Mozilla

Articles by Brendan Eich

Sort by:

View:

  1. Video, Mobile, and the Open Web

    [Also posted at brendaneich.com.]

    I wrote The Open Web and Its Adversaries just over five years ago, based on the first SXSW Browser Wars panel (we just had our fifth, it was great — thanks to all who came).

    Some history

    The little slideshow I presented is in part quaint. WPF/E and Adobe Apollo, remember those? (Either the code names, or the extant renamed products?) The Web has come a long way since 2007.

    But other parts of my slideshow are still relevant, in particular the part where Mozilla and Opera committed to an unencumbered <video> element for HTML5:

    • Working with Opera via WHATWG on <video>
      • Unencumbered Ogg Theora decoder in all browsers
      • Ogg Vorbis for <audio>
      • Other formats possible
      • DHTML player controls

    We did what we said we would. We fought against the odds. We carried the unencumbered HTML5 <video> torch even when it burned our hands.

    We were called naive (no) idealists (yes). We were told that we were rolling a large stone up a tall hill (and how!). We were told that we could never overcome the momentum behind H.264 (possibly true, but Mozilla was not about to give up and pay off the patent rentiers).

    Then in 2009 Google announced that it would acquire On2 (completed in 2010), and Opera and Mozilla had a White Knight.

    At Google I/O in May 2010, Adobe announced that it would include VP8 (but not all of WebM?) support in an upcoming Flash release.

    On January 11, 2011, Mike Jazayeri of Google blogged:

    … we are changing Chrome’s HTML5 <video> support to make it consistent with the codecs already supported by the open Chromium project. Specifically, we are supporting the WebM (VP8) and Theora video codecs, and will consider adding support for other high-quality open codecs in the future. Though H.264 plays an important role in video, as our goal is to enable open innovation, support for the codec will be removed and our resources directed towards completely open codec technologies.

    These changes will occur in the next couple months….

    A followup post three days later confirmed that Chrome would rely on Flash fallback to play H.264 video.

    Where we are today

    It is now March 2012 and the changes promised by Google and Adobe have not been made.

    What’s more, any such changes are irrelevant if made only on desktop Chrome — not on Google’s mobile browsers for Android — because authors typically do not encode twice (once in H.264, once in WebM), they instead write Flash fallback in an <object> tag nested inside the <video> tag. Here’s an example adapted from an Opera developer document:

    <video controls poster="video.jpg" width="854" height="480">
     <source src="video.mp4" type="video/mp4">
     <object type="application/x-shockwave-flash" data="player.swf"
             width="854" height="504">
      <param name="allowfullscreen" value="true">
      <param name="allowscriptaccess" value="always">
      <param name="flashvars" value="file=video.mp4">
      <!--[if IE]><param name="movie" value="player.swf"><![endif]-->
      <img src="video.jpg" width="854" height="480" alt="Video">
      <p>Your browser can't play HTML5 video.
     </object>
    </video>
    

    The Opera doc nicely carried the unencumbered video torch by including

     <source src="video.webm" type="video/webm">
    

    after the first <source> child in the <video> container (after the first, because of an iOS WebKit bug, the Opera doc said), but most authors do not encode twice and host two versions of their video (yes, you who do are to be commended; please don’t spam my blog with comments, you’re not typical — and YouTube is neither typical nor yet completely transcoded [1]).

    Of course the ultimate fallback content could be a link to a video to download and view in a helper app, but that’s not “HTML5 video” and it is user-hostile (profoundly so on mobile). Flash fallback does manage to blend in with HTML5, modulo the loss of expressiveness afforded by DHTML playback controls.

    Now, consider carefully where we are today.

    Firefox supports only unencumbered formats from Gecko’s <video> implementation. We rely on Flash fallback that authors invariably write, as shown above. Let that sink in: we, Mozilla, rely on Flash to implement H.264 for Firefox users.

    Adobe has announced that it will not develop Flash on mobile devices.

    In spite of the early 2011 Google blog post, desktop Chrome still supports H.264 from <video>. Even if it were to drop that support, desktop Chrome has a custom patched Flash embedding, so the fallback shown above will work well for almost all users.

    Mobile matters most

    Android stock browsers (all Android versions), and Chrome on Android 4, all support H.264 from <video>. Given the devices that Android has targeted over its existence, where H.264 hardware decoding is by far the most power-efficient way to decode, how could this be otherwise? Google has to compete with Apple on mobile.

    Steve Jobs may have dealt the death-blow to Flash on mobile, but he also championed and invested in H.264, and asserted that “[a]ll video codecs are covered by patents”. Apple sells a lot of H.264-supporting hardware. That hardware in general, and specifically in video playback quality, is the gold standard.

