Ogg Vorbis Update: Thomson Trouble

3.1415926535 writes: "In this article on C|Net, Thompson Multimedia's vice president of new business Henri Linde openly threatens the Vorbis project. The quote is, 'We doubt very much that they are not using Fraunhofer and Thomson intellectual property. We think it is likely they are infringing.'" Considering Ogg Vorbis is GPL, you'd think they'd already know.

We need a fight

I had mentioned this in a reply to another post, but I want to start a new thread on this one... Perhaps it's time to make a stand here. What do we have here? IMHO, we have a corporation (two, actually) that's apparently decided to play hardball, never mind whether or not they have a legal leg to stand on. Notice that this "issue" didn't come up until CMGI withdrew its support for Ogg Vorbis, probably because CMGI could have coughed up the cash to mount a defense, whereas the developers probably can't. So Thomson is going to try to steamroll over a potential competitor. Are there legitimate patents on audio compression technology that Ogg Vorbis is violating? Possibly, but I don't think that's Thomson's intent. They're out to crush a competitor that they see as vulnerable now and a potential threat later. Even if the applicable patents could cover this area, are they too broad? Once again, we don't really know, but we aren't going to learn anything if Ogg Vorbis is forced to roll over and die. IMHO, we need a very public, very nasty, very enlightening fight here, if Thomson is willing to oblige. Why? Because we'd better define how patents and intellectual property are going to be applied in the coming decades. I'm not saying who's right or wrong, but we'd better dtart defining some boundaries whiel there's something left to defend. Get the debate off of Slashdot and out into the mainstream. There are lots of ways to do that, and I'll leave it to others to define exactly how, but we need to make more people begin to think about what role patents and IP are playing in new technologies, such as computers and the Internet. I will pose this question: If the current interpretation of patents was in use about 100 years ago, would one company have been able to patent automobiles? If it was in effect 50,000 years ago, would someone have been able to patent fire? What about the wheel? I can see it now. "A device for the smooth transport of passangers and cargo across the surface of the earth by means of rounded edges and continuous motion almong the ground."

Fraunhofer

Fraunhofer owns patents on damn near everything
sound related. Everything they do is cross patented. It's a research lab. This is how they
make money. They're extremely good at it. If .ogg hits them in the pocket, expect litigation.

Here's a list of their patents pertaining to MP3.
http://mp3licensing.com/patents.html [mp3licensing.com]

Re:So naive.

Another factor is that a suit by Thompson will hardly appear as an act of good faith to a judge. If Thompson can download and examine the source code, they can hardly claim ignorance when they lose the suit.

50 cents? HAHA

Try more like 2-6.00 US dollars a copy of any shipping software product.

Plus the 10k US yearly minimum.

It's quite a racket they have going with mp3. They did a good job pulling the wool over most people's eyes.

Re:Ominous.

In this case, if you abandon your principles, survival becomes irrelevant. What does survival in this context mean?

Re:How can MP3 be stopped?

You forget the law that will be passed three years from now and where the traffic patroller will have the right to ask for your driver's licence, your seat belt and your car player...

Highway UCITA...

Beowulf cluster of Fraun. execs?

Y'know, I've always wanted to play those clock-speed dependent 286 games again.

But seriously, Ogg's been Ogged [tuxedo.org]. Frauenhofer is making a kamikaze attack without regard to future repercussions. The irony is wonderful.

I forsee the rattling will continue. The Ogg Vorbis format will exit beta and enter into the Internet's various mirroring services and freenet-style anti-censorship services, the company Xiph will get sued out of existance, the CODEC will survive, plugins for Xamp and Winamp will abound, business as usual will continue. Anyone remember why we should be using PNGs instead of GIFs?? right. do you? same deal.

Re:So, what's next???

Nice try, but I think you're completely off base on several things on this list.

For example, Apple has completely re-written Display Postscript and created Quartz to be Adobe-free (to avoid paying licencing fees) for Mac OS X. No patent infringement there. Furthermore, there have been third-party GPL'd Postscript interpreters for years; maybe a decade at this point. The other file format examples that you provide are equally impossible.

Sun doesn't have anything to do with JavaScript; you'd think that /.ers could figure that out after 5 years. Nothing in Java is patentable (it's a language and a spec for a class library), so submarine patents are unlikely. There are multiple sources for JVMs (both Sun and IBM make JVMs for Win32 and Linux), so if Sun starts to charge, people will stop using it.

