6 ms·

Why Mozilla will support H.264 on OS that comes with it

90 points by lillycat 15y ago

fpgeek 15y ago
To put it more simply, what is the greater threat to the open web: platform-dependent mobile apps or patent-encumbered video codecs?
- lillycat 15y ago
  Without doubt platform-dependent mobile apps: by nature patents expire.
  - ZeroGravitas 15y ago
    I agree in general, but note that those pushing patented codecs aren't idiots and have an upgrade treadmill ready for users where at each decision point it'll make sense to upgrade to a newer codec with new patents. Having said that, the silliness of patented codecs seems to have been recognized for lower bitrate use cases (image, speech, audio) so I guess it's only a matter of time for video. I think a Googly approach where the encoder and decoders are assumed to be software like anything else in Chrome and updated every 6 weeks would be interesting to see.
    - bad_user 15y ago
      Even though there are better alternatives to MP3 from both a technical and legal perspective, and even though Apple has been pushing for AAC within iTunes, nothing is close to replacing MP3 as the defacto standard for distributing audio files. There are multiple reasons for this, like the huge cost of re-encoding everything that was once encoded in MP3, the broad hardware and software support or the fact that MP3 is good enough for most purposes. Broadband plays a role here as I'd bet that if it weren't so prevalent today, then people would actually want formats that sound better at lower bitrates. As it is, Amazon doesn't have a problem in selling songs encoded at 256 kbps, although you won't hear a difference when comparing that to an 128 kbps encoded AAC file from iTunes. An upgrade scheme only works if files encoded with the newer codecs don't have a problem being played on older software/hardware, because here's the thing ... H.264 is already good enough as far as video codecs go, just as MP3. And this is doable, but the newer patents will most likely touch the encoders, not decoders and the newer patents themselves will get more fragile over time, as you can't really innovate on top of the same technique forever. All in all I agree with the general sentiment ... proprietary platforms are more dangerous than patents and native apps are more dangerous to the future of the Internet than proprietary codecs.
      - ZeroGravitas 15y ago
        Just as a counterpoint that illustrates the patent upgrade treadmill, you say that nothing has come close to displacing MP3 for distributing audio files, yet if you're distributing that audio as the soundtrack for a video then MP3 has been quite firmly displaced by AAC due to bundling by the MPEG group (even though pirates held out with MPEG-4/DivX & mp3 for years). You could even argue that the fact that it always comes with such (relatively) large size video files makes the improvements of AAC over MP3 even less relevant, but still the "standard" has moved on and everyone needs to keep up. I also believe that 3G phones needed to support AAC but I could be just making that up. I've seen one "mp3 player" that dropped AAC support, but I can't see that happening for any device that also wants video. The direct result will be paying separate royalties on two mostly overlapping audio formats.
      - FrankBooth 15y ago
        Transcoding is still expensive for video, but not for audio. There's certainly some cost to transcoding all the world's MP3s, but given that a single modern CPU can encode and decode audio at 100-500x realtime, it's fairly small.
  - aw3c2 15y ago
    Platforms seem to get irrelevant quicker.
  - wmf 15y ago
    Patents on specific technologies expire, but there are always new patented technologies. I don't see us getting out from under patents for decades if ever.
- rimantas 15y ago
  Neither.
antonyme 15y ago
H.264 is arguably more open than WebM. It has been developed by standards committees over the course of many years, is published and backed by ITU, ISO, IEC and MPEG. A massive number of devices support it, there are loads of independent implementations, it is well documented and understood. It has been blessed as the standard codec in many media formats (BluRay, HD DVD) and for broadcast (ATB, DVB, etc) and is one of the most common video formats on the web today. WebM was originally a proprietary codec called VP8, developed by a single vendor (On2 Technologies) with a single implementation and no real specification. It was purchased by Google, and an ersatz specification was extracted from the implementation. I admire Google for what they are trying to achieve. But given both Windows and Mac ship with H.264 decoding support, and it is fast becoming the de facto format for mobile, the decision seems to be a no brainer.
- ZeroGravitas 15y ago
  I agree about the openness shortcomings of VP8 but suggest most are to do with it being a last-ditch attempt to get something out to compete with H.264 in time rather than conscious decisions taken by Google to minimize openness and maximise control (though some have portrayed it as such). I guess the big question is whether H.265(6,7) could be argued to be more open than VP9(10,11). I get the feeling the former will continue down the patent route and the latter will be more likely to be a well-engineered open source success story (hopefully emulating the Opus codec effort, who by the way, opted for the code as specification route too), but you never know.
  - 0x09 15y ago
    MPEG has made (I think) the right choice in splitting off a track for royalty-free web video (WebVC/IVC) while continuing their traditional generational codec development (HEVC). The former allows them to utilize expired patents from their long history as well as use their weight to negotiate royalty-free grants for current tech, in order to set an unrestricted baseline. Making this a separate effort rather than enforcing those conditions on the main standardization process allows them to continue evaluating the state of the art for all other use cases. I think this is as close as we'll get to having our cake and eating it too while patents on algorithms are still recognized.
cageface 15y ago
Mozilla's browser usage statistics show mobile devices at almost 2x desktop browser usage next year yet stats available here show mobile as about 10% of the market: http://en.wikipedia.org/wiki/Usage_share_of_web_browsers#Historical_usage_share http://en.wikipedia.org/wiki/Usage_share_of_web_browsers#His... Which of these is right?
- ZeroGravitas 15y ago
  The Mozilla numbers show mobile devices will outnumber desktop devices. The Wikipedia numbers show web usage. People surf more on desktop machines, and don't (need to) upgrade them as much so both figures are correct. The real issue for Mozilla is that you can get great benefits (shared history, passwords etc.) if you use the same browser on both platforms and sync them. Luckily Apple and Microsoft's lock-in approach seems to be holding them back here, but I'd imagine Android Chrome will be a big driver for Desktop Chrome (and possibly vice versa).
Porter_423 15y ago
At the end of 2010 I think the war of H.264 was reignited and I am surprise that this time by Google.But still H.264 has some big advantages such as such as it supports efficient hardware acceleration and its also an efficient decoder of the browser.
MatthewPhillips 15y ago
Money quote: > 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. Mozilla fought the good fight, and lost. Instead of standing on a street corner with a bullhorn, Mozilla is going to move on to fight the next battle. Those would be WebRTC and Google/Netflix's Web DRM initiative.
- tomjen3 15y ago
  What the hell is wrong with real time capabilities (I won't comment on open source DRM, except to say it would make an okay April first joke)? The web needs them, if it is to become the future we always wanted.
  - wmf 15y ago
    WebRTC is good; the question is whether it should use VP8 or H.264. Apple will argue for H.264 while Mozilla will argue for VP8. Since WebRTC has not been deployed yet there is no status quo to create path dependence.
bsimpson 15y ago
Didn't H.264 become free forever on the Web due to fear of competition from WebM? Doesn't sound like a total loss.
- wmf 15y ago
  Encoders and decoders have to pay royalties. But MPEG-LA says they're not going to charge more (legally questionable due to exhaustion) fees beyond that.
silon3 15y ago
So, will there be a European/free-world build for Linux/XP Firefox?
cpeterso 15y ago
Any guesses on the minimum number of (known!) patents that must be violated to implement an H.264 decoder? I assume the answer must be greater than zero, otherwise someone would have written an clever/convoluted H.264 decoder that sidesteps known patent issues.
- FrankBooth 15y ago
  The AVC/H.264 patent pool contains over a thousand "essential" patents. So, at a guess, at least hundreds.
ars 15y ago
I hope they also support ffmpeg libraries on linux to do the same thing. (libavcodec and libavformat) It would be really nice to have VDPAU video playback in the browser!

