Another Patent Pool Forms For HEVC

An anonymous reader writes: A new patent pool, dubbed HEVC Advance, has formed for the HEVC video codec. This pool offers separate licensing from the existing MPEG LA HEVC patent pool. In an article for CNET, Stephen Shankland writes, "HEVC Advance promises a 'transparent' licensing process, but so far it isn't sharing details except to say it's got 500 patents it describes as essential for using HEVC, that it plans to unveil its license in the third quarter, and that expected licensors include General Electric, Technicolor, Dolby, Philips and Mitsubishi Electric. The group's statement suggested that some patent holders weren't satisfied with the money they'd make through MPEG LA's license. One of HEVC Advance's goals is 'delivering a balanced business model that supports HEVC commercialization.' ... HEVC Advance and MPEG LA aren't detailing what led to two patent pools, an outcome that undermines MPEG LA's attempt to offer a convenient 'one-stop shop' for companies needing a license." Perhaps this will lead to increased adoption of royalty-free video codecs such as VP9. Monty Montgomery of Xiph has some further commentary.

Comments

How is hardware VP9 support developing?

[Anonymous Coward]

I'm involved in developing a service with video functionality and we would love to switch from h264 license hell to VP8/9 (the amount of time and money we've spent on legal on this issue is infuriating, to say the least). The main problem isn't that it is lagging somewhat on quality/bitrate, but the availability of hardware decoding support across common mobile devices. Especially on mobile devices hardware decoding is required, not only for performance but at least as much for power usage. Anyone have any insights into how this is developing?

Re: How is hardware VP9 support developing?

[chowdahhead]

WebM has a royalty-free hardware VP9 decoder and encoder. I don't know if any SOCs are incorporating it yet but Tegra X1 has 4k VP9 support supposedly.

Re: How is hardware VP9 support developing?

[Anonymous Coward]

But that's true for all codecs, so what's the difference? You might as well use the VP9 now because it has no obvious license problem. In contrast, HEVC now has two competing patent pools and uncertain licensing requirements in the future.

So You are Saying

[wisnoskij]

That to use a simple, SINGLE, encoding algorithm like HEVC takes licencing thousands of patents?

Re:So You are Saying

[Aighearach]

Most algorithms are not patentable in the US either, post-Alice. I don't doubt thousands of patents have been issued, but that doesn't automatically make them valid.

Re:So You are Saying

[HiThere]

No, but it does mean you risk going bankrupt fighting an invalid patent.

Re:So You are Saying

[king neckbeard]

That MPEG2 had hundreds of patents would suggest that there is a problem. That makes it sound as if basically every step had at least one patent, possibly more. If that's the case, then meaningful competition is going to be impossible.

Re:So You are Saying

[ColaMan]

"method for adding two numbers using instructions and registers commonly available on x64 processors"

Re:So You are Saying

[Dutch Gun]

To answer your question: Yes. Whatever you think of patents (personally I despise software patents and think they're a cancer on our industry), these are not single algorithms, nor are they in any way simple. This is very sophisticated software. At least scan through the Wikipedia entry linked in the summary to get a rough idea of the complexity of these monsters.

Modern video formats are comprised of a vast collection of different algorithms and techniques, and part of the encoding process is determining how best to apply those various techniques to create the best compression while maintaining a specific desired perceptual quality. It's perhaps best to think of a video codec as a family of many different video encoding, decoding, and storage techniques.

Re:So You are Saying

[tlhIngan]

That MPEG2 had hundreds of patents would suggest that there is a problem. That makes it sound as if basically every step had at least one patent, possibly more. If that's the case, then meaningful competition is going to be impossible.

Every standard has patents. When you make standards by committee, whether it be video encoders, Wi-Fi, Ethernet, what have you, it's really a give-and-take of patent and technology holders trying to squeeze their thing into the standard. Sort of the "I'll let your patent go in, if you'll vote for mine in too"

Now, MPEG patents are typically not a problem - because MPEG set up the MPEG-LA that serves as a patent pool - instead of trying to negotiate with 500 patent holders on licenses, you go to the pool administrator, pay the appropriate fee according to the fee schedule and walk away with the licenses.

It's why you rarely see patent holders suing licensees over patent violations - Motorola suing Microsoft being about the only one for h.264 - every other licensee has paid MPEG-LA and thus licensed the appropriate patents.