HP charging for it's printer drivers (apart from the cost of the printer) is crazy. What would you do with a printer without a printer driver?

/. charging is, I would bet, quite likely within 5 years. As VC money dries up, companies are going to need to find some other way to pay for providing content (and such).

-jon

Question: the memes of Ogg!

As the project itself is not under the umbrella of any corporation, who exactly is being sued? The visible maintainers/authors? Who says it can't just be picked up by anyone else?

Can't anyone pick up where it left off? How can you charge a piece of information with a crime?

Is this not a way in which OSS can almost circumvent the system by simply not being part of it? The software will exists as it's own entity, and simply be serviced by whoever wants to work with it.

Re:It's very likely...

Did Microsoft license Fraunhofer / Thompson's patents when they created their Windows Media (.wma) format?

If so, Microsoft would have another reason to be happy to see .ogg files disappear.

But if not, that's proof that it is possible to build a decent encoder without the patents. At least if you have barns full of money and don't need to worry about nasty legal threats.


Torrey Hoffman (Azog)

It's very likely...

It's actually quite likely that Vorbis infringes on several of Fraunhoffer/Thompson's patents (now, whether you agree with the patents or not is a seperate issue).

The fact is that Fraunhoffer has pattent MANY different technologies used in perceptual encoding. Basically everything in MP3 and AAC is covered by strong patents. They even have many patents on other perceptual coding techniques not used in those formats.

One of the biggest impediments to commercial development of competitors to MP3 has been the Fraunhoffer patent collection - which makes it difficult to do any type of perceptual encoding without infringment. Pretty much the only other companies that can get away with it are people like Lucent (w/ PAC/ePAC) that also have their share of perceptual audio coding licenses that they can cross-license w/ Fraunhoffer so they don't sue each other into oblivion.

The chances that the Vorbis guys have discovered some completely revolutionary method of encoding that doesn't infringe on any of these patents is unfortunately very slim.

Of course, I wouldn't expect to see Thompson do anything about this until it becomes a real threat. No reason to waste money on lawyers otherwise.

Re:isn't there a law?

>Isn't there a law against harassing suits? If not, there should be...

I think we should harass suits more often.

Re:FUD not against OV Team, but against others

An astute observation.

Actually we are trying to set things up such that xiph.org is the only easy target in any litigation that might come Vorbis's way. The idea, obviously, is to mimimize potential liability of any industry adoptees--- not because it's strictly necessary, but because the industry is more likely to adopt if we can give them additional armor-plated warm fuzzies.

Monty

Re:The horse's mouth: Time to calm down folks

Oh, erm, actually I was only talking to the folks that seemed ready to flee if MPEG said "boo". Sort of a "stop running, good, now breathe deeply, yes, good" in calm soothing tones.

Now if you're pissed and feel like getting a few hundred developers to march on Thomson headquarters with flaming torches (maybe we could get the Large Hot Pipe Organ to play) to make our opinions known (loudly, spectacularly, but peacefully), I'll volunteer right now for propane torch duty so long as someone volunteers crash space for the sleepover!

Monty
.

Re:How can MP3 be stopped?

As an individual, sure, you have relatively little to worry about. Except that all the really *fun* artists and companies will spend more time worrying about the bottom line than getting music to you to enjoy.

Oh and don't forget about all the restrictions on when you can use your music. SDMI may yet require you to listen on only a single 'approved' player. Buy a copy for your desktop, another for your portable, another for the backup disc... SDMI and FhG are on friendly terms.

We're making Vorbis for everybody, and that includes all the companies and artists, big and small. The small fry (who can't afford mp3) have the most to gain. The big fry are happy too because this whacks a chunk of money right out of expenses. It levels the technological playing field and makes it easier for everyone to make money at the same time it keeps the technology free (and interoperable: easy to use)

Mmmm. Makes me just want to go lie down in a peaceful, sunny field of flowers (and code, of course).

Monty

Re:So, what's next???

Here are some that might be on the horizon, although some or all may be a stretch, and some may already be happening. What do you all think?

I agree, some are a stretch. But some are exactly what I'm talking about:

• Adobe and .ps/.pdf
• Microsoft and .doc

Although I use postscript and pdf all the time, I worry a little. I realize that they've been so useful to me because Adobe has published the formats and allowed implementations to thrive. They could pull a Unisys, though.

And MSFT's .doc format? Just Say No.

Whenever you depend upon something that you use by permission, not by right, you are creating a dependency that may cost you in the future.

