Software patents, and patents in general, are a plague on the industry and should be abolished due to a lack of empirical evidence to justify their existence. As a thought experiment they make sense, but thought experiments often diverge from reality.
Patents don’t solve for innovation. What they solve for is time for marketing and market adoption of a novelty, which is what people (investors) mistakenly refer to as innovation.
Here is a debate on IP with two different libertarian perspective: https://www.youtube.com/watch?v=Ep2-ohgFOys `Abolish the IP laws` side provides some pointers to related research and empirical evidence against IP.
I'm hoping to add support for Opus to the upcoming PineBuds Pro, but no guarantees I'll be able to accomplish it at my current skill level. So far I couldn't even get the first SDK to compile (https://wiki.pine64.org/wiki/PineBuds_Pro).
I would have thought that something like that would have been all offloaded to hardware or in a weird DSP assembly (Like how Qualcomm have their newer aptx codecs as Hexagon DSP code), but I see a LDAC decoder as a regular ARM object in there.
Context: Opus was specifically designed to avoid known existing patents, except for ones that are available royalty-free [1]. If Fraunhofer and Dolby believed Opus infringed on their patents they could have said so a decade ago, but instead they've waited for it to become widely deployed.
This is also a peculiar choice from Dolby because the patents that do cover Opus are licensed with an exception for companies that sue saying they have patents covering Opus [2]. Dolby has at least one product that uses Opus [3] so if this pool goes ahead and sues someone then it looks like Dolby is open to being sued themselves.
[2] "If you ... file a Claim for patent infringement against any entity alleging that an Implementation in whole or in part constitutes direct or contributory patent infringement ... then any patent rights granted to you under this License shall automatically terminate retroactively as of the date you first received the grant."
It looks like Fraunhoffer and Dolby are going after hardware devices using the Opus codec.
'The planned Opus program will focus on hardware devices and will not be directed towards open-source software, applications, services, or content.
'“We are pleased to act as administrator in the development of this important patent pool which aims to support efficient, fair, and transparent access to hundreds of patents covering technologies used in the Opus codec,” said Giustino de Sanctis, CEO of Vectis IP. “The reasonable royalty rates of this pool will balance the interests of both the program’s innovators and end-user device manufacturers whose products benefit from use of these patented technologies.”
'Vectis IP expects to launch the program in the coming months and for the program to include access to hundreds of patents with a royalty rate in the range of €15-20c per unit. Royalties for willing licensees will be applicable from the program launch, with further incentives available for licensees that promptly execute a pool license.
'Vectis IP welcomes other innovators with patents essential to Opus to submit applications for consideration in the program at opus@vectis.com.'
In other words: Yeah, we know that the IETF put out an IPR on Opus years ago. We don't care. We smell money.
> we know that the IETF put out an IPR on Opus years ago
As in, over a decade ago.
(Though they are covering more recent work in their patent pool, explicitly including stuff that they declare the planned program will not be directed toward.)
After so many years that idea should be laughed at at court. Patents are one thing but waiting for your "free" codec to gain usage across the world only then to reap the benefits should be called fraud.
It’s also just a sign of how broken the system is. These companies didn’t contribute anything to the creation of opus. The codec would have still existed since it was designed explicitly to avoid existing patents.
Unfortunately I don't know of any legal theory that would actually cause a judge to rule this as fraud. Unlike trademarks for instance, there's no requirement for you to actively defend a patent to have it be valid.
Estoppel. The development process was highly publicized and the relevant entities surely knew about the call for patent claims.
And possibly laches. Patents are highly dependent on the technological context of when they're written. The Opus development process deliberately sought to avoid existing patents, and trying to reverse these judgements a decade later is disingenuous.
Vultures coming to take claims to a codec which, I understand, was fairly carefully built to be actually open.
That it is only persecuting hardware inplementations gives me little relief. Low power hardware blocks should be readily available & usable, be a great help to making computers & sound devices better, with low power consumption.
This is, as much as anything, a huge blow against making bluetooth generally good & better; a great potential general use codec, encumbered! Alas! Make sure we cant use anything but locked-in overpriced proprietary codecs. Keep standards from actually developing healthily, fighting a 10+ year old codec from growing. Everyone associated with this patent pool should be blocked from the OIN patent pool immediately & with malice. Alas OIN is too old to have defenses against these here new crust of aggro lawyering stasists.