90 points by lillycat 15y ago

To put it more simply, what is the greater threat to the open web: platform-dependent mobile apps or patent-encumbered video codecs?

lillycat 15y ago
Without doubt platform-dependent mobile apps: by nature patents expire.
- ZeroGravitas 15y ago
  I agree in general, but note that those pushing patented codecs aren't idiots and have an upgrade treadmill ready for users where at each decision point it'll make sense to upgrade to a newer codec with new patents. Having said that, the silliness of patented codecs seems to have been recognized for lower bitrate use cases (image, speech, audio) so I guess it's only a matter of time for video. I think a Googly approach where the encoder and decoders are assumed to be software like anything else in Chrome and updated every 6 weeks would be interesting to see.
  - bad_user 15y ago
    Even though there are better alternatives to MP3 from both a technical and legal perspective, and even though Apple has been pushing for AAC within iTunes, nothing is close to replacing MP3 as the defacto standard for distributing audio files. There are multiple reasons for this, like the huge cost of re-encoding everything that was once encoded in MP3, the broad hardware and software support or the fact that MP3 is good enough for most purposes. Broadband plays a role here as I'd bet that if it weren't so prevalent today, then people would actually want formats that sound better at lower bitrates. As it is, Amazon doesn't have a problem in selling songs encoded at 256 kbps, although you won't hear a difference when comparing that to an 128 kbps encoded AAC file from iTunes. An upgrade scheme only works if files encoded with the newer codecs don't have a problem being played on older software/hardware, because here's the thing ... H.264 is already good enough as far as video codecs go, just as MP3. And this is doable, but the newer patents will most likely touch the encoders, not decoders and the newer patents themselves will get more fragile over time, as you can't really innovate on top of the same technique forever. All in all I agree with the general sentiment ... proprietary platforms are more dangerous than patents and native apps are more dangerous to the future of the Internet than proprietary codecs.
    - ZeroGravitas 15y ago
      Just as a counterpoint that illustrates the patent upgrade treadmill, you say that nothing has come close to displacing MP3 for distributing audio files, yet if you're distributing that audio as the soundtrack for a video then MP3 has been quite firmly displaced by AAC due to bundling by the MPEG group (even though pirates held out with MPEG-4/DivX & mp3 for years). You could even argue that the fact that it always comes with such (relatively) large size video files makes the improvements of AAC over MP3 even less relevant, but still the "standard" has moved on and everyone needs to keep up. I also believe that 3G phones needed to support AAC but I could be just making that up. I've seen one "mp3 player" that dropped AAC support, but I can't see that happening for any device that also wants video. The direct result will be paying separate royalties on two mostly overlapping audio formats.
    - FrankBooth 15y ago
      Transcoding is still expensive for video, but not for audio. There's certainly some cost to transcoding all the world's MP3s, but given that a single modern CPU can encode and decode audio at 100-500x realtime, it's fairly small.
- aw3c2 15y ago
  Platforms seem to get irrelevant quicker.
- wmf 15y ago
  Patents on specific technologies expire, but there are always new patented technologies. I don't see us getting out from under patents for decades if ever.
rimantas 15y ago
Neither.

