Slashdot Mirror
Free Software Camps Wading Into VP8 Patent Fight
An anonymous reader writes "As reported by Slashdot, Nokia recently notified the IETF that its RFC 6386 video codec (aka VP8, released by Google under a BSD license with a waiver of that company's patent rights) infringed several dozen of its patents; furthermore, Nokia was not inclined to license them under FRAND (fair, reasonable, and non-discriminating) terms. While the list provided by Nokia looks intimidating, Pamela Jones at Groklaw discovered that many appeared to be duplicates except for the country of filing; and even within a single country (e.g. the U.S.), some appeared to be overlapping. In other words, there may be far fewer distinct patented issues than what appears on Nokia's IETF form. Thom Holwerda at OSNews also weighed in, recalling another case where sweeping patent claims by Qualcomm and Huawei against the Opus open source audio codec proved to be groundless FUD. The familiar name Florian Mueller pops up again in Holwerda's article."
Thanks for the nostalgia and for reviving SCO in the guise of Nokia. It was nice of you to dig out Florian for a reprise too...
"I've got more toys than Teruhisa Kitahara."
Nokia, the new SCO?
Of course does the list contain duplicates. Patents are usually limited to specific jurisdictions. Patents granted by the US patent office are only valid in the US. That is why companies file the same patent in many jurisdictions. That does not make the list less intimidating, because for VP8 to be free they still need to be invalidated individually in many courts around the world. Just take a look at the recent case of Microsoft vs Motorola for how tricky this is. A US judge agrees with Microsoft, while a German judge agrees with Motorola.
So when it comes time for MS to dish out more FUD Florian shows up? What a fucking surprise.
At least when I see that name, I can ignore the quotes and comments as being nothing more than a paid shill spouting BS.
They still charge for the encoder.
Because VP8 is good enough and FREE, is why the big hassle is there. Giving up some performance to get out from under the MPEG-LA's thumb is well worth it.
They still have their palms out if you want to encode video for public consumption. It isn't about screwing the consumer so much as preventing the consumer from becoming a producer.
Oh, no! You have walked into the slavering fangs of a lurking grue!
And GPUs only support H264, not VP8.
RIGHT NOW, GPUs only support H.264.
If VP8 became a standard that might change.
When talking about codecs one has to think long term.
We still use MP3 after more than 15 years.
Uhm.... wow. Seems really wrong somehow. But I think I see what you're trying to say.
"The systems" are in place to support "the thing." The copyright system is in place to support creators of creative works. The patent system is in place to support the creators of inventions. But somehow, there are people who game the system to make the system itself how they make their money.
The systems themselves should be the method by which government protects works of practicing entities. But instead the systems themselves are the things because they are non-practicing entities. And yes, that's very wrong and far from the intent of the systems.
As for Muslims? Sorry, but I don't get it. What I see, too often, is nothing like this. On the whole, I see a collection of people (not even sure it's a majority of muslims) who wish to respect and practice their traditions and cultures and for others to be sensitive to them. Most of the time, I'm okay with it. My experience is, "yes, they stand out in a crowd" but they do me no harm. Talk about people who fake or exaggerate being offended so that they can leverage that into some sort of power or influence? That happens all over.
Have a nice day! :) -- Picture of Mohammad smiling
MPEG-LA made a binding agreement that licensing of H.264 decoders would be free forever. They have already said that agreement would also apply to H.265, which is due to be formalized soon. VP8 is only about as efficient as H.264, H.265 is considerably more efficient.
The push for VP8 started when MPEG-LA wanted to charge a fee for licensing the decoder. It's now several years after that became a dead issue. The state of the art of video encoding has moved far past VP8. Why spend so much time and effort on an outdated codec?
They still charge for encoding. Perhaps more importantly, do you think that they made H.264 decoding free out of the goodness of their good little hearts, or because Google called their bluff?