On one hand I believe the theoretical patent is very important to promote innovation. If you, the inventor, can get guarantees on your sales for 10 years (which is reasonable, imo) then it'll encourage people to make more stuff. I don't think patents are wrong.
However, the way patents are filed is wrong. There's so many issues with first to file. It's a good system when your office sees 100 patents a year. All of them are likely to be novel, and likely to be made by independent inventors. The problem we have now is there are entire companies (IBM, Google, RAND, TI, etc) that exist solely to farm patents. In fact, there are so many patent farms disguised as businesses I challenge you to sit, think of something, and search for a patent for it. I guarantee you one exists. Moreover, pursue the patent office if you never have. The definition of invention, especially in tech, is so vague it may as well not exist. Vomit some smart sounding stuff on paper with drawings and you've got yourself a patent that, if written correctly, will almost certainly produce a profit. Hell, take a cat and draw a stereo speaker on top of it. Call it "animal sound machine". I guarantee you if it hasn't already been done someone will do it. Better, draw a cat with a USB stick duct taped to it and call it "IP over Feline". Devise a stupid protocol. Then sue every idiot who thought it'd be funny to make one. This is seriously how lame modern patent law is. You might even be infringing on a patent with IP over Feline because someone already "invented" transportable storage.
I don't have a solution to the problem. Part of me wants to say one patent, per inventor, per term (10 years). You have the option of canceling a patent to put a new one in. But this seems like it's too strong. Maybe there's another way that legitimate inventors for an appropriate amount of time, and also disincentivizes patent trolls through some measure. Maybe we could start with "common good" patents (as determined by a board of smart people) being invalidated.
For the rest of the universe, if you read patents then you are liable for willful infringement instead of accidental infringement, the penalties for willful infringement are much higher. This is why you should never read through patents without a specific purpose in mind if you have any plans of filing your own.
A big problem is that you don't get a choice. It would be one thing if you could just take or leave the patented technique and benefit or not from it. If you could then the holder of the patent monopoly could only extract the marginal value of the patent. Unfortunately, patent holders frequently act to create hold up to extract far more than their share and prevent users of technology from being able to choose to not do business with them (at least without severe collateral damage).
You might find this now 11 year old writeup informative:
But even that assumes that there is some actual valid patent, in this case no patents are identified on the website, they're simply "calling" for patents.
On the other side, inventors are going to invent anyway for the sake of it and to have something to sell. They could be outspent and out marketed by established companies and end up working for them. Similar dynamic to startup entrepreneurs which get acquihired by an incumbent.
The point of the patent was less about ensuring people invent, and more about ensuring people invent in the open so other people can (eventually) build on it. The thought goes, that without patents, people will be super secretive about their inventions and the general public will never find out how they work, and thus nobody would ever be able to improve on other people's inventions.
And now ironically, companies that keep their inventions secret use the patent system to hinder open development of protocols, file formats, and software.
> A patent is a type of intellectual property that gives its owner the legal right to exclude others from making, using, or selling an invention for a limited period of time in exchange for publishing an enabling disclosure of the invention.
Patents have always been about "promoting the progress of science and useful arts" by promoting disclosure of inventions. The two are intricately linked. More disclosure = more progress.
I’m skeptical too but what OP describes would “promote progress” (by promoting openness vs secrecy). I agree that I’ve always had it explained as “promote progress [by incentivizing creation] but it doesn’t technically specify what he’s implying or what we are.
> On the other side, inventors are going to invent anyway for the sake of it and to have something to sell.
Are they? Inventors need to eat and sleep, and that's hard to do if someone else with a higher marketing budget and fewer qualms about lying to make a sale rips off your invention a day after launch day.
For example the situation between AWS and many open source server-side tools. Recently several companies based around open sourced software with auxiliary monetization plans have chosen to abandon open source in favor of more restrictive licenses that prevent cloud providers from re-branding their software and selling it as their own.
> They could be outspent and out marketed by established companies and end up working for them.
Or they could decide inventing isn't worth the hassle and they'd rather be a Dentist or something that pays well.
If I as a small company actually decided to sue a large company, I think that the response would be "hang on, here are 20 of our patents you are infringing on - you actually owe us money". Larger companies treat patents like trading cards, sadly.