Cellphone companies, though, haven't created a patent pool which is why you have everyone suing everyone else.

This HEVC alliance will be interesting because it'll be interesting if them and the MPEG-LA will respect each other's licenses. If not, this could easily tank HEVC if people can't figure out how they're supposed to license the patents. VP9 could easily squeeze in the gap despite HEVC being acknowledged as the next-gen codec - but if the licensing agreements can't be worked out between MPEG-LA and HEVC-A, then people wouldn't dare use it.

Re:So You are Saying

[Ben Hutchings]

MPEG set up the MPEG-LA that serves as a patent pool

No, they did not. There are a lot of the same players there, of course, but they have no formal relationship.

Re:So You are Saying

[fsterman]

... these are not single algorithms, nor are they in any way simple. This is very sophisticated software. At least scan through the Wikipedia entry linked in the summary to get a rough idea of the complexity of these monsters.

I actually read through some of the patents Nokia was threatening VP8/9 with and they really are not sophisticated at all, they are just written in the most confusing possible way. For example, the following paragraph is from a Nokia patent that basically describes the selection of neighboring pixels:

selecting a first reference video pixel in the first video block and a second reference video pixel in the second video block, the first reference video pixel and the second reference video pixel being other than the first boundary video pixel and the second boundary video pixel and the first reference video pixel and the second reference video pixel being placed closer to a central portion of each of said video blocks than the respective boundary video pixel, in such a way that the reference video pixels and the boundary video pixels are situated on a straight line, the straight line being transverse to the boundary, drawn from the first reference video pixel to the second reference video pixel, wherein the first and the second boundary video pixels are located between the first and the second reference video pixels on the straight line,

I was planning on busting all of the Nokia patents myself, but then I got busy :p

Re:So You are Saying

[Dutch Gun]

I actually read through some of the patents Nokia was threatening VP8/9 with and they really are not sophisticated at all, they are just written in the most confusing possible way.

Oh, don't misunderstand... I'm betting that what was patented is actually not all that complex in principle. And naturally, being patents, they're written as broadly and confusingly as possible. That doesn't mean the software as a whole is not extremely sophisticated. Try reading an open source video or audio codec and you'll see how complex it really is in practice.

Great

[Anonymous Coward]

More patent encumbered crap.

we need to stop feeding this legal beast.

Let the stupidity continue

[JonathanP.Bennett]

If enough of this sort of stupid continues, perhaps the backlash will be enough to finally kill the software patent beast.

Daala

[mlkj]

May Daala save us all.

Re:Daala

[Burz]

Vorbis made it into a lot of products that were not Apple or MS, from Sandisk to Samsung.

Daala is shaping up to be excellent as well, but its biggest competition may be VPx in the long run... Google announced they would start 18-month release cycles for major VPx codec revisions after 10. That creates a Chrome-like effect on the mindshare of early adopters, so it should be interesting. Of course, who is backing Daala? Mozilla... who may get dragged into release-cycle competition with Google on another product. :)

Re:Daala

[serviscope_minor]

Google announced they would start 18-month release cycles for major VPx codec revisions after 10. That creates a Chrome-like effect on the mindshare of early adopters, so it should be interesting.

I'm skeptical that that is a good thing: video encoding is a closed, well defined task compared to web browsing. It's also reached the stage where the changes are somewhat incremental. HEVC is a step on from H.264, but the gains aren't immense for instance. Also, modern codecs are encoder heavy which means that given the decoder algorithm there is a lot of leeway to improve the encoding by working on the encoder. By way of example, look how much the lowly MP3 128kbit/s MP3 improved over the years as encoders got better.

What this means is there's not all that mch advantage to changig the codec every 18 months. In fact, there are a lot of disadvantages. Time was unless you wrre targeting Linux (which had/has FFMPEG), getting a video to play on a random platform was an exercise in frustration, especially if Windows and MacOS were involved.

Eventually everyone settled on MP4. I can now create an MP4, and it will play almost anywhere, on any PC operating system, on any mobile one, including "ancient" versions of android which are still common in the wild, a Raspberry PI, a web browser, with or without a flash plugin. Oh actually even random set-top boxes and TVs will play MP4s from a USB memory device.

An 18 month release cycle for codecs will not allow for that kind of ubuquity.