This is so boring, really. I really consider today's tech industry just a huge pile of fraudulent investor. All these patents fights are over a software algorithm shows that there's no real innovation here: just plain old incremental releases that are developed and researched completely different entities (after a certain size, R&D division is like almost a different company) have nothing patentable on them, not in the original intention behind the whole idea of patents. This whole patent-wars are completely wasteful and useless, but the corporate lobby prevent any attempt of legislation that aim to eliminate corporate patents over trivial matters, so we stuck with these companies spending millions of dollars on lawyers and patent fights, for whose benefit? Lawyer benefit.
There has to be a point where it becomes so unbearable the whole idea of patents must be abolished and any company who participated in this fight must be also dismantled and assets to be redistributed.
Over the last few decades, Nokia has spent more money in R&D than almost any other company in the world.
They do spend quite a lot but they're not top of the heap even just in technology companies. IBM, and Microsoft both spend considerably more on research.
Nokia has spent roughly $4-5 billion per year but it's been dropping steadily from about $5B in 2009 down to about $3.7B last year. A very substantial sum to be sure but not out of line with other large tech companies and they've been forced to spend steadily less due to their financial position. Kind of amazing that they can't seem to develop a hit phone when they spend 5X what Apple does on R&D. Makes you wonder what the heck they are doing.
Also, GPUs tend not to be as fixed-function as they used to be. Some of the nastier mobile parts are still pretty inflexible; but Nvidia has certainly been pushing the nearly-general-purpose compute capabilities of their PC line down into their Tegra products, and they probably won't be the only one.
There are certainly a lot of existing and near future parts that are fixed function and will never support anything else; but if the pressure remains, newer fixed function devices might support it, and programmable devices will always be a firmware update away from doing so.
Of course does the list contain duplicates. Patents are usually limited to specific jurisdictions. Patents granted by the US patent office are only valid in the US. That is why companies file the same patent in many jurisdictions.
That's absolutely correct. However, there's another type of "duplicate" that Groklaw mentions, which is a continuation application:
But even in the US-only context, you see the same title more than once. Those are continuations. What's that? It's where the really elaborate machinations live, and the submarine patents. It's how you argue with the USPTO when it doesn't approve your patent application or if it does how you keep adding to what you already got approved from the USPTO, something you can do over and over to time indefinite... Sigh. In other words, that's where you try to get more than you had in the beginning, maybe as you see what others are inventing so you can sue them. Blech. I so hate patent law.
Well, of course PJ hates patent law, since her understanding of it is so flawed. In reality, continuation applications are new applications, tied to the parent, that are typically filed to claim an invention in a slightly different way, to fix errors in the claims, or to address other unclaimed aspects of the invention. For example:
1) You claim A+B+C+D+E. The patent office determines it's patentable because even though A+B+C is known, no one has ever done D+E and it's not obvious to do so. So, in a continuation, you claim just D+E, leaving out the unpatentable cruft;
2) You claim A+B+C+D+E, but without realizing it, you accidentally said that A is ~B. Whoops. Pay the filing costs all over again, and fix the error;
3) Your patent application describes A+B+C+D+E but also F+G+H+J. You did this to save drafting and filing costs by bundling it all into one giant application, but you didn't have enough money to pursue both inventions initially. Now, a few years later, you file a continuation claiming F+G+H+J so that you can pursue it.
PJ's complaints, however, are incorrect: these aren't submarine patents, since continuations are published almost immediately. You also can't add anything to what the parent application contained. Change anything beyond a misspelled word, and the application is rejected as containing new matter. And finally, you can't do it for time indefinite, since the continuations all expire on the same date as the parent. If you file an application on Jan 1, 2010, any patents issuing from it or the continuations expire on Jan 1, 2030. So, while you can keep filing continuations right up until Dec 31, 2029, they all expire on that final day, so you really aren't extending any rights temporally.
For someone who writes a lot about patents and talks to a lot of professors and patent lawyers, PJ sure misses some fundamental details.
VP8 Hardware codecs are up to generation six, freely licensed also - to 80 chip companies so far, and in production by many major vendors. So there's that.
Help stamp out iliturcy.