Probably do, at that. However, patents have been known to stifle innovation at critical junctures in history. Patent pools seem to have been one way to patch that over wrt "common good" kinds of patents.(though at the time it was too little too late)
>The problem we have now is there are entire companies (IBM, Google, RAND, TI, etc) that exist solely to farm patents.
Google exists "solely" to farm patents? I'd think they exist to sell advertising, but that's just my uninformed opinion as a layman. Am I missing something here?
What's the point of this wall of text ? It discusses nothing about the article. The article is about opus and whether this firm has a valid basis to claiming patents that are used by opus but were presumably not known to opus.
Even better, "none to independently derive". Patents must be non-obvious. I would take independent invention as evidence that the patent should never have been granted in the first place.
They must be non-obvious to an expert in the field.
In theory that's a strictly tighter constraint than non-obvious, although it's impossible to actually implement.
However, independent invention isn't necessarily a proof that the particular patent should not have been granted.
Often, it is not the solution but its problem space that's obvious. There tends to be a finite domain of reasonable solutions, and solutions converge to that domain.
In which case you probably want some sort of thicket busting legislation.
Are you saying that anyone who independently derived the mechanism — before the first machine was made public or the patent diagrams were published — should also be granted the patent?
Once your rival has manufactured and sold their sausage making machine you can take it apart to see how it works. Perhaps you just claim you’ve never done this?
I said things will still end up in court - and the end-product in the example that you're giving will be too similar to the patented invention to stand up in court.
> Are you saying that anyone who independently derived the mechanism — before the first machine was made public or the patent diagrams were published — should also be granted the patent?
You seem incredulous that a "first to invent" system could work, but US moved away from that system less than a decade ago.
Edit: Never mind the first post was about independent derivations. That's not the same thing. My mistake.
As a secondary point to this, is AV1 also at risk of this kind of rent-seeking behaviour? It's also claimed to be patent-and-royalty-free, as does Opus. These companies who step on tech innovation to collect their pound of flesh are disgusting to me.
Do we know how successful their claim is. I've not followed that too much.
It makes sense to go after hardware vebdors, because that is where the money is and shaking down open source software users is a lot of work and bad press.
I've yet to hear of them suing Intel, who recently shipped hardware AV1 encoding in their GPUs. So either Intel is paying up or the pool doesn't want to poke the bear.
There are known patents on Opus held by Xiph.org, Broadcom and Microsoft, all made freely available. But as per https://opus-codec.org/license/ if you try patent litigation against anyone, your license to use any of these free patents is revoked.
So there's not many ways this pool can work. A company can only join it if they want to give up on ever using Opus themselves.
> It's a really crazy move from Dolby since they have at least a few products that use Opus
Seems like the strategy here is going to be to use FUD to leech small amounts of money from individual hardware device sales from the companies that can afford to pay it. It's just rent seeking. So long as they never actually sue anyone for their use of Opus, it seems like they'll be in the clear as far as the Opus license goes.
Edit: wouldn't it be pretty easy to call this bluff? you don't even have to try to win the court case, you just have to get them to sue you and immediately lose their license, or to admit that they can't sue you because they don't want to lose their own license, and the whole thing falls apart.
A more cynical view is that this is not about the opus licensing fees at all but rather to discourage use of opus in favor of Dolby's own alternatives.
>This is probably behind Oracle's move to remove Opus from Virtualbox recordings and revert to Vorbis, as of v.7.
Do you mean Oracle is being defensive and doesn't want to be sued ? Or that Oracle is going to join this patent pool and go after others, and hence won't be able to use Opus themselves without being counter-sued ?
Okay, maybe I'm missing something here, but how? Per wikipedia (1), opus is licenesed under bsd license. More relevant, on their own site(2) they tout opus as "a totally open, royalty-free" codec.
My question is, how can someone just step in, say "this mine, give money" and the world just goes with it? Did the opus developers (inadvertendly) use some patented tech? Do hardware makers use it? Or do Dolby and Fraunhofer have patents so broad that they cover anything encoding audio?
> My question is, how can someone just step in, say "this mine, give
money" and the world just goes with it?