    Google is in my opinion not going to ship mobile browsers this year or next that fail to play H.264 content that Apple plays perfectly. Whatever happens in the very long run, Mozilla can’t wait for such an event. Don’t ask Google why they bought On2 but failed to push WebM to the exclusion of H.264 on Android. The question answers itself.

    So even if desktop Chrome drops H.264 support, Chrome users almost to a person won’t notice, thanks to Flash fallback. And Apple and Google, along with Microsoft and whomever else might try to gain mobile market share, will continue to ship H.264 support on all their mobile OSes and devices — hardware-implemented H.264, because that uses far less battery than alternative decoders.

    Here is a chart of H.264 video in HTML5 content on the Web from MeFeedia:

    MeFeedia.com, December 2011

    And here are some charts showing the rise of mobile over desktop from The Economist:

    The Economist, October 2011

    These charts show Bell’s Law of Computer Classes in action. Bell’s Law predicts that the new class of computing devices will replace older ones.

    In the face of this shift, Mozilla must advance its mission to serve users above all other agendas, and to keep the Web — including the “Mobile Web” — open, interoperable, and evolving.

    What Mozilla is doing

    We have successfully launched Boot to Gecko (B2G) and we’re preparing to release a new and improved Firefox for Android, to carry our mission to mobile users.

    What should we do about H.264?

    Andreas Gal proposes to use OS- and hardware-based H.264 decoding capabilities on Android and B2G. That thread has run to over 240 messages, and spawned some online media coverage.

    Some say we should hold out longer for someone (Google? Adobe?) to change something to advance WebM over H.264.

    MozillaMemes.tumblr.com/post/19415247873

    Remember, dropping H.264 from <video> only on desktop and not on mobile doesn’t matter, because of Flash fallback.

    Others say we should hold out indefinitely and by ourselves, rather than integrate OS decoders for encumbered video.

    I’ve heard people blame software patents. I hate software patents too, but software isn’t even the issue on mobile. Fairly dedicated DSP hardware takes in bits and puts out pixels. H.264 decoding lives completely in hardware now.

    Yes, some hardware also supports WebM decoding, or will soon. Too little, too late for HTML5 <video> as deployed and consumed this year or (for shipping devices) next.

    As I wrote in the newsgroup thread, Mozilla has never ignored users or market share. We do not care only about market share, but ignoring usability and market share can easily lead to extinction. Without users our mission is meaningless and our ability to affect the evolution of open standards goes to zero.

    Clearly we have principles that prohibit us from abusing users for any end (e.g., by putting ads in Firefox’s user interface to make money to sustain ourselves). But we have never rejected encumbered formats handled by plugins, and OS-dependent H.264 decoding is not different in kind from Flash-dependent H.264 decoding in my view.

    We will not require anyone to pay for Firefox. We will not burden our downstream source redistributors with royalty fees. We may have to continue to fall back on Flash on some desktop OSes. I’ll write more when I know more about desktop H.264, specifically on Windows XP.

    What I do know for certain is this: H.264 is absolutely required right now to compete on mobile. I do not believe that we can reject H.264 content in Firefox on Android or in B2G and survive the shift to mobile.

    Losing a battle is a bitter experience. I won’t sugar-coat this pill. But we must swallow it if we are to succeed in our mobile initiatives. Failure on mobile is too likely to consign Mozilla to decline and irrelevance. So I am fully in favor of Andreas’s proposal.

    Our mission continues

    Our mission, to promote openness, innovation, and opportunity on the Web, matters more than ever. As I said at SXSW in 2007, it obligates us to develop and promote unencumbered video. We lost one battle, but the war goes on. We will always push for open, unencumbered standards first and foremost.

    In particular we must fight to keep WebRTC unencumbered. Mozilla and Opera also lost the earlier skirmish to mandate an unencumbered default format for HTML5 <video>, but WebRTC is a new front in the long war for an open and unencumbered Web.

    We are researching downloadable JS decoders via Broadway.js, but fully utilizing parallel and dedicated hardware from JS for battery-friendly decoding is a ways off.

    Can we win the long war? I don’t know if we’ll see a final victory, but we must fight on. Patents expire (remember the LZW patent?). They can be invalidated. (Netscape paid to do this to certain obnoxious patents, based on prior art.) They can be worked around. And patent law can be reformed.

    Mozilla is here for the long haul. We will never give up, never surrender.

    /be

    [1] Some points about WebM on YouTube vs. H.264:

    • Google has at best transcoded only about half the videos into WebM. E.g., this YouTube search for “cat” gives ~1.8M results, while the same one for WebM videos gives 704K results.
    • WebM on YouTube is presented only for videos that lack ads, which is a shrinking number on YouTube. Anything monetizable (i.e., popular) has ads and therefore is served as H.264.
    • All this is moot when you consider mobile, since there is no Flash on mobile, and as of yet no WebM hardware, and Apple’s market-leading position.