Can We Legislate Past the H.264 Debate? 310
Midnight Warrior writes "We could solve the H.264 debate if a country's legislature were to mandate that any patents that contribute to an industry-recognized standard were unenforceable in the application of that standard. Ideally, each standard would also be required to have a 'reference design' that could be used without further licensing. This could also solve problems with a ton of other deeply entrenched areas like hard drives, DRAM, etc. RAND tries to solve this strictly within industry, but both the presence of submarine patents and the low bar required to obtain a patent have made an obvious mess. Individual companies also use patent portfolios to set up mutually assured destruction. I'm not convinced that industry can solve this mess that government created. But I'm not stupid; this clearly has a broad ripple effect. Are there non-computer industries where this would be fatal? What if the patents were unenforceable only if the standard had a trademark and the implementer was compliant at the time of 'infringement'? Then, the patents could still be indirectly licensed, but it would force strict adherence to standards and would require the patent holders to fund the trademark group to defend it to the end. In the US model, of course."
Re:No. Just pay up (Score:3, Informative)
Indeed, Linux and Ubuntu linked in the summary have had H.264 support for years. It's a no brainer for them to pay a license for their OEM distributors to be able to sleep at night. I don't really see that as a "debate" though.
With regards to codecs like Theora I've tried a couple of times to get encode HD video in it but it always comes out blurry. I've seen videos like the HD Theora video on the big buck bunny website so I know it's possible however there doesn't seem to be much information out there on the correct settings. For example how would one convert a clear h.264 HD video to Theora and have it come out with the same quality.
Violation of TRIPS (Score:5, Informative)
We could solve the H.264 debate if a country's legislature were to mandate that any patents that contribute to an industry-recognized standard were unenforceable in the application of that standard.
I have read until here. What you propose is unfortunately not allowed under the TRIPS agreement [wikipedia.org], which requires that once that a patent has been granted, the holder must be able to enforce it. While there can be exceptions to this rule, I highly doubt that a country trying to get rid of the H.264 patents that way will get away with it.
If a legislative body wants to fight these patents, the best thing it can do is to require the use of unencumbered technologies in the government.
Re:Uh huh (Score:5, Informative)
Ahh... But you don't know all the places that it requires LICENSING.
Use it to produce a home movie, you're okay.
Use it to produce a indie movie, even with "pro" grade equipment and you're not.
Use it to produce a demo reel for your work, and you're not.
Only parts of the generation or playback licensing have been paid for- you're on the hook for everything else and they'll enforce if you hit a certain threshold (about $100k of revenue of any kind generated from it...). They'll come mug you for money at that point and it's NOT cheap.
Re:No. Just pay up (Score:5, Informative)
Actually, you're not licensed to do it even BEFORE the 150k people viewing it- that's just the threshold at which they have chosen to ENFORCE their IP rights. You technically still need a license for it (Check the licensing details on your gear, even the pro-grade stuff will tell you that you need a separate license for professional uses of the gear. They're not kidding.).
And stating that it's relevant to video sharing sites- they're an enabler, but YOU are the one on the hook, not they (because there's yet another license THEY have to have to do what they're doing...) and you're still needing that license in addition to the one they're paying.
As for Theora being better than h.264... No, you'd be right about that. It's on a rough par with MPEG4- VP8's closer to what you're looking for and if rumor's right Theora 2 will be in that space. Having said that, I'd prefer a web (and others...) standard to be something that's utterly unencumbered. All it'd take is for one player to play grab-em like Unisys did with LZW and you owe money all over the place. It could just as easily as not happen with h.264.
Re:What's an "industry-recognized standard"? (Score:1, Informative)
H264 is far away from the theoretically best compression. So we just need one socialist to prove that they can do more, to make something that beats H264, and give it away for free.
That's not how it works. 99% of progress is evolution, not revolution. A codec that beats H264 will inevitably share at least some characteristics with H264, and if those characteristics are patented -- even if the designer of the new codec has never even heard of H264 -- they're not allowed to distribute it in the US. (And you're conveniently ignoring the fact that patents are socialism.)
Can't ever be done in the US .. here's why (Score:2, Informative)
Article the seventh [Amendment V]
No person shall be held to answer for a capital, or otherwise infamous crime, unless on a presentment or indictment of a Grand Jury, except in cases arising in the land or naval forces, or in the Militia, when in actual service in time of War or public danger; nor shall any person be subject for the same offence to be twice put in jeopardy of life or limb; nor shall be compelled in any criminal case to be a witness against himself, nor be deprived of life, liberty, or property, without due process of law; nor shall private property be taken for public use, without just compensation.
---> It's the last half sentence here ... ; nor shall private property taken for public use, without just compensation http://www.half-bakery.com/
Re:Uh huh (Score:5, Informative)
No, actually it doesnt.