MPEG-LA made a binding agreement that licensing of H.264 decoders would be free forever.
Ah no, if you want to provide an encoder or decoder there are royalties, and even if the recipient has a licensed decoder the sender must pay if you offer a subscription or PPV service. The only thing they've promised is that Internet Broadcast AVC Video like YouTube does not require extra payment for the use of the codec on top of what you've already paid for the decoder (typically through Windows, OS X, Chrome or Flash) and Google has paid for the encoder. P.S. If you do have an AVC encoder, you will typically find that the license terms say you can only use it for "personal and non-commercial use", anything else requires a special license from MPEG LA and it doesn't come cheap.
Live today, because you never know what tomorrow brings
In a few months time (aka this year) outside of the USA MP3 is going to be patent free. Even in the USA in a couple of years decoding will be patent free. MP3 is going to be around for a very very long time because there is a huge amount of material encoded in it, so it has momentum and the patents are expiring rapidly.
Sigh.
If you sell individual videos more than 12 minutes long, you play MPEG-LA a royalty of of 2% or $0.02 per sale, whichever is less.
If you run a paid subscription service, and you have more than 100,000 subscribers, you pay MPEG-LA a royalty between $0.10 and $0.25 per subscriber per year.
If you broadcast your shows on TV, you pay either a one time fee of $2500 for each encoder, or between $0.005 and $0.01 per viewer per year.
If you make your videos available for free (even if they are ad supported) you pay no royalty.
If you sell videos less than 12 minutes long you pay no royalty.
If you run a subscription service with less than 100,000 subscribers you pay no royalty.
If that "prevents you from becoming a producer", you might want to rethink your business model.
(Source: the AVC/H.264 terms.)
How can I believe you when you tell me what I don't want to hear?
Also, GPUs tend not to be as fixed-function as they used to be. Some of the nastier mobile parts are still pretty inflexible; but Nvidia has certainly been pushing the nearly-general-purpose compute capabilities of their PC line down into their Tegra products, and they probably won't be the only one.
I've been paying a little attention to the possibility of shader-based decoding in the AMD open source drivers since AMD probably can't release UVD specs (their fixed function hardware) and so far it's been a dud, yes certain things are extremely well suited to GPUs and is done in shaders by the proprietary driver as well but other parts run very poor so they've been looking at mixed CPU/GPU solutions but then you have all sorts of latency/memory synchronization issues. In short, if you're not doing it all on the CPU you want some form of fixed function hardware and I imagine that is true for VP8 as well.
Live today, because you never know what tomorrow brings
From what I understand VP8 was based on the reference code for MPEG4 but modified to avoid all known patented methods. That seems inherently risky. Why not instead invest in SNOW or other wavelet encoding methods and leap ahead of the current MPEG standard?
Cwm, fjord-bank glyphs vext quiz
It isn't about screwing the consumer so much as preventing the consumer from becoming a producer.
And controlling the playback devices, and therefore the means of production and distribution in the video arena.
My God, it's Full of Source!
OUTSIDE_IP=$(dig +short my.ip @outsideip.net)
Florian, is it you?
And I say that as a booster who was happy with an n810 and the Qt work just a few years ago. Sorry about that cancer you got, Nokia (or was it MS?), but it's changed you and it's fatal.
My God, it's Full of Source!
OUTSIDE_IP=$(dig +short my.ip @outsideip.net)
They still charge for encoding. Perhaps more importantly, do you think that they made H.264 decoding free out of the goodness of their good little hearts, or because Google called their bluff?
Neither. Honestly, they did it because that is what the players in the MPEG-LA wanted. They knew that a new standard codec was needed. The knew Microsoft was giving away VC-1 for free, and without a free play-back situation for H.264, the proprietary, single vendor (and arguably better quality) VC-1 would become the standard - remember the Blu-Ray standard allows VC-1 in addition to H.264 and MPEG-2.
