holder.
  • If a person chooses not to purchase an operating system such as Windows or Mac OS that includes patented codecs, but instead downloads a free operating system such as Ubuntu, they cannot view any content using patented codecs unless they obtain the codecs illegally or purchase them through software such as Fluendo Complete Playback Pack, which can be prohibitively expensive.
  • Related to the above point, if a company is designing a new product but wants to keep costs as low as possible and also provide entirely redistributable software to its users, it cannot include any patented codecs. This is an issue for projects such as One Laptop per Child. For more information see their page on restricted formats and standards for software freedom.

With these problems in mind, there is one advantage that proprietary codecs might currently have over royalty-free codecs. They are more well-supported by the dominant operating systems, video editing tools, Adobe Flash Player, and some hardware. This is perhaps why, as John Dowdell says, the “world prefers modern” (he means “proprietary”) “codec technology”: installed user base. Fortunately, the situation is changing. Mozilla will soon distribute Theora support to millions of users (in Firefox 3.5), which will make Theora a much more compelling platform for video.

From what I can tell, you can virtually always replace a proprietary codec with one of Theora or Dirac and get comparable size and video quality. This is why I believe we can start the transition away from proprietary codecs right now. If you believe that royalty-free codecs are inferior to proprietary codecs on a technical level, please provide some information to support that.

Adobe should add royalty-free codecs, such as Theora and Dirac, to Flash Player and their Flash creation tools. This would be very beneficial for both Adobe and the general public. Adobe could gradually migrate developers away from proprietary codecs, eventually allowing Adobe to drop proprietary codecs from Flash Player altogether, eliminating the licensing fees Adobe has to pay. The general public would benefit through lower costs on consumer electronics and video tools, whose price would no longer include a proprietary codec royalty.

I can’t see a reason why Adobe wouldn’t include royalty-free codecs. When will these be added to Flash Player? If they won’t be, why not?

6 Responses to “Solving the codec problem”


  • Denver,

    I hadn’t see JD’s post, otherwise I would have answered myself (not sure if he would listen though, because it’s always harder for a man to understand something when he’s paid not to).

    Thanks a lot for this response, it’s spot on. The world will indeed be better place with Open, royalty-free codecs, and you demonstrate it.

    Frankly, proprietary codecs where good until we had decent Open codecs such ad Dirac and OggTheora. They solved a problem for a while, but it’s time for them to go away and be replaced by Open codecs, which enable more people to do more things in more ways.

    –Tristan

  • You know what would have even more influence than the Flash Player? Quicktime. Quicktime powers all the media (or most of it) on Apple tech, and is installed on a very large number of Windows machines as well.

  • “Adobe should add royalty-free codecs, such as Theora and Dirac, to Flash Player and their Flash creation tools.”

    Hi Denver, that’s an option for the future (and I’m not privy to the decision-making process), but your essay fails to say how most would consider the additional runtime download costs to be worthwhile. Would content providers move back beyond VP6 to a VP3 codebase, and if so, why? These are the types of questions that would have to have compelling answers in order for the world to change to meet your personal goals.

    “Adobe could gradually migrate developers away from proprietary codecs….”

    That isn’t the way the world works. People make their own best choices. There’s a bit of authoritarianism and mind-control in the world, but we’re fighting to minimize such faith-based influences.

    “The general public would benefit through lower costs on consumer electronics and video tools, whose price would no longer include a proprietary codec royalty.”

    The consumer cost structure would likely be in higher download times and greater CPU costs for lesser-quality video. Content creators would have significant costs in changing their existing workflow and production tooling. Easy to see the costs, hard to see any benefits.

    Tristan wrote: “I hadn’t see JD’s post, otherwise I would have answered myself (not sure if he would listen though, because it’s always harder for a man to understand something when he’s paid not to).”

    Spoken like someone who gets paid to aver things without backing them up.

    Go read and comment at my complaint at your continued intemperate, unreasoned speech. It ain’t about codecs, it’s about insinuation being a weak and despicable debating technique.

    jd/adobe

  • Would content providers move back beyond VP6 to a VP3 codebase, and if so, why? These are the types of questions that would have to have compelling answers in order for the world to change to meet your personal goals.