Patents are not real property. They are monopoly privileges *created* by the state, and in fact they represent "expropriation" to begin with. Understanding this fact is critical.
The proposal in this article is backwards for exactly that reason. We have a problem created by state interference in the economy, and rather than propose that the state simply *quit creating the problem* we propose even more interference.
Entirely backwards. The solution here is the opposite of the proposal. Simply abolish patents instead.
Re:No. Just pay up (Score:5, Informative)
Exactly. It's like creating furniture with a SawStop table saw and the patent holders expecting to get a cut of everything you make with it.
You should buy the right to use the patented technology, and that should be the end of it.
Re:Uh huh (Score:5, Informative)
Correct.
Producing the movie does not require a license. Distributing the movie for pay would require a license.
Producing with H.264 does not require paying a license fee. Demo reels generally are not distributed for pay or in quantities large enough to meet the thresholds for which licensing fees kick in.
Engadget had a good article [engadget.com] that dealt with much of the H.24 licensing FUD that is going around.
Two Cents A Dance (Score:5, Informative)
They'll come mug you for money at that point and it's NOT cheap.
It's dirt cheap.
Retail sale, disks or downloads:
Where an end user pays directly for video services on a title-by-title basis ...royalties for video greater than 12 minutes (there is no royalty for a title 12 minutes or less) are...the lower of 2% of the price paid to the Licensee (on first arms length sale of the video) or $0.02 per title
Paid subscription services:
Where an end user pays directly for video services on a subscription-basis (not ordered or limited title-by-title), the applicable royalties per legal entity payable by the service or content provider are:
100,000 or fewer subscribers per year. No royalty.
100,000 to 250,000. $25,000
250,000-500,000. $50,000.
500,000 to 1 million $75,000.
Over $1 Million. $100,000.
Broadcast, Cable and Satellite:
where remuneration is from other sources, in the case of free television...satellite and/or cable Transmission, and which is not paid for by an End User), the licensee (the broadcaster) may pay...according to one of two royalty options:
(i) a one-time payment of $2,500 per AVC transmission encoder
or (ii) annual fee per Broadcast Market
starting at $2,500 per calendar year per Broadcast Markets of at least 100,000
but no more than 499,999 television households
$5,000 per calendar year per Broadcast
Market which includes at least 500,000 but no more than 999,999 television households
and $10,000 per calendar year per Broadcast Market which includes 1,000,000 or more television households.
Free distribution over the Internet:
In the case of Internet broadcast for which the End User does not pay remuneration for the right to receive or view, i.e., neither title-by-title nor subscription), there will be no royalty during the first term of the License (ending December 31, 2010), and after the first term the royalty shall be no more than the economic equivalent of royalties payable during the same time for free television.
The Cap
In the case of the sublicenses for video content or service providers, the maximum annual royalty ("cap") for an enterprise (commonly controlled legal entities) is... $5 million per year in 2010.
$5 million a year for as many free H.264 video downloads (over 12 minutes) as Google has the capacity to host.
License terms.
Five years. 10% increase cap on renewals.
SUMMARY OF AVC/H.264 LICENSE TERMS [mpegla.com]
Re:What's an "industry-recognized standard"? (Score:4, Informative)
Ironically, CDMA is the technically better standard, since GSM under the hood (at the time) was actually TDMA, a modulation/channel sharing technique known to have significant sideband emission due to the frequent switching. (also CDMA tends to fail gracefully, by steadily increasing the error rate as the channel becomes fuller, rather than simply dropping calls.)
errata (Score:3, Informative)
$5 million a year for as many free H.264 video downloads (over 12 minutes) as Google has the capacity to host.
Strike this.
The $5 million cap applies to mega-corporations like Disney distributing content through many commonly owned channels and services.
Re:Uh huh (Score:1, Informative)
One thing I'd like to see (besides the abolition of software patents altogether) is being able to demand of a patent holder that they tell you whether your product infringes on their patents.
That way, for instance the Linux foundation could demand of Microsoft, or the Xiph.org foundation could demand of the MPEG licencing consortium, that they give a list of patents they infringe on, and they'd get a reasonable time (say, a year) to respond.
That would go some way towards both alleviating the submarine patent problem, and stopping FUD tactics with vague threats about possibly suing in the future if a competitor gains traction.
F.U.D. (Score:5, Informative)
And stating that it's relevant to video sharing sites- they're an enabler, but YOU are the one on the hook, not they (because there's yet another license THEY have to have to do what they're doing...) and you're still needing that license in addition to the one they're paying.
Shorts under 12 minutes long are royalty free.
Period.
Amateur or professional production.
Free or paid distribution. It doesn't matter.
Royalties on SALES of disks or downloads are 2% of the retail price or 2 cents a title, whichever is LOWER.
MPEG LA doesn't give a damn about your wedding videos.
