4 ms·

While none of this is false, it's a bit of a misdirection to imply that the 'mighty MPEG' was a monolithic actor preventing this course of action from taking pl

While none of this is false, it's a bit of a misdirection to imply that the 'mighty MPEG' was a monolithic actor preventing this course of action from taking place. In fact, the individual collaborators that MPEG (or ITU) brought to the table were the same sorts of hardware companies, software companies, universities, research institutes, and media companies that are surprisingly familiar among AV1's contributors, and MPEG's licensing scheme -- the patent pool being its primary innovation -- was a way to seek mutually assured use of each other's work. The story here isn't that the Big Evil companies behind the MPEG codecs were beaten by the Good companies behind the Alliance for Open Media -- the story is that the pool of companies is kinda-sorta the same or comparable, with plenty of them playing both sides, a fact not lost on the founder of MPEG [1][2]. [1] http://blog.chiariglione.org/2018/01/28/ http://blog.chiariglione.org/2018/01/28/ [2] https://news.ycombinator.com/item?id=16261313 https://news.ycombinator.com/item?id=16261313

adrianratnapala 9y ago
> it's a bit of a misdirection to imply that the 'mighty MPEG' was a monolithic actor preventing this course of action from taking place. In fact, the individual collaborators True but if we are comparing models for getting stuff invented collaboratively (open source vs. licensed IP) then this hardly misdirection. Indeed the fact that the same sort of actors found it possible/useful/necessary to go down the open route makes TD-Linux's point all the more interesting. This is important because when policymakers talk among themselves they often just assume that IP and such are necessary for fostering innovation.
- rayiner 9y ago
  > This is important because when policymakers talk among themselves they often just assume that IP and such are necessary for fostering innovation Would AV1 have been possible in 1993 when MPEG-1 came out? How would you have bankrolled the standard at the time?
  - xiphmont 9y ago
    > Would AV1 have been possible in 1993 when MPEG-1 came out? No. There weren't enough transistors in the world. I'm not being silly. Something like Av1 happens now because Moore/Dennard allow it to.
    - niftich 9y ago
      It seemed to me like they meant the question politically, not technically, as in, would AOMedia have been possible? I think it's widely understood that the technical innovations in sixth-generation codecs like AV1 were simply not computationally feasible on circa ~1993 devices, so I highly doubt that's what was meant. Therefore I'm not sure how transistor density relates to funding models for DSP innovation.
      - xiphmont 9y ago
        The answers aren't entirely separable. Hardware capable of any kind of video wasn't accessible to the larger community and it sure as hell wasn't affordable. When Xiph.Org started (1994), even audio required high-end hardware. Yes, sound cards were available for PCs, but big enough hard disks were not. When only the elites have access, the standards are made by elites.
  - jmvalin 9y ago
    Getting something like AOM would have been easy back in 1993 because the costs would have been much lower. Back then complexity had to be really low, which means most of the complicated modern tools were off the table. Coming up with something equivalent to MPEG-1 would have required just a handful of engineers over maybe a year. In terms of IPR, there would also have been much less to check than today. OTOH, the minefield was moving really fast at the time, which could have added some complications. In the end, I think the main reason nobody bothered with something AOM-like is that few people realized the huge problem of patents on standards.
    - niftich 9y ago
      ~1993 is around when CompuServe and Unisys began negotiating about the implications of patented LZW in GIF files, but the issue didn't become public until 1994, but the community rebounded spectacularly by developing PNG in 1995. So perhaps the cutoff year of 1993 is right before the threat of non-pooled patents was well-publicized. There was a lot of video codec innovation back then too, but everything aside from the MPEG or ITU-T codecs was proprietary, and everyone took out patents. To Monty's point, the sheer number of endpoints capable of consuming digital video whose consumption somehow results in income for the publishers was just not quite there, making an alternate push for deriving revenue from DSP IP than patent licensing fees much less likely.
- peoplewindow 9y ago
  Money is necessary. IP rights are a general and flexible tool to funnel money into research. AV1 is basically funded by Google as a quasi-charitable endeavor, as far as I can tell. Whilst there may be several apparent funders like Mozilla, they in turn trace their funding to Google. "Research funding by Google" is never going to be a tactic that policymakers take seriously, and rightly so, even if it happens to be doing a lot of good work in this particular time period.
  - xiphmont 9y ago
    >AV1 is basically funded by Google as a quasi-charitable endeavor Nothing charitable about it. They want to make money. So do a bunch of other companies. And they've realized the way to win that game is to relinquish control over the fundamental/infrastructural pieces. That also has some beneficial aspects beyond 'a rising tide lifts all boats', but I feel more comfortable appealing to reliable motivations. The benefits to others aren't an accident, and they're important to e.g. us at Xiph, but let's not ascribe industry interest to anything more charitable than 'enlightened self-interest'.
xiphmont 9y ago
MPEG is, for purposes of this discussion, a system that member companies have learned how to game. It is so thoroughly gamed that it has become dysfunctional and an impediment to progress. The open/RF development strategy is now consistently beating it despite MPEG having a consistent royalty revenue stream and a massive patent war chest. At this point, MPEG, as a self-interested system, is only trying to slow things down and keep control. No, it's not monolithic inside, but the worst actors have solid control and it warps the behavior of the entire organization. We saw that coming a long, long time ago and yet barely in time.

