Cisco Releases Open Source "Binary Module" For H.264 In WebRTC 95
SD-Arcadia writes "Mozilla Blog: 'Cisco has announced today that they are going to release a gratis, high quality, open source H.264 implementation — along with gratis binary modules compiled from that source and hosted by Cisco for download. This move enables any open source project to incorporate Cisco's H.264 module without paying MEPG LA license fees. Of course, this is not a not a complete solution. In a perfect world, codecs, like other basic Internet technologies such as TCP/IP, HTTP, and HTML, would be fully open and free for anyone to modify, recompile, and redistribute without license agreements or fees. Mozilla is fully committed to working towards that better future. To that end, we are developing Daala, a fully open next generation codec. Daala is still under development, but our goal is to leapfrog H.265 and VP9, building a codec that will be both higher-quality and free of encumbrances.'"
Good luck with that... (Score:5, Insightful)
Re: (Score:3)
Why not sell full h.264 converters in countries other than USA, Australia, New Zealand and the other handful of countries that have such software patents?
Bring the datacentres to the UK or other EU countries.
At the same time Mozilla set up bank accounts and trading addresses in Europe and release a browser with h.264 support. Companies get away with it for tax, so why not for patents?
Re: (Score:2)
Why not sell full h.264 converters in countries other than USA, Australia, New Zealand and the other handful of countries that have such software patents?
I didn't think New Zealand had software patents.
Re:Good luck with that... (Score:4, Insightful)
Re:Good luck with that... (Score:4, Informative)
Legally, there's a reasonable limit on how long you can wait (6 years under some theories). That being said, indeed can can never prove non-infringement, and it's equally valid for free codecs than it is for encumbered codecs. Paying the MPEG LA tax does not shield you from trolls, or even from companies that participated in the standard and aren't part of the patent pool (usually, not all declared IPR holders are represented in a pool).
Re: (Score:2)
Legally, there's a reasonable limit on how long you can wait (6 years under some theories).
There have been codec legal fights much longer than 6 years after a patent was granted.
Laches (Score:3)
Good luck with that double edged sword. (Score:2)
I often wonder how you can guarantee something doesn't violate any prior art. Since there's no requirement for how long an existing "invention" can wait before the patent applicant "all-of-a-sudden" independently discovers that an existing idea has never been pushed through to be patent process, and since there are so many ideas out there, it would be quite hard for there to be a guarantee that something didn't think of something first. "Undisclosed source code" as it's called happens all the time. You don
Re: (Score:2)
Re: (Score:3)
You are forgetting Daala is developed at the IETF and Mozilla by some of the same people that made the patent free Opus audio codec.
Which really is 'best of breed':
http://www.opus-codec.org/comparison/ [opus-codec.org]
So have I have at least some fait.
Re: (Score:2)
So to prove this Opus codec is the best, you link to... the website opus-codec.org? I could employ that same methodology to prove that Surface is the best tablet ever, the NSA only does what it needs to keep us safe, and that cigarettes are beneficial in many ways.
Re: (Score:1)
So to prove this Opus codec is the best, you link to... the website opus-codec.org?
"Prove"? He's clearly just directing readers to more information about the codec. (He mistakenly thought Slashdot users could read, research, and make their own judgments.) What better place to start than the official website?
Would you rather he linked to Wikipedia or some blog entry? Would that make you feel better? Probably not ...
"So to prove this Opus codec is the best, you link to... the website wikipedia.org? I could employ that same methodology to prove that Surface is the best tablet ever,"
Did I
Re: (Score:2)
There are several reputable third-party listening tests linked at the bottom of the page, below their own, admittedly unjustified, graphs.
Re:Good luck with that... (Score:5, Informative)
I recommend reading Monty's Daala demos 1 [xiph.org], 2 [xiph.org], 3 [xiph.org] and 4 [xiph.org]. We're not just building a similar codec, but making radical changes to many fundamental components of a video codec.
Daala development is just like "good math"... (Score:4, Insightful)
The Daala development is covering new grounds (yes, that's correct), and doing so in a public way. Just like the proof for the Fermat theorem was extremely useful because it created a LOT of new, *good* math (that has applications on stuff as seriously important as the entire field of cryptography) and not because it proved the Fermat theorem, Daala is already important even if the end result ends up not being the best codec under the sun. However, if you go by the result in Opus, it WILL be of extremely good quality.
