A Competing HEVC Licensing Pool | Hacker News Reader

A Competing HEVC Licensing Pool

xiphmont.livejournal.com

80 pointsKwanEsq22 comments

Just want to say, thank you Monty and Co for the work you all do at Xiph. Ogg, Vorbis, Theora, Speex, FLAC, Opus, Icecast, and Daala... it's a seriously impressive portfolio. It probably would not be a stretch to say that billions of people use these daily, often without even realizing it (which is often the best kind of tech).

And thank you to Red Hat, Mozilla and all the others who sponsor this work too. It really is incredibly important to have these open codecs, not only for their own sake, but for the significant pressure it puts on the MPEG cartel. As bad as they are, without stuff like Xiph I think they would be much, much worse.

> It probably would not be a stretch to say that billions of people use these daily.

I think it is a stretch. There are maybe 4 billion people that access the Internet currently and most of those have limited access. Only a fraction of those know about, much less use one of those technologies. To be in the billions, 50% of people accessing the Internet would have to use one of those technologies every day. I just don't believe that is the case.

BTW, this has nothing to say of the technologies themselves. It's great that they exist. The faster we can get rid of royalty the better IMO.

If they use Youtube or Skype, they use some of them. Any online/app audio/video chat thing will probably use a mixture.

Depending on how your office VOIP phone works, or the PBX it's connected to, or your POTs carrier, you might be using those technologies. Maybe the CCTV at your work uses them too...

If you are building a multimedia thing, and you don't want to worry about paying for patents, you might reach for one of these. Your users won't know.

Wikipedia (6th largest website in the world) serves only unencumbered formats. Apple's Siri uses Speex. As does Flash. This stuff shows up in more places than you think, precisely because it allows innovation without asking permission.

My new favorite metaphor : Schroedinger's Cash Box

As another point of reference, you can compare the licensors in the MPEG-LA H.264 patent pool:

http://www.mpegla.com/main/programs/AVC/Pages/Licensors.aspx

with those in the MPEG-LA HEVC patent pool:

http://www.mpegla.com/main/programs/HEVC/Pages/Licensors.asp...

I'm really excited about Daala, in ways that VP8 and VP9 never excited me—VP8 seemed like a damaged version of MPEG4. Although it looks like MPEG4 will be "good enough" for the web for a long time to come, it seems like HEVC development is weighed down by some perverse incentives.

See this post from "Diary Of An x264 Developer" in 2010: http://x264dev.multimedia.cx/archives/360

Meanwhile, Opus came along and outperformed HE-AAC, which makes me think that more good things are going to come from open standards.

If you haven't looked at in a while, you should try VP9 again. It is much faster, especially if you try the new 1.4.0 release candidate.

Also keep in mind that MPEG 4 includes two separate video codecs, ASP and AVC, with wildly different performance.

> Apparently all we have to do is stand back and let the dominant players commit suicide while they dance around Schroedinger's Cash Box

Don't know how true this is with regards to MPEG, but I really enjoyed this quote

MPEG is funded directly by their organization (ISO) and doesn't really have a (edit: financial) dog in this fight.

One thing that's easily mixed up: MPEG sets standards. MPEG-LA is a wholly different entity to organize cashing in on them.

MPEG's policies on creating standards are basically set up to enable the MPEG-LA and new patent pools like this one to make money, though. They have a policy of not caring about whether the technology is patented or not - there could be a small tweak that'd turn it from patent-infringing to patent-free and they'd ignore it - which means that all the members can and do cram as many of their patents into MPEG standards as they can find excuses for.

Can someone explain why this is terrible for the licensors? If they have patents that apply, can't they pretty much setup whatever ridiculous roadblocks and tolls they desire?

Or are the open source codecs truly patent-free, as in they don't accidentally step on something that could be "infringing" legally?

The whole point of paying the MPEG-LA protection racket is that you don't get sued over patents. If there's another patent pool that claims to have their own patents embodied by one or more of the major proprietary codecs, that protection racket starts to look more like the sham it is. Someone can always come out of the woodwork and claim you're infringing one of their patents.

The open source codecs may accidentally step on something that could be infringing, and MPEG-LA and similar entities are always quick to point that out. The implied message is that you're safe when you use codecs under license with MPEG-LA maintained pools.

This demonstrates that the MPEG-LA also is no guarantee for patent safety (and if you read their fine print, they never actually claim to be), and so far this new organization is using similar strategies (not listing the patents, and right now, not even the patent owners) which are essentially FUD.

That this blows up in the media is not so much terrible for licensors or licensees, but for the pool forming bodies that sit in the middle to seek rent and for the codecs that did benefit from that 'safety'.

All of the new codecs the author proposed as alternatives have the same issue: companies who took no part in creating the codec may someday decide that the codec infringes their patents and make everyone using the codec pay for a license.

Abstractly, sure - all software is vulnerable that way. However, specifically, if you create a new codec that is heavily based on previous ones, you are clearly more at risk, than if you create a more original codec using novel methods. The latter is safer.

Doesn't it make sense that a codec that's heavily patented would be more likely to have all of the patents known, rather than one that hasn't been looked at, and is likely to accidentally infringe one of the patents in the minefield that exists?

I don't think so. The more novel something is, the higher the chance that no one thought of it, and so could not patent it.

If something is patented, that means people know of it, and will try to patent similar things or parts of it that haven't been patented yet, making it riskier.

For Daala, they've purposefully avoided 'technologies' that have been used by other codecs before. This severely reduces the risk of coming across such a patent. Xiph also apply for patents themselves, to ensure it's protected for the future. Below is an introductory video about daala, that talks about this a bit.

https://www.youtube.com/watch?v=Dmho4gcRvQ4

In the case of NETVC, all contributors are required to submit IPR declarations. This doesn't prevent other uninvolved parties from popping up with patents later, but is still far better than the "patent-blind" approach that MPEG takes.

The same applies to every other codec (or piece of software) as well.

MP3 had Sisvel who came late with legitimate claims (as far as patents go). HEVC now has this second pool coming out of nowhere.

If anything, the promise of MPEG-LA, that they manage to build a reliable licensing regime, is broken - and I think it's good that this starts to show.

In particular it pleases me that this new group pulls the same stunts on MPEG-LA that MPEG-LA used on others (see their attempt to build a VP8 pool): vague statements of how surely there must be some more patents that they'll maintain soon.

Reply on news.ycombinator.com