adrianratnapala 9y ago

> it's a bit of a misdirection to imply that the 'mighty MPEG' was a monolithic actor preventing this course of action from taking place. In fact, the individual collaborators True but if we are comparing models for getting stuff invented collaboratively (open source vs. licensed IP) then this hardly misdirection. Indeed the fact that the same sort of actors found it possible/useful/necessary to go down the open route makes TD-Linux's point all the more interesting. This is important because when policymakers talk among themselves they often just assume that IP and such are necessary for fostering innovation.

rayiner 9y ago
> This is important because when policymakers talk among themselves they often just assume that IP and such are necessary for fostering innovation Would AV1 have been possible in 1993 when MPEG-1 came out? How would you have bankrolled the standard at the time?
- xiphmont 9y ago
  > Would AV1 have been possible in 1993 when MPEG-1 came out? No. There weren't enough transistors in the world. I'm not being silly. Something like Av1 happens now because Moore/Dennard allow it to.
  - niftich 9y ago
    It seemed to me like they meant the question politically, not technically, as in, would AOMedia have been possible? I think it's widely understood that the technical innovations in sixth-generation codecs like AV1 were simply not computationally feasible on circa ~1993 devices, so I highly doubt that's what was meant. Therefore I'm not sure how transistor density relates to funding models for DSP innovation.
    - xiphmont 9y ago
      The answers aren't entirely separable. Hardware capable of any kind of video wasn't accessible to the larger community and it sure as hell wasn't affordable. When Xiph.Org started (1994), even audio required high-end hardware. Yes, sound cards were available for PCs, but big enough hard disks were not. When only the elites have access, the standards are made by elites.
- jmvalin 9y ago
  Getting something like AOM would have been easy back in 1993 because the costs would have been much lower. Back then complexity had to be really low, which means most of the complicated modern tools were off the table. Coming up with something equivalent to MPEG-1 would have required just a handful of engineers over maybe a year. In terms of IPR, there would also have been much less to check than today. OTOH, the minefield was moving really fast at the time, which could have added some complications. In the end, I think the main reason nobody bothered with something AOM-like is that few people realized the huge problem of patents on standards.
  - niftich 9y ago
    ~1993 is around when CompuServe and Unisys began negotiating about the implications of patented LZW in GIF files, but the issue didn't become public until 1994, but the community rebounded spectacularly by developing PNG in 1995. So perhaps the cutoff year of 1993 is right before the threat of non-pooled patents was well-publicized. There was a lot of video codec innovation back then too, but everything aside from the MPEG or ITU-T codecs was proprietary, and everyone took out patents. To Monty's point, the sheer number of endpoints capable of consuming digital video whose consumption somehow results in income for the publishers was just not quite there, making an alternate push for deriving revenue from DSP IP than patent licensing fees much less likely.
peoplewindow 9y ago
Money is necessary. IP rights are a general and flexible tool to funnel money into research. AV1 is basically funded by Google as a quasi-charitable endeavor, as far as I can tell. Whilst there may be several apparent funders like Mozilla, they in turn trace their funding to Google. "Research funding by Google" is never going to be a tactic that policymakers take seriously, and rightly so, even if it happens to be doing a lot of good work in this particular time period.
- xiphmont 9y ago
  >AV1 is basically funded by Google as a quasi-charitable endeavor Nothing charitable about it. They want to make money. So do a bunch of other companies. And they've realized the way to win that game is to relinquish control over the fundamental/infrastructural pieces. That also has some beneficial aspects beyond 'a rising tide lifts all boats', but I feel more comfortable appealing to reliable motivations. The benefits to others aren't an accident, and they're important to e.g. us at Xiph, but let's not ascribe industry interest to anything more charitable than 'enlightened self-interest'.

xiphmont 9y ago

MPEG is, for purposes of this discussion, a system that member companies have learned how to game. It is so thoroughly gamed that it has become dysfunctional and an impediment to progress. The open/RF development strategy is now consistently beating it despite MPEG having a consistent royalty revenue stream and a massive patent war chest. At this point, MPEG, as a self-interested system, is only trying to slow things down and keep control. No, it's not monolithic inside, but the worst actors have solid control and it warps the behavior of the entire organization. We saw that coming a long, long time ago and yet barely in time.