Doesn't mean we shouldn't try (Score:2)
Thanks for the warm show of support at the end. You're right that it's a huge task and that we'll need all the luck we can get.
Re: (Score:2)
Re: (Score:2)
A modern video codec that exceeds the performance of H.265...? And one that gains the support of hardware vendors to build it into systems? Good luck
There are about thirty licensors of H.264 and HEVC technologies.
Most of them global industrial giants like Mitsubishi, Philips, Samsung, Toshiba, LG. Their principal licensees are on the same scale and for all practical purposes control together they control 100% of a vertically integrated video hardware market.
The geek may have a codec in development. Panasonic and Samsung will have 4K UHD video gear in production.
Why not just build on VP9? (Score:4, Insightful)
Isn't VP9 supposed to be unencumbered by patents anyways?
Re: (Score:1)
Please read up on the history. Short version: "Mokia" blocked a standard not based on h264.
Re: (Score:2)
Because it is most likely not quite as good as h.265, and it is not quite clear it is actually unencumbered, either.
Re:Why not just build on VP9? (Score:4, Informative)
Why isn't this comment already rated -5 Troll? Pathetic even for a troll, 1 second Google search shows it is BSD licensed.
Phillip.
Re: (Score:3)
And you trust a Google search to be impartial in this!?
Go Monty (Score:1)
Huge respect for Chris Montgomery. Instead of using his talents to make a huge pile of money he works so we can have open codecs! I also have watched his video destroying myths about digital audio [xiph.org] about ten times, it's great.
Re: (Score:1)
Bribery, lobbying & politics
Re: (Score:2)
What are those countless others before? Theora was the first serious free codec I can think of.
YouTube uses WebM (Score:2)
Misconceptions (Score:5, Interesting)
reeks, even (Score:1)
agree with your perspective
Re:Misconceptions (Score:5, Interesting)
As I understand it from reading the article and the comments, Cisco will subsidize the patent licenses if you use the binary. If you prefer, you can use the source code, but then you will have to deal with the patent licensing yourself.
"Nathan – We will select licensing terms that allow for this code to be used in commercial products as well as open source projects. In order for Cisco to be responsible for the MPEG LA licensing royalties for the module, Cisco must provide the packaging and distribution of this code in a binary module format (think of it like a plug-in, but not using the same APIs as existing plugins), in addition to several other constraints. This gives the community the best of all worlds – a team can choose to use the source code, in which case the team is responsible for paying all applicable license fees, or the team can use the binary module distributed by Cisco, in which case Cisco will cover the MPEG LA licensing fees. Hope that answers the first part of your question – Nadee, Cisco PR "
Re: (Score:2)
The binary module is by far the more interesting part. There are already entirely capable open-source h.264 decoders. Another one would not be interesting on its own. Getting a fully licensed binary is much more useful.
Re: (Score:2)
They're releasing both the source and the binary... I get the feeling the binary release and the way that Cisco hosts it and Firefox downloads it on-demand has something to do with how the licensing works, that you wouldn't be covered by Cisco's patent payments if you compiled the source yourself.
Re: (Score:3)
Right now, the site has neither binaries nor source, but I'm pretty sure both will be available at the same time. The only point of the binaries there is that since they are served by Cisco, then Cisco can handle the patent licensing. The license is non-transferable (this is not something Cisco controls), so you can't download once and put it into your product, each product has to downloading it on its own for the license to apply. Since it's open source, anyone can also just build it themselves, then then
Good luck, Daala (Score:2)
Looks interesting [xiph.org] and all, but it'll take a lot to convince me it's a serious competitor to H.264.
Re: (Score:2)
It's not a competitor to h.264, it's aiming to be a competitor to h.265.
Re: (Score:2)
You're right. I misread
our goal is to leapfrog H.265 and VP9, building a codec that will be both higher-quality and free of encumbrances.
to mean it intends merely to be a generation ahead of the current-generation truly-open video codecs, which would be a much weaker claim.
My opinion towards it remains the same, though: truly-open codecs don't have a great track record, but I really hope they succeed.
Re: (Score:2)
though: truly-open codecs don't have a great track record
Like opus which beats pretty uch everything else?
Re: (Score:2)
They've been doing well for audio, both Vorbis and Opus are good codecs. But audio is a lot simpler. It's a very different story for video.
