Sisvel patent pools formed for VP9 and AV1

[nevcairiel]

Originally Posted by mandarinka

If you say that to somebody who might have a legitimate claim to the IP

But they stated themselves that they don't even know yet what patents are applicable. That means there could also be none. So they are charging money before they even know if they have a case to do so. How is that good practice, or not trolling?

[mandarinka]

Originally Posted by nevcairiel

But they stated themselves that they don't even know yet what patents are applicable. That means there could also be none.

Eh, I don't think that is quite correct. The list is not published/finalised, so it's probably a case of the usual business communication "we don't commit to anything yet". There are four major companies quoted as backers in their press release, so it is safe to assume that their IP for example will land there.
They say the list will be published when the call for other pool members ends...

[bstrobl]

VP9 has already existed for several years. Seems kind of convenient for them to pop up less than a year after AV1s introduction, while covering VP9 in their terms as well. If it just covered AV1 it may not seem that out of place but at this stage they have had enough time, especially since VP9 already had other patent pools pop up to try to claim their rights.

[soresu]

AV1 was openly developed, with all tools expressed in the specification.

The standardisation process towards bit sream freeze was telegraphed well in advance for any interested patent parties to have found applicable patents and demanded compensation or withdrawl.
For this consortium to have waited until software and hardware was well into development implies a definitive desire to increase the pressure on AOMedia contributors in order to defray the cost of said software and hardware development.

I think it does bare in mind remembering that not only Google was supposed to be contributing legal protection to AV1 by reviewing patent IP, there were other significant parties involved like Cisco and Microsoft.
Wasnt ANS left out of AV1 because of precisely this problem?

[soresu]

I guess at least there is no small corporate backing behind AOMedia to provide legal assistance here, the founding members alone can contribute a very significant defence if needed, and I would assume their own IP review pre bit stream freeze was probably significant enough to address any less than iron clad attempt to shake them down.

[Quikee]

Originally Posted by soresu

Wasn't ANS left out of AV1 because of precisely this problem?

No.. it was left out because it wasn't a good fit for hardware. But Google filed a patent for "ANS in video compression", which is what pissed people of.

[utack]

https://aomedia.org/the-alliance-for...dia-statement/

[mandarinka]

Originally Posted by soresu

Wasnt ANS left out of AV1 because of precisely this problem?

I don't think so. What I heard is that the guy behind it wanted to contribute it, but actually, Google at the same time he was in the AV1 development discussions with them, they tried to patent his idea for themselves behind his back (like, wtf?!). I'm not sure if he even succeeded at preventing that, but if the snafu kept ANS out, it was not the inventor's fault.

[dapperdan]

https://xiphmont.dreamwidth.org/84214.html

Xiphmont's post about the ANS patent thing has all the nuance that gets lost in the retelling of this story.

[soresu]

Interesting read about the performance claims not adding up, I got the initial impression that it was some big advance in efficiency over CABAC with less complexity.

As you say it pays to read up on these things, and yes I always did get Monty's impression that Google's patent filing was specifically a pre-emptive protection against patent trolling over AV1, as indeed alot of AOMedia's mandate was during AV1s initial development.

[mandarinka]

Originally Posted by dapperdan

https://xiphmont.dreamwidth.org/84214.html

Xiphmont's post about the ANS patent thing has all the nuance that gets lost in the retelling of this story.

For saying how he doesn't want to defend Google he sure seems to. It baffled me a bit when he basically addressed the whole matter by telling Duda to shut up go smoke cannabis instead. Maybe it's the classical case of poor joke that wasn't really meant to offend but was too dumb, but well... it involuntarily kind of started shifting my impression of the guy as I read that.

Also I think that the compression efficiency or usefulness of the algorithm has zero importance in the matter of Google's patenting attempt at all, so why bring it up.

[schweinsz]

Originally Posted by mandarinka

I'm not sure it's as clear-cut. For example, the alt-ref thing in VP9/AV1 is really a copy of b-frame mechanism that is designed in a roundabout way to not look totally like b-frames, at a glance. Is that an example of "quite happy to avoid IP that's not available under terms they can accept" in a good sense? Sounds more like use the idea but dodge the responsibility to me, but well, maybe they changed enough details to get away with it.
As for the unclearness - it's perfectly possible the patents just weren't discovered or they weren't careful enough in trying to stay clear. I don't think it's a responsibility of patent holder to reach out to everybody in the world to informa them about it, that's pretty much not viable after all.
Also you call a video compression patent a monopoly and paint is as some injustice. But look at it from the inventor's point of view. They invested money and work. They are probably willing to offer it to multiple parties and not having a monopoly. But Google/AOM will only accept it for free. How can you think that refusing to grant the technology to them in such a situation is unjust? AOM's insistence of being royalty free is their choice, they have no right to push that onto IP holders that aren't okay with that. It's normal to want to be compensated for your property/work.

I think we are starting to hijack this thread though, so better stop this debate (which has been here over and over in the past).

The b-frame is invented more 30 years ago, while the multiple references and the multiple hypothesis inter prediction are invented more than 20 years ago. To the best of my knowledge, the hierarchical b coding structure is invented more than 24 years ago.

[hajj_3]

HEVC patent pool news: https://www.streamingmedia.com/Artic...ls-136123.aspx

Also 'unifiedpatents' has now challenged a total of 42% of Velos Media patent pool's h265 patents: https://www.unifiedpatents.com/insig...y-unpatentable

[hajj_3]

Sisvel has published a list of their patents.

List of VP9 patents: https://www.sisvel.com/images/docume...ntList_VP9.pdf
List of AV1 patents: https://www.sisvel.com/images/docume...ntList_AV1.pdf

[hajj_3]

2 companies have now licenced sisvel's patents: https://www.sisvel.com/news-events/n...nsing-platform

[benwaggoner]

Originally Posted by hajj_3

2 companies have now licenced sisvel's patents: https://www.sisvel.com/news-events/n...nsing-platform

I note that Mitsubishi Electric is the original holder of all the Xylene S.A. patents asserted in the docs, which are a good chunk of them.

I don't recall having heard about Tremmen Tecnologica. Their site is Spanish only and they don't have a Wikipedia entry even.

[hajj_3]

Interdigital has filed a lawsuit against Lenovo regarding VP9 & AV1 patent infringement: https://insight.rpxcorp.com/litigati...ments/15358454

These are the patent numbers and anticipated expiration dates (according to google patents):

US10,250,877 - 2035-05-22
US8,674,859 - 2030-06-22
US9,674,556 - 2027-10-25
US9,173,054 - 2029-12-22
US8,737,933 - 2029-12-22

[Zarxrax]

From: https://insight.rpxcorp.com/litigati...ments/15358454

The ’877 Patent is entitled “Method and Device for Coding an Image Block,
Corresponding Decoding Method and Decoding Device” and issued on April 2, 2019, to
inventors Philippe Bordes, Pierre Andrivon, and Philippe Salmon.

From: https://en.wikipedia.org/wiki/AV1

On 25 June 2018, a validated version 1.0.0 of the specification was released.

Sounds legit.

[rwill]

'Issued' does not equal 'Filed' Mr. Gump.

[Avclover]

Interestingly, a new patent pool called Avanci video announced last week that they will ask video streaming operators to take a license for AV1 among other modern day codecs. 25 licensors are listed...

https://www.avanci.com/video/