Steve

In the UK this is called Libel

Thompson is libeling the authors of Vorbis. Could the EFF be talked into taking them to court, rather than waiting for Thompson to make a claim of patent infringement? Attack is the best form of defence.

Making groundless statements in order to sabotage a company's prospects is a serious crime, at least in the UK, and the fines are unlimited. Since the code is open the CEO or whoever it was has no defense: he should have known whether the patents were infringed or not before opening his mouth.

TWW

Re:Patents, the joy, the pain...

IANAL, but I believe that one kind of patent (the statutary patent) already allows this. The point of such patents is to simply say, "I discovered this." You can't charge license fees, but neither can anybody else, and you go on record as the inventor.

Of course, a better idea would be to patent something, and license it under a variant of the GPL for patents, eg "You can use this to produced a derived work, and patent that, but you must license that patent under the PGPL".

I'm unsure of whether existing patent law would allow this kind of thing, though.

Not quite

This will be a real-world test of GPL and the power of the internet.

You are right that the OV people have no resources, money, or lawyers, but the OV people aren't critical either. OV can live and thrive *without* the OV people.

So what can happen? Thompson stifles the top 10 developers? 100? 1000? Will they target everyone who's downloaded the code?

Lets say the top 10 developers are sued to, essentially death. That doesn't mean they lose; OV could be found non-infringing. At which point *any* developer could pick up the pieces and continue.

If it is found infringing, well, that's all folks. The code was infringing... Fix, and release again, I guess.

A commercial big daddy will help it nothing in proving the code is not infringing, I don't think. It can only provide resources. In the end, I hope OV survives, and that we have a better solution, that the GPL reigns powerful, and Thomson gets egg on their face.

Geek dating! [bunnyhop.com]

Ominous.

Considering Ogg Vorbis is GPL, you'd think they'd already know.

They very probably do. Of course, that's not to say that Ogg V does use mp3 technology, but it is certainly possible that it has been influenced by it.

What we are looking at now is a commercial company deciding to go after a defenseless GPL project. How will Ogg V survive? They don't have any resources or money or lawyers, do they? In the real world, that's what it takes to survive, and Thompson's know it.

Perhaps it would be good for Ogg V to get a commercial 'Big Daddy' that will defend it under the GPL.

But I fear that would be impossible. They would only defend it if they owned it, of course, or if they had some power over it.

If Ogg V is reliscensed under a more commercially friendly license, such as MozPL, It may survive.

What do you want more? Your principles or Survival? That is what it comes down to, I fear.

KTB:Lover, Poet, Artiste, Aesthete, Programmer.

The positive side

There is a positive. At this point, while the Open Source community is quite supportive of this technology, the majority of the world is still fixated on MP3.
Likely, the lawsuit will be dismissed, or at least won by Ogg Vorbis, but the damage to MP3 will be that Ogg Vorbis will suddenly be well known to people who aren't neccessarily going to hear about it in the community.
It's free advertising via a nuisance lawsuit... sounds like a case of "Any publicity is good publicity to me." (Excuse the cliche)

Ogg goes nowhere without hardware.

I can walk over to my local Circuit Shitty today and buy a $199 Philips [expanium.com] or $150 D-Link [dlink.com] portable CD/MP3-CDR(W) player (there's even an off-brand "Classic" MP3CDRW player for $99), a $299 Aiwa [aiwa.com] car unit, a $299 Philips mini-system (sorry, no link. I think the model is FWM55M37), and an Apex [nerd-out.com] or Aiwa (model XD-DV370, I think) or Raite or similar DVD/VCD/SVCD/MP3 player for ~$200, and have MP3 capability with media compatibility across all typical listening environments for under $1000.

Ogg can't chain me to my computer or even to a PDA and expect to thrive. Ogg should spend some time bringing their codec to the typical embedded A/V processors found in the new generation of cheap OEM DVD and CD chipsets for consumer electronics (like the ESS VideoDrive 4308 and 4318, found in most of the DVD/VCD/MP3 combo players)

-Isaac

Re:So naive.

I think the point was not "being GPL will give it some kind of special legal protection" but "if they were competent, and they thought it was infringing, they'd just download the source and figure it out themselves." Clearly, being GPL doesn't mean much- the license itself has not even been significantly tested in court. But it does lend a layer of transparency to the project that makes these types of threats both more irritating in their bluster and arrogance and more pathetic in their ignorance.
~luge

So naive.