H.264 is a great codec. It took a lot to get it up and going and it took a lot to get it standardized and documented. It should be obvious to anyone that the parties should be rewarded for the effort. The idea that Google should be allowed to just copy (yeah, VP8 is a bad, bad, bad copy of H.264) it and get away with it is absurd.
They still charge for the encoder. Because VP8 is good enough and FREE, is why the big hassle is there.
This is what you owe MPEG LA for the use of a licensed H.264 encoder:
Personal and internal business use: 20 cents a unit, max. Paid by hardware and software manufacturers selling more than 100,000 units each year. Realistically, given the enterprise caps and volume discounts, 10 cents a unit or less.
User pays for content by subscription: Nothing until you have more than 100,000 subscribers.
User pays for content by title: 2 cents a title, max.
Internet distribution free to user is free to distribute. AVC Patent Portfolio License Briefing
It is not too difficult to find VP8 aware video editors. But video cameras, video hardware of kind, for any purpose?
anything else requires a special license from MPEG LA and it doesn't come cheap.
You are right, if you sell videos over 12 minutes in length, you have to pay either 2% of the video in licensing fees, or $0.02, whichever is less. That is back-breaking bankruptcy-inducing fees right there. As mentioned above in the thread - if your business model can not support that, you should re-think it. Remember, this is only if you are selling videos some how, and getting paid directly for each one by the person buying it.
If you run a subscription service with more than 100 000 paying subscribers, you are going to be out 1/10th of a cent per subscriber per year or less. Again, go straight to bankruptcy, right?
Doesn't Google/YouTube have to still produce H.264 though? Unless you're assuming YouTube will go only VP8. If that's the case then everyone else will have the ability to play it so they can watch their cat videos.
I'm not defending their actions, just don't see how YouTube is any different from any other content creator in your scenario.
Since you can't redistribute even 1 cent is too much.
"Why should google payoff a company without proof it infringes"
Two good reasons:
- The cost to pay them off can be less than the legal costs of fighting, espicially if they manage to get an injunction.
- Even if it is blatantly obvious to anyone in the industry that the patents are junk, judges are not technical experts, so there is always some element of risk.
Because being a Microsoft stooge is so fun.
* Carthago Delenda Est *
The exact same extortion is occuring now, threatening podcast producers with some bullshit patent trolling - only *after* it was finally becoming financially viable for the real creative artists.
Nokia is not a design company. They're an engineering company.
They make most of their money selling phones. As such they are a consumer electronics company. Nokia doesn't make the majority of their money selling engineering. They may have an engineering culture but that isn't the same thing. Similarly Apple is a consumer electronics company that arguably has a design culture.
(Actually you can make a very credible argument that Apple really is a software company that bundles their software with nicely designed commodity hardware. One could put Android on an iPhone or Windows on a Mac but if Apple did that Apple's profit margins would disappear faster that you could say shareholder lawsuit because everything that truly makes their products different in important ways is in the software. A Mac running windows isn't much different from a PC from Dell or HP.)
Nokia's R&D also is more engineering-focused. Hardware, signals processing, accessibility, etc. They'll be trying to cram a large swiss army knife worth of tools into a phone, or coming up with new antenna designs to improve signal transmission and reception, or finding new materials to make lighter, thinner phone casings.
All of which is meaningless unless they can come up with products people want to buy or technology they can license to others who do have products people want to buy. I think large companies absolutely should have research labs working on some long term basic science and engineering problems. But Nokia is seriously in danger of going out of business. They have lost over $4 billion in the last two years and their share price is around $3.28 per share as I write this. Their assets are falling faster than their liabilities. While they still have a lot of cash, they have been hemorrhaging cash for the last year or so. Their market cap is around $12 billion which Apple, Microsoft, Samsung or Google could buy with petty cash were they so inclined. (I doubt any of them would except maybe Microsoft)
Google has already announced (but not effectuated) that they will not support H.264 in Chrome in the future, only Ogg and VP8, so, no, they will probably not continue to support H.264 on YouTube. "The One Google Way or Fuck Off" is coming quite soon.