I think there's also the other thing: MP3 is still really common. It's not the best audio codec, and there's a suite of others which handily outperform it in listening tests. However, unless you have golden ears and excellent headphones, MP3 is nearly perfect now at 128kbit/s. Storage has also been increasing, and since our ears haven't been improving it's reached the stage where the comibnation of quality, size and storage capacity means there's not much motivation to change something so ubiquitous.

I think the same will happen to H.264. Encoders will improve to squeeze the most out of it. Ultimately, quality is tailing off: in many cases you can get good quality on the majority of screens without insane storage costs. As storage increases, the motivation to switch codecs will diminish.

Tha's not to say new codecs won't have their place. It's just that I think H.264 has hit the point where entrenchment is possible (previous ones were severely lacking) so newer codecs will always be somewhat limited to niches.

DVD patents expiring

[crow]

At least the patents on DVDs are expiring if not already expired. The first DVD player was sold in 1996, and patents can be good for up to 20 years from the filing date, so it would seem that by late next year, all necessary patents should have expired. (Patents are only 17 years from the issue date, so any patents that were actually issued at the time of the first players would have expired.)

I'm sure that they've added on patents for various RW formats, and probably for some new tricks in encoding, but that wouldn't impact playback.

MP3 patents have mostly expired, though one US patent expires later this year.

So for any application using MPEG-2 or MP3, you shouldn't be facing a big patent hurdle. If you want the lower bitrates found with newer codecs, the pain will be with us for a while to come.

Re:DVD patents expiring

[evilviper]

At least the patents on DVDs are expiring if not already expired. The first DVD player was sold in 1996, and patents can be good for up to 20 years from the filing date, so it would seem that by late next year, all necessary patents should have expired.

This is HORRIBLE legal advice. Patent laws were different before 1996, that's why MP3 patents are still around (and will be until 2017) despite the fact that specifications were published back in 1991!

In the United States, "patents filed prior to 8 June 1995 expire 17 years after the publication date of the patent, but application extensions make it possible for a patent to issue" quite a few years after initial filing.

MP3 patents have mostly expired, though one US patent expires later this year.

I wish that was true, but it's certainly not:

http://www.tunequest.org/a-big...

Re:DVD patents expiring

[crow]

There are other MP3 patents that don't expire this year, but their validity is in question.

And the surviving patents are, in most cases, USA only.

Re:Its a shame WebM sucks

[SeaFox]

Has anybody tried the WebM encoders? They STINK, its either a bunch of CLI gobbledygook or its some half assed support in some other encoder.

If you've talked to anyone on the pirating scene (the ones who actually know what they're doing), controlling x264 through command line is normal. Trying to use a GUI on an encoding test for an anime fansubbing group gets you laughed at.

After H.265

[sonicmerlin]

What comes next? H.266? Is anyone working on it? Is it even possible?

Re:After H.265

[Kjella]

What comes next? H.266? Is anyone working on it? Is it even possible?

Of course it is, the question is if anybody will care. We know there are many better image compression routines than JPEG like JPEG 2000, JPEG-XR and WebP, but it's "good enough" nobody cares and it is now absolutely guaranteed patent free. Same with PNG, there are arguably better compression algorithms but it works. Everywhere but the US the MP3 patent has expired and in 2017 the last patent will expire there too. In 2018 the MPEG2 patents including AAC end, which I think will make AAC the de facto codec for lossy audio.

Video is another ball of fur, there's Theora and VP8/9 and WebM but none have gained any real traction and from what I understand not even MPEG2 is patent free yet while H.264 is 10+ years from becoming patent free. In that kind of marketplace what's another patented codec? I assume the gains for each generation will be less though and with bandwidth and disc capacity increasing we might not care that much anyway. I know it's not the JPEGs slowing down my browsing experience...

Re:After H.265

[sanf780]

The problem is HW support. Video is so computing intensive that modern CPUs have trouble keeping the bandwidth. Even if you could run video encoding or decoding on a CPU, it probably is more efficient to run on HW. So, that is the reason why people stick to an old format for a long long time. H264 and MP3 are not going to disappear next week.

Re:After H.265

[serviscope_minor]

In 2018 the MPEG2 patents including AAC end, which I think will make AAC the de facto codec for lossy audio.