VP9 (Score:2)
Re:Open Source Binary Module (Score:5, Informative)
FTA: "Cisco is going to release, under the BSD license, an H.264 stack, and build it into binary modules compiled for all popular or feasibly supportable platforms, which can be loaded into any application (including Firefox)."
From your comment: ..."since it lacks copyleft provisions to actually make the source open."
Looks like the source will be open, since they are releasing the stack under the BSD license. Looks like people will be able to do anything they want with it, including making baby mulchers, angel summoning portals, and *gasp* video player implementations. Oh, HORRORS, people might not submit their code back to Cisco after attributing their source to them (as simply doing so will allow people to find, oh, I don't know, the source that Cisco is offering for free under a BSD license?).
The only issue is with the fact that Cisco is having to provide a shield using the BSD license between MPEG LA and the rest of the world, while paying a hefty licensing fee for the privilege. However, using a BSD license means they cannot have any unreasonable hold over the source once it is out in the open. If anything, Cisco is a good guy in this (god, did I just say that?).
Re: (Score:2)
I'm not saying Cisco isn't being the good guy here. They are. For their own gain, sure, but the result is the same. That's quite separate from my dislike of the BSD license.
Re: (Score:2)
So, by your definition then, public domain is open source?
Re: (Score:2)
The source code is open source but patent encumbered, so you can download it alter, compile and distribute the code acording to the license. Copyright is not a problem here. The problem is the code is covered by about a billion patents owned by the mpeg la group so you cant distribute the binaries or sadly the code freely under a different set of laws not covered by most open source licences. Cisco owns a license for the patents covering this codex, so they are hosting the source code repository and the bin
Re: (Score:2)
Re: (Score:2)
My guess is that they're distributing the binary and paying the fees associated with its distribution. The source code is coming: https://github.com/cisco/openh264 [github.com] but it's only useful to tick the "open source" checkbox. The important thing is that they're paying the fees and that the binary module is under a BSD license.
Re: (Score:2)
There will be 2 ways to use this. You can either download the source code and use it in your code (in which case you have to pay the h.264 patent royalties yourself) OR you can have your program download the binary blob from Cisco directly in which case any download your end users make is covered by Cisco's license for the H.264 patents.
Cisco have also said they will accept (and host) binaries for OSs and architectures where ports are done by the community, meaning that if someone happens to want a binary f
Monty's comments (Score:5, Informative)
Beyond the official announcements, I strongly recommend reading Monty's comments [livejournal.com] on the issue.
Open source still requires license fees (Score:5, Informative)
As pointed out in the comments on the Cisco blog post by a Cisco PR rep, if you use the source code (as opposed to the binary) you are responsible for any resulting licensing fees. Cisco is only covering the fees for those who use the binary.
Re: (Score:2)
That's actually missing a key piece of information. Patent licenses can be very narrow in scope, which allows an owner to charge different parties for different aspects how a device uses a patented technology.
Does anybody know why I wouldn't need a separate licensing deal with MPEG LA if I built a web application using WebRTC? Their typical licensing agreement (as used in Windows and Flash, for instance) does not extend to third-party applications that use the codecs through APIs.
Re: (Score:2)
Why would you need a citation for that? That is how licensing works. Cisco can not realistically offer a license for code you build yourself, only for binaries they provide.
Re: (Score:1)
For the same reason that long division should be free: it's an algorithm.
Re:Why free? (Score:5, Insightful)
Because the H.264 video format is only worth money because of the network and incumbency effects, not because it is better. A video format is a natural monopoly [wikipedia.org]. VP8 is just as good as H.264, and free, but that is not enough to displace H.264 because H.264 has a monopoly via the network effect.
If we were talking about a program such as Photoshop, where the barriers to entry is most determined by your ability to make a better photo editor, it would not be the same thing. There is good reason that the other examples in the summary are "TCP/IP, HTTP, and HTML", all of which are not terribly hard to replace, but which have powerful positions because of the network and incumbency effects.
Re: (Score:3)
VP8 is just as good as H.264,
This is not actually true.
Re: (Score:1)
Re: (Score:2)
Why should a codec be free? Or when you say "in a perfect world" do you really mean "I want it, therefore its wrong if I dont have it" ?
If you reduce barriers to using the technology, then more people can receive its benefits and create wonderful things from it, from which yet more people benefit. TCP/IP, HTTP, HTML, SMTP, etc. worked out well in that regard.
I'm not against patents, but certainly it's better for everyone but the patent-holder if the technology is free.