In other words, such an encoder can never be used to produce CC-BY content. Would you accept a compiler that could not be used to produce GPL code? Then why would you use an encoder that cannot produce Free content?
Give me Classic Slashdot or give me death!
Here's my problem: because H.264 is a patented technology, you can't use it without the permission of the patent holders. So maybe today they are charging a couple of pennies per individual video, but how can I trust that this won't go up significantly tomorrow?
My understanding is that the H.264 patents won't expire until somewhere around 2027 or so. That is a long time to be at the mercy of patent holders.
Also, the technology being patented is a problem for free software projects like Firefox. I would like to see at least one video codec with acceptable performance that free software can use freely. Even if H.264 was licensed free-as-in-beer, there are restrictions on it that make it impossible for free software projects to use.
Google's lawyers spent a long time looking over VP8 before Google tried to set it free. So far no challenges to VP8 have really succeeded (MPEG-LA got some money, but failed to stop VP8 or get royalties, and that really must be considered a failure for MPEG-LA). I'm hoping and expecting that this challenge will, in the end, not succeed either.
If I'm right, what happens? Then VP8 becomes a free, lower-performing alternative to H.264. H.264 retains its status as the favorite codec at Apple, all those mobile devices still have H.264 built-in, and MPEG-LA can still collect the royalties. As you noted in your post, the royalties are not unreasonable.
It will be a similar situation as Vorbis and MP3. I consider Vorbis to be a success; it didn't kill MP3, but it did provide a useful alternative, and it kept the MP3 royalties from getting completely crazy. (Vorbis is actually technically superior to MP3, so I once had hopes it might "win" but it never happened.) I expect a similar story from VP8: it will never displace H.264 as the top format, and years from now people will sneer at it for "failing" to do so... but it will give Google and other companies a bargaining chip when MPEG-LA tries to raise royalties too much. They can make a serious threat to migrate their business away from H.264 and to VP8 if the royalties go too high.
If H.264 really was the only game in town, the industry would have to pay whatever rates MPEG-LA chose to set. And in the end, that means the consumers would pay.
lf(1): it's like ls(1) but sorts filenames by extension, tersely
Chrome dropping H.264 was announced over two years ago with absolutely no follow up. It's entirely possible they just changed their minds on the whole thing.
Possibly, on the other hand, if they changed their minds, VP8 makes no sense at all, then it is just a meaningless posturing trying to peddle an inferior product on an unsuspecting public. That'd be insane.
FRAND != free. the articles say no such thing.
besides, it's not as if wheels aren't constantly re-invented in the software world.
Didn't Nokia also make a bunch of money selling cell towers and other network stuff?
Yes. About half their revenue comes from cell phones and related products and about half comes from infrastructure via Nokia Siemens Networks which is a 50/50 joint venture. They have a third division for their mapping technology (NAVTEQ) but its revenue is relatively inconsequential at about 1/10th the size of the other two divisions. The infrastructure division is relatively profitable though joint ventures tend to be unstable in the long run. (commonly the partners eventually have differing goals and one or the other takes over) It's not a good situation when your most profitable division is half owned by someone else. Worse, Nokia's prospects in cell phones now depend heavily on Microsoft and that hasn't exactly worked out amazingly well.
I think Nokia has a business model that made sense a few years ago when people weren't so picky about the software on their cell phones. (everyone's software kind of sucked, not just Nokias) Nokia was geared to sell to AT&T and Verizon and the other carriers rather than worrying about the needs of the end users. They tried lots of phones but they never worried too much about the phone interface because their real customers (the carriers) didn't care. Apple changed expectations about the interface for phones and Nokia was caught flat footed and consumers haven't warmed much to their latest offerings.
Nokia historically has made pretty good hardware but their software has traditionally been mediocre at best. Their internal software projects were not getting much consumer interest and so they threw in with Microsoft. I don't think they are dead yet but if their stock price falls too far I wouldn't be shocked to see the cell phone division get bought by Microsoft or see the company broken up into its constituent divisions.