Subscription services with less than 100,000 paid subscribers are also royalty free. Your "viewer supported" Free Culture magazine on DVD+R is a go.
Own a cable service or TV station in a market of less than 100,000 households?
The one time fee for an AVC transmission encoder is $2,500.
SUMMARY OF AVC/H.264 LICENSE TERMS [mpegla.com]
Re:No. Just pay up (Score:3, Informative)
Unless format B is lossless, you will degrade the quality when transcoding from format A to format B. By definition, a lossy encoding will result in the loss of some information [wikipedia.org] from the source video. I haven't experimented with this, but I would guess that codecs that are similar to each other (i.e., both lose similar information) would result in less degradation with an output at roughly the same bitrate. If the codec is completely different (loses information unrelated to the information that the first codec discarded), I would imagine you would need to increase the bitrate closer to the point of being lossless in order to maintain the quality.
Re:MPEG-LA prevents non-commercial use (Score:5, Informative)
And submarine patents do exist
The term submarine patent [wikipedia.org] originally referred to a scenario in which you'd file a patent, let the procedure stall indefinitely, and only complete the process once you have someone to sue. Formerly in the U.S., you'd then get the full patent term starting from when the patent was actually granted, which could give you extra years of patent life. It also meant that nobody could possibly know about the patent, because it didn't get published until the end. But you still got most of benefits of having the patent.
The rules for granting patents in the United States were changed years ago, so this is no longer possible. Submarine patents thus do not, strictly speaking, exist anymore. But the term has caught on to just mean "patents no one knows about", and in that sense of course you still do have submarine patents.
Re:No. Just pay up (Score:3, Informative)
But I don't get how information that is digital is supposed to "rot" or somehow degrade simply because you are taking it from format A to B unless the settings in B are less than A. Care to explain and back it up with citation?
I can explain it. While digital information can be copied as many times as you want and still be identical to the original, the same is not true about some transcoding.
There are two compression types - lossy and lossless.
Lossless compression (zip, rar, flac, lagarith) compresses the original data so as to remove redundancies, but still have all of the information, for example, you can compress "AAAAAAAAAA" into "10A" and save 7 bytes, but still be able to expand back into the original form. However, this type of compression does not achieve high ratios.
Lossy compression throws some of the original data out. This is done to increase the compression ratio, but it is now impossible to expend back int othe original, some data will be missing. MP3 file sounds different than the original WAV, while a FLAC file sounds the same. This is because mp3 is lossy compression. However, mp3 file uses less space than flac file.
Video uses a lot of data, RGB 1920x1080 video takes 5.9MB per frame and there are at least 24 frames per second, so uncompressed HD video takes ~142MB per second (SD 720x576 would be ~28MB/s). Lossless compression does not help here much, you need to use lossy compression. And all widely used video codecs are lossy, they throw out a lot of information to make the 1080p HD movie fit in a Bluray disc or ~11GB file. Or to make a SD movie fit in a DVD or even in a CD.
Lossy codec creates artifacts, that is, details that are not in the original video and are just the product of the codec throwing out information. When you watch a low bitrate video that has a lot of motion you get blocks. Blocks are one type of artifact.
Now, when you try to transcode, the first thing you do is decode from the original codec to uncompressed video (which can be passed straight to the other codec without storing it first). The encoder looks at the decoded video and again tries to throw out information to make the result fit in the specified space. This encoder has no way of telling which details in the video are real and which are just artifacts, therefore, it tries to encode all of them just as good. In the encoding process it produces its own artifacts and now you have a video that is of lower quality. A third generation video will be even lower quality.
This is the same as with analog, where each generation has not only the noise and distortions of the medium that it is recorded on but also noise and distortions of all of the previous generations.
Only if you want to contribute it (Score:3, Informative)
"I'm sorry sir, your idea has become too successful, we're going to have to take it away from you now."
You've got a mistranslation, here. It's more like:
"I'm sorry sir, you contributed your patented idea to standard X, you're not allowed to use your patent against anyone implementing X."
or even:
"I'm sorry sir, you contributed your patented idea to standard X, you're not allowed to use your patent against anyone using the reference implementation of X."
So, in short, nobody would have the force of their patent involuntarily stripped, it'd be a voluntary part of participating in some standardization activities.
At a minimum, I suspect it'd be helpful if someone came up with some good, marketable shorthand for a standard that is "open" in these ways.
Re:Back in 1988... (Score:1, Informative)
Maybe it is the simple fact that saying: ... would be impolite.
"You're wasting everyone's time with a pipe dream. Please return to reality, this is not your hazy dorm room, we have jobs and want to get back home."
Reality is not a schoolhouse rock song, you are not going to petition the government during that meeting. No one in that room is going/can't to call a high level EU meeting to deliberate merits of patent law. Reality is a little more complicated and using up time in a meeting talking about make-believe would have zero results on the end of the day.