    Theora started as VP3, but has improved by leaps and bounds since then. Have you actually compared Theora and VP6?

    Even if VP6 or H.264 is better than Theora, there are other royalty-free codecs, which I mentioned. Compare H.264 or VP6 with Dirac; I think the quality will be comparable.

    I don’t think these are just my personal goals. I’m pretty sure all consumers want to pay less and have more flexibility in choosing video tools.

    The consumer cost structure would likely be in higher download times and greater CPU costs for lesser-quality video.

    Again, you’ve made a claim without any supporting evidence. Please provide references that show this is the case.

    You also ignore the fact that Adobe added H.264 support to Flash Player in 2007. If Adobe can make room for a royalty-ridden codec like H.264, surely it can make room in Flash Player for some royalty-free codecs like Theora and Dirac.

    Content creators would have significant costs in changing their existing workflow and production tooling. Easy to see the costs, hard to see any benefits.

    You must have missed the 5 big bullet points in my post. It’s easy to see the benefits of switching to royalty-free codecs (ie. they lack the problems proprietary codecs have). Sure, there will be small switching costs. But the benefits outweigh the significant costs of maintaining our dependence on proprietary codecs.

    Perhaps we should all have stuck with IE because switching to a better browser incurred a small switching cost. How awful would the web be then?

  • Good background on ongoing codec comparisons is with search term “codec shootout”.

    Theora improvements over the 2001 VP3 donation seem to have been minimal:
    http://en.wikipedia.org/wiki/Theora#Performance

    H.264 had a _lot_ of grassroots support before Adobe took the gamble on licensing and distributing it. Particularly, many production shops had already moved to H.264 workflows.

    And for codecs, it’s companies like Adobe which pay for the consumer use. The big desire for consumer support of patent-unencumbered codecs seems to come from that small minority of toolmakers — not consumers, not content creators. Toolmakers are important, but not so much as consumers and creators.

    jd/adobe

  • Good background on ongoing codec comparisons is with search term “codec shootout”.

    Most of the comparisons I’ve found are over 2 years old, which means they don’t compare implementations of the finalized Dirac spec (released January 21, 2008). The ones I’ve looked at are mainly these:

    http://en.wikipedia.org/wiki/Comparison_of_video_codecs#Freely_available_codecs_comparisons

    The recent comparisons I’ve seen tend to be sloppy in their analysis of Dirac and Theora, failing to provide details on which Dirac implementation they used (Schrödinger or dirac-research) and other test setup information.

    Theora improvements over the 2001 VP3 donation seem to have been minimal:
    http://en.wikipedia.org/wiki/Theora#Performance

    Fair enough; it may or may not have improved since then. I think Theora will work fine for low-bitrate video while Dirac will cover the high-bitrate market.

    H.264 had a _lot_ of grassroots support before Adobe took the gamble on licensing and distributing it. Particularly, many production shops had already moved to H.264 workflows.

    Dirac was used by the BBC during the 2008 Beijing Olympics (see And now, Dirac from the Olympics, a new free codec!). Also, there are two independent, interoperable software implementations of Dirac and at least one hardware implementation. Dirac is already supported by VLC, FFmpeg, and GStreamer. I think these facts show Dirac is ready for wider deployment. For more information, see the Dirac Wikipedia page.

    And for codecs, it’s companies like Adobe which pay for the consumer use. The big desire for consumer support of patent-unencumbered codecs seems to come from that small minority of toolmakers — not consumers, not content creators. Toolmakers are important, but not so much as consumers and creators.

    Perhaps this is where we disagree. Toolmakers are profoundly important. Even though they are a small group compared to consumers and content creators, what they produce impacts everyone far more than what consumers and content creators produce. For example, Firefox has significantly enriched the web by having very competitive standards support, thus forcing Microsoft to keep their browser interoperable and up-to-date. Had it not been for Firefox (the tool) and the royalty-free standards that enabled it, we would be stuck using old standards or, worse yet, depending on IE-only features. AJAX and other useful technologies simply wouldn’t have been developed.

    This is why the tools are important: they define what consumers and content creators can do. Restrict what tool makers can do and you ultimately restrict the features that consumers and content creators will have available to them, limiting what they can create and consume.

Leave a Reply to John Dowdell