It's called violence. Might is right [1] means the world goes along
out of fear, not agreement. Violence need not appear physical, but is
always backed up by physical violence in the end. There's no irony
that so-called "intellectual property" is backed up by thugs with
guns. Whatever noble aims patents began with they are now nothing more
than a form of violent "political language" against the intellectual
classes, designed, as Orwell said, to make lies sound truthful and
murder respectable, and to give an appearance of solidity to pure wind
[2].
civil litigation is not violence just because the police exist. This is a terrible take. Stop making non violence into 'violence' just because its something you don't like.
Hardware codecs are just generic applications of VHDL and integrated circuits. Making integrated circuits work is a legitimate challenge, but that's an integrated circuits patent, not a codec patent.
The opus software implements several ideas covered by patents. The software is open source, but using it and selling it is potentially encumbered in patents.
In this case, they seem happy enough for people to keep using it under the open source license (I'm not sure the patent pool has any direct power over the software), and seemed more focused on where the money is: people building this into hardware.
I'm with you though. They shouldn't have described it as royalty-free.
All known software patents that cover Opus are licensed under royalty-free terms.
However, Fraunhofer and Dolby are apparently thinking that they are holding patents that are relevant for Opus. It will be interesting to see what these patents are.
Wikipedia points to the 3-clause BSD license. Looking up the text of that, there's no mention of patents.
Software licenses primarily deal with copyright -- the ability to control how a concrete representation of an idea can be copied or performed. Patents deal in the abstract idea itself.
This is an attempt to prevent hardware for multi channel opus, for use in film/TV, and eventually inhibit the open source CODECs. Because only Dolby is allowed to have hardware for this.
Maybe we can argue that since there's a war going on and all that, we can no longer tolerate this sort of leeching. Attack it as unpatriotic. I mean, it is. It's the least way it's disgusting, but it's still perfectly true.
I'd love to see a well written defence of their position. Frauenhoffer and Dolby both employ smart people. If they put smarts out into the world, which people leveraged to do smart++ then there's a component of "ok, so pay the tax" on this but given the IETF side, even with IETF IPR statements, this sounds like they misunderstood the RAND goal.
or is EUR 20c per unit for hardware 'reasonable' -on Apple volumes, thats a lot of money.
It’s not a hardware feature and it’s not essential to Apple’s business. If the trolls insist that it is, Apple can and should nuke the extremely limited software support they have for it.
>or is EUR 20c per unit for hardware 'reasonable' -on Apple volumes, thats a lot of money.
Apple thought they should only paid $1.5 / unit to Qualcomm for all of Qualcomm's patents. 20c per unit is a lot for single type of patents with limited coverage. Their "reasonable" range is probably 1c per unit.
Opus is a decade old, any patents would have to be filed before then or Opus would be prior work, and they're issued only for 20 years, right? At least for the software, which they claim not to be going after (but who'd trust a software patent lawyer). Any software implementations such as in VirtualBox, who just switched back to Vorbis from Opus, would be perfectly safe in at most ten years, perhaps already today. Am I understanding this right, at least for the software part?
Most of Opus is quite a bit older than that too a decade is for the final standard but the project began in earnest in 2007 and many components came from earlier work.
It's best understood through the lens of FUD against both a specific competitor codec and a competing standards methodology.
Fraunhofer and Dolby make money from patent royalties on standards. If standards no longer want patents or royalties, then they are out of business.
It's obviously going that way long term, but in the meantime a bit of "Linux is cancer/communism" type FUD will more than pay for itself in extending their patent royalty collecting time.
This is the correct way to understand it. Dolby sells a bunch of codecs, and Fraunhofer sells USAC (which is basically Opus with some patented stuff added on).
Their problem is that Opus serves many of those use cases for free. They can't allow that to happen, so scaring potential customers away is good for business.
If Opus was a kind of tree, a maple, that's got very popular in the country, and a gang of sketchy "enterpreneurs" came to a judge and claimed that all these maples in the country belong to them now, because they've come to possession of a piece of toilet paper with a word "maple" on it, would the judge order to cut down all the maples?
Isn't Opus used by WebRTC, i.e. this would break the many video conferencing systems which have made WFH and online conferences possible for the last couple of years? That's a rather large class of customers to troll.
The Case Against Patents:
https://research.stlouisfed.org/wp/more/2012-035