antonyme 15y ago

H.264 is arguably more open than WebM. It has been developed by standards committees over the course of many years, is published and backed by ITU, ISO, IEC and MPEG. A massive number of devices support it, there are loads of independent implementations, it is well documented and understood. It has been blessed as the standard codec in many media formats (BluRay, HD DVD) and for broadcast (ATB, DVB, etc) and is one of the most common video formats on the web today. WebM was originally a proprietary codec called VP8, developed by a single vendor (On2 Technologies) with a single implementation and no real specification. It was purchased by Google, and an ersatz specification was extracted from the implementation. I admire Google for what they are trying to achieve. But given both Windows and Mac ship with H.264 decoding support, and it is fast becoming the de facto format for mobile, the decision seems to be a no brainer.

ZeroGravitas 15y ago
I agree about the openness shortcomings of VP8 but suggest most are to do with it being a last-ditch attempt to get something out to compete with H.264 in time rather than conscious decisions taken by Google to minimize openness and maximise control (though some have portrayed it as such). I guess the big question is whether H.265(6,7) could be argued to be more open than VP9(10,11). I get the feeling the former will continue down the patent route and the latter will be more likely to be a well-engineered open source success story (hopefully emulating the Opus codec effort, who by the way, opted for the code as specification route too), but you never know.
- 0x09 15y ago
  MPEG has made (I think) the right choice in splitting off a track for royalty-free web video (WebVC/IVC) while continuing their traditional generational codec development (HEVC). The former allows them to utilize expired patents from their long history as well as use their weight to negotiate royalty-free grants for current tech, in order to set an unrestricted baseline. Making this a separate effort rather than enforcing those conditions on the main standardization process allows them to continue evaluating the state of the art for all other use cases. I think this is as close as we'll get to having our cake and eating it too while patents on algorithms are still recognized.

cageface 15y ago

Mozilla's browser usage statistics show mobile devices at almost 2x desktop browser usage next year yet stats available here show mobile as about 10% of the market: http://en.wikipedia.org/wiki/Usage_share_of_web_browsers#Historical_usage_share http://en.wikipedia.org/wiki/Usage_share_of_web_browsers#His... Which of these is right?

ZeroGravitas 15y ago
The Mozilla numbers show mobile devices will outnumber desktop devices. The Wikipedia numbers show web usage. People surf more on desktop machines, and don't (need to) upgrade them as much so both figures are correct. The real issue for Mozilla is that you can get great benefits (shared history, passwords etc.) if you use the same browser on both platforms and sync them. Luckily Apple and Microsoft's lock-in approach seems to be holding them back here, but I'd imagine Android Chrome will be a big driver for Desktop Chrome (and possibly vice versa).

Porter_423 15y ago

At the end of 2010 I think the war of H.264 was reignited and I am surprise that this time by Google.But still H.264 has some big advantages such as such as it supports efficient hardware acceleration and its also an efficient decoder of the browser.

MatthewPhillips 15y ago

Money quote: > 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. Mozilla fought the good fight, and lost. Instead of standing on a street corner with a bullhorn, Mozilla is going to move on to fight the next battle. Those would be WebRTC and Google/Netflix's Web DRM initiative.

tomjen3 15y ago
What the hell is wrong with real time capabilities (I won't comment on open source DRM, except to say it would make an okay April first joke)? The web needs them, if it is to become the future we always wanted.
- wmf 15y ago
  WebRTC is good; the question is whether it should use VP8 or H.264. Apple will argue for H.264 while Mozilla will argue for VP8. Since WebRTC has not been deployed yet there is no status quo to create path dependence.

bsimpson 15y ago

Didn't H.264 become free forever on the Web due to fear of competition from WebM? Doesn't sound like a total loss.

wmf 15y ago
Encoders and decoders have to pay royalties. But MPEG-LA says they're not going to charge more (legally questionable due to exhaustion) fees beyond that.

silon3 15y ago

So, will there be a European/free-world build for Linux/XP Firefox?

cpeterso 15y ago

Any guesses on the minimum number of (known!) patents that must be violated to implement an H.264 decoder? I assume the answer must be greater than zero, otherwise someone would have written an clever/convoluted H.264 decoder that sidesteps known patent issues.

FrankBooth 15y ago
The AVC/H.264 patent pool contains over a thousand "essential" patents. So, at a guess, at least hundreds.

ars 15y ago

I hope they also support ffmpeg libraries on linux to do the same thing. (libavcodec and libavformat) It would be really nice to have VDPAU video playback in the browser!