I doubt that. Vorbis is at least as good as AAC and very popular: I think just about every game out there now encodes audio in Vorbis format and it's patent and royalty free and yet MP3 is still king. The thing is MP3 is good enough:

http://listening-test.coresv.n...

If you're prepared to spend 30% extra on storage space (and a bit less on decoding power), MP3 is as good as the better codecs. Storage space is increasing year on year, audio perception stays the same. There are always a few people who have vast audio collections, but most of us don't. IOW, it's mostly passed the point where that 30% matters, which is why MP3 seems to be entrenched for the long haul.

Re:Its a shame WebM sucks

[Anonymous Coward]

Anyone doing any kind of serious work with an encoder (bulk especially) would laugh at you for the "gobbledygook" comment. In order for people to be able to enjoy videos in a format, we need these tools to give them things to watch. Without efficient CLI commands, your codec is nigh on useless. And for personal use it's also largely irrelevant; users use what their camera does, which is probably calling that CLI command, or the even more technical APIs those interfaces expose. GUI work for video editing is important, but once you have your clip you still need those tools for your GUI to even function.

Broken window fallacy

[tepples]

Anonymous Coward wrote:

[Patent FUD] encourages the use of older higher-bandwidth codecs which encourages provision of higher bandwidth internet connections.

Textbook broken window fallacy.

Re:Broken window fallacy

[Kjella]

[Patent FUD] encourages the use of older higher-bandwidth codecs which encourages provision of higher bandwidth internet connections.

Textbook broken window fallacy.

Isolated speaking, yes. However, you can consider it a cross-subsidy to enable other and presumably more worthy causes for high speed broadband than watching YouTube. Or you can assume that at some point we'll want higher bandwidth anyway for 4K TV so using a less efficient codec now means you're doing most of the roil-out for later when you'll use a more efficient codec. It's hardly unusual that creating less favorable conditions for some individuals may benefit the group as a whole or that different short-term incentives benefit the long term result.

Personally I consider broadband to be the electricity, telephone and running hot water of the 21st century, if you're not online you're more or less detached from contemporary society. Not that we all need to be on gigabit fiber, but dial-up just isn't cutting it anymore. Not like necessity of life because people lived without 100+ years ago and I guess in poorer parts of the world many do but it'd just not the way I'd like to live now.

duh. it's called greed!

[Gravis Zero]

The group's statement suggested that some patent holders weren't satisfied with the money they'd make through MPEG LA's license.

so there are 500 patents and at least one of the patent owners is an insufferable greedy asshole? what's next, are bears going to start shitting in the woods too?!

Re:Its a shame WebM sucks

[Anonymous Coward]

Only some of us, I'm straight, thank you very kindly(I might be bi if he's cute enough). Anime fans have a tendency to be weird, but we also have a tendency to have contacts across the globe with translators that are practiced and able to deal with common idiom and work very quickly. You try finding another group which can record a 30 min program, translate it, sub it, encode it, and upload it in about 2 hours. I've found a number of them among anime fans, and some of them release multiple languages all at once.
I'd value their opinion on how to translate and subtitle a video fast, I'd value their opinion on dealing with encoding formats, I'd value their opinion on a number of things, once you got done weeding out the possers and weabos.

why does the poster thing this helps VP9?

[YesIAmAScript]

How do you call VP9 royalty-free in the same article as the rest of this info.

There is not currently a patent pool for VP9. That doesn't mean it's in a better position than HEVC, given there could be a "freelance" patent pool for VP9 any day now.

Any standard which becomes successful attracts leeches. VP9 is no exception.

Re:why does the poster thing this helps VP9?

[Anonymous Coward]

Any standard which becomes successful attracts leeches. VP9 is no exception.

So, if this is true for all codecs, then why would you choose the codec with an obvious case of licensing confusion and uncertainty, such as HEVC has now. Why wouldn't you choose the codec with clear licensing and merely the possible chance of licensing problems, as VP9 has now. In terms of licensing VP9 is plainly the better bet.

Re:why does the poster thing this helps VP9?

[fsterman]

How do you call VP9 royalty-free in the same article as the rest of this info.

There is not currently a patent pool for VP9. That doesn't mean it's in a better position than HEVC, given there could be a "freelance" patent pool for VP9 any day now.