Re: (Score:2)
Why should a codec be free? Or when you say "in a perfect world" do you really mean "I want it, therefore its wrong if I dont have it" ?
Why should someone who writes the software to implement a codec algorithm not be able to give their work away for free, just because someone claims to 'own' the algorithm because law pixies say they do?
This is very good news! (Score:1)
I especially like how they're not charging for the NSA code either.
Title has the quotes on the wrong phrase (Score:2)
Oh, goody, I can "consume" silent movies now... (Score:4, Interesting)
If this license and "modules" covers encoding as well as decoding, AND The Motion Picture Experts Group Licensing Association doesn't decide to rescind their moratorium on charging license fees for the amazing innovative "actually sending the video over the internet" parts of the patent pool, at least this will allow some basic participation in online using the lower-quality "baseline" profile (only, as far as I know) without paying a poll tax to use their Intellectual Precious.
However, I would also assume this doesn't include AAC or MP3 patents either, so unless "consumers" start using Opus (preferable - Opus is awesome) for their audio codec [and it's packageable with h.264 video in some legally-usable manner], you'll still be limited to providing "silent movies".
If the MPEG License Ass. was serious about killing VP[0-9]+, they'd explicitly waive license fees for any implementation of the h.264 encoding and decoding algorithms that are implemented in "software" intended to be executed on a general-purpose CPU (i.e. no dedicated "hardware decoding/encoding") and which is released under an OSI-approved license (if they really wanted to troll, they could mandate that it be a share-alike license like the GPL, just to make a subset of people throw a tantrum). This wouldn't cost them anything (people with money who are "selling" software and/or making dedicated hardware for encoding and decoding h.264 would still be paying them anyway), but there'd still be a legally-free path for everyone else to participate using h.264 (i.e. Mozilla et al could implement software encoders and decoders to distribute). This would eliminate the need for any of the current-generation "alternative"/free codecs entirely, leaving them only daala to have to compete with later, and completely undermining what little momentum Google has bothered to get going on the VP* codecs.
Given how much sense this would seem to make, though, I wouldn't expect the MPEG License Ass. to even consider it.
Re: (Score:2)
You understand what "WebRTC" is, right?
Re: (Score:2)
Yes...but you've got a point. I hadn't immediately recognized that this was really just for "Internet Video-Telephone" use rather than "Internet Television" (i.e. Youtube). There are a lot of people that seem to also be commenting on the assumption that this has something to do with watching TV on youtube (e.g. the <video> tag)...which it hypothetically could, though it seems obvious that's not what Cisco cares about.
The point about not having proprietary-soun
Re: (Score:3)
Well, keep in mind that an MTI video codec is mostly intended to serve the purpose of preventing complete failures to negotiate. Also, the MTI that 's being proposed in the IETF is H.264 baseline, which is a far sight worse than VP8 by pretty much every metric imaginable. If H.264 baseline is selected as MTI, then I would imagine that the existing implementations will continue to offer VP8 in preference to H.264 baseline, and fall back to H.264 baseline only as an emergency backup "codec of last resort".
As
Opus (Score:2)
Firefox has supported .opus files in <audio> for more than a year now and it works quite well. (I'm a raving Opus fanboy since over a year ago - I started running what would become Firefox 15 during its "Aurora" stage just because I knew it had .opus support. I've had .opus up on my "HTML5 <audio> Test Page [dogphilosophy.net] for quite some time. (I still need to add an .alac sample on there - if it tu
Free binaries + free Backdoors! (Score:4, Interesting)
Re: (Score:2)
Eich often does a good job of intelligently addressing questions in the comments. I strongly encourage looking through them to learn more.
In reply to one question about the binaries he replied:
...because the BSD-licensed source code is available at http://www.openh264.org/ [openh264.org], you and others can verify the compiled bits come from that source, no malware or spyware added. We will organize community auditing of this sanity check, and the binary modules will be cryptographically signed so Firefox can verify their integrity.
And another,
great question, and it applies to Firefox, Chrome, and other browsers. But in the case of Firefox for Linux at least, and for Cisco’s OpenH264 binary modules, we can audit: get matching revision of the open source, compile with the same (bootstrapped from open source) clang or gcc toolchain, and compare bits.
It appears we can have a good amount of confidence that what's in the code is what's in the binary.
The absurdity of the US patent situation. (Score:2)
Modify TCP all you want (Score:1)