Any standard which becomes successful attracts leeches. VP9 is no exception.

How do you call VP9 royalty-free in the same article as the rest of this info.

There is not currently a patent pool for VP9. That doesn't mean it's in a better position than HEVC, given there could be a "freelance" patent pool for VP9 any day now.

Any standard which becomes successful attracts leeches. VP9 is no exception.

Carefully avoiding all known patents puts them into a better position, even if the position is just a smaller number of patents.

End User

[Whiteox]

OK So I'm not into this very much. I use some splitters and re-encode ripped video on the odd occasion with Handbrake. What got me though with HEVC is that the requirements for playback needs a multi-core processor.
I downloaded a 30 min video file (FTA torrent) and I was surprised with the smaller file size (about 30-40% improvement), but pissed off at not having a player for it. After searching around, I got VLC updated to play the file which looked promising at first. Unfortunately, it bombed out as it lost the audio/video sync within a few minutes and started to jitter.
Undie-turred, I downloaded a different source and found the same issue. The files just wouldn't play on a single processor successfully.
ATM I can't support this codec as I have a few old, single core laptops in bedrooms that are used to play video from my home server.

This guy thinks VP9 is winning already:

[Burz]

http://www.nojitter.com/post/2...

That stat about VP9 meeting 60% of Youtube delivery is interesting.

Re:Its a shame WebM sucks

[Kjella]

I've can't decide if you're a troll or just lack sane opinions, you seem to hate on most things except AMD for which you have a major boner. The average person doesn't use an encoder, ever. The only reason they care about decoding formats is because they download stuff off the Internet want their MKV to work on their gizmo, not just their computer. Both "DivX 3.11 ;-)" and MKV gained popularity that way.

Ordinary users upload videos to YouTube, but they don't have any say in what codec/settings/resolution/bitrate Google chooses to use. The people who edit video want it in their save/export dialog of whatever editing software they use, which Google can do fuck all about and those who transcode for "the scene" are 99% fine using a CLI. They just don't give a shit about legality so if WebM beat H.264 they'd use it and it'd be popular. Absolutely nobody cares about encoder GUIs.

Re:Its a shame WebM sucks

[SeaFox]

DaFuq? What does anime piracy have to do with shit, you like the majority eats Pokki and watches...sheeit, can't even name any of that crap for a comparison...uhhh...Battletech?

Again look at the formats that HAVE taken off...what do they have in common? EVERYBODY CAN USE IT, sure there are tools that can use CLI, fuck I'm sure you can encode MP3s in CLI...but nobody give a fuuuck, the majority are using easy GUIs...of which jack and shit exists for WebM. Fuck even the last Handbrake I checked had HEVC support...no support for WebM.

"Everybody can use it" isn't about creating, it's about consuming. If you want your format to take off you need it supported on consumer electronics devices. Again -- look at pirates. You know why h264 took so long to take off? It's because there weren't a whole lot of stand-alone devices that supported it several years back. Same with MKV container format. Been around for quite awhile, not supportted commercially. This is why the whole live action pirating community continued to put out Divx/AVI files like it was still 2002 until a few years ago. It was because there was a large installed base of DivX DVD players and first gen consumer electronics with PC file playback that wouldn't support MKV containers, and anything more than Main Profile h264.

The anime fansubbers didn't care about standalone players because once they stopped hardsubbing they required a PC for playback anyway (for SSA subtitle rendering). This allowed them to keep following new advances in encoding -- they weren't having to maintain backwards compatibility with people still running XBMC on actual original XBoxes. So they changed to h264/MKV back in 2005 and a few years back started doing shit in 10-bit h264. But they still do it in CLI, and the high-quality live action pirates, too. Nobody is wailing for a GUI but people too lazy to learn to type.

WebM ain't taking off because h264 has already established itself as a standard outside a web browser. It's the convenience to the users that counts.

Greed vs Greed

[John Allsup]

It's really just a bunch of selfish caveman playing with these magical Toys the modern world has, and stubbornly fighting over territory. FFS the world needs to grow up, and this is why.

Re:Its a shame WebM sucks

[fsterman]

I think there are two distinct worlds, people who handle the distribution of video and content creators. For content creators, they need highly polished GUIs ... like those provided by Final Cut Pro and iMovie. There are also batch video conversion tools that are entirely oriented around the GUI.