The fact the Ogg Vorbis is GPL will only be relevant once the case begins, if it begins at all. The purpose of any litigation has nothing to do with truth or fairness, and everything to do with protecting Thompson's business. As the article says, the threat of litigation is enough to scare would-be hardware and software and content providers away. The idea is that Thompson has a big enough legal war chest to win even if they lose.

"This is what hardball is like" is what their representative said. Essentially, it's plutocracy as normal.

How to read a patent.

When looking for patent infringement you look at the Claims rather than the specific Embodiment. The idea is that the claims stake out the intellectual territory covered by the patent and the Embodiment tells you how to build a specific example. You aren't interested in the Embodiment in the patent, you are interested in some other Embodiment (in this case Ogg/Vorbis) and whether it sits on territory staked out by the Claims.

The Claims are a series of mostly concentric circles. The outer ones are going to be very broad and will almost always have prior art against them. Inner ones will be progressively more specific. So the first thing is to look at the published state of the art before the patents were applied for and decide which of the Claims are actually real. Then you can look at Ogg/Vorbis and see if any of those Claims cover their work.

Paul.

Someone needs to stick to the name brand crack

Troll. Your mama would spank you.

I was never a VQF developer. One of my professors in Japan, Sadoaki Furui, is head of NTT Human Interfaces lab, which developed the original VQF, then called TwinVQ. The fact that we knew each other was the limit of the overlap (and we met exactly because we both worked in compression).

Any technical commonality (there is some) is coincidental. Both TwinVQ and Vorbis draw more from speech encoding technology than mp3 does.

(Incidentally, VQF has noise problems because of the way the lossy/nonuniform vector quantization interleaves MDCT scalars. Vorbis works differently).

Vorbis was never a 'limited hack of VQF'. Please, *do* go inspect the original CVS snapshots as well as the previous generations of Xiph.org codecs, Squish, '95' and Stormbringer, all of which predate TwinVQ.

But, eh, I just fell for arguing eith a fool. Now I feel all dirty.

Monty

Re:Ogg goes nowhere without hardware.

It's happening. Don't forget it took WMA a year and a half to make it into its first handheld. We're already running under Dadio, the iObjects [iobjects.com] handheld operating system software licensed by handheld manufacturers based on ARM (eg, Cirrus Maverick). And we haven't even had an official 1.0 release.

My Iomega HipZip plays .oggs, but I'm not allowed to give out that firmware version yet. For those of you who have the HipZip, had you wondered why the xiph.org twirlfish logo was in the 'about player' menu? :-)

Monty

Re:Question: the memes of Ogg!

Ogg is and always has been under Xiphophorus, a registered S-corporation in Somerville, Massachusetts. Xiphophorus holds all the copyrights.

Xiph.org would be the target, not any contributors.

Monty

Theory behind patents

Actually, the thoery is that if there were no patents, companies would keep everything as trade secrets. By giving them legal protection if they explain what they've done, everyone can build on their innovation, at least once their patent runs out. Also, patents help companies recoup R&D costs, thereby encouraging them to do more research

Of course, since patents now last 17 years, and are awarded for things that could never have been protected by being trade secrets (One-click, anyone?), or for that matter even required R&D, the situation is ridiculous.

So, what's next???

We should see these things coming a mile away? Here are some that might be on the horizon, although some or all may be a stretch, and some may already be happening. What do you all think? Are there any others?

• Adobe and .ps/.pdf formats
• Microsoft and their .doc format
• Sun with Java(script)
• Napster
• Bluetooth
• The zero-emmission vehicles that run on that special juice
• Sony et al. with Hi-quality CDs
• Free isps like Juno
• HP with printer drivers (now their free, but once market dominance is assured...you pay)
• hotmail et al.
• /. (wouldn't you pay $1.00/month for a access/posting privileges/a +2 bonus/10 karma points/etc.)
• The applix anywhere java office suite
• Free on-line disk space/storage companies

I'm not totally happy with these items--none seem as obvious as the mp3/.gif examples. Oh well, it goes to show that there is an important difference between free and Free.

patent system encourages innovation?

So let me get this straight. We have patents so people won't use existing technology, and instead will have incentive to create new solutions to problems. Now patents are used to make sure new technologies DON'T appear. When did this huge disconnect come about?

Missing the Important Bits (Again)

You have to wonder if either the submitter or the editors actually read this article.

Sure, Thompson Multimedia is doing some entirely predictable sabre-rattling. Anybody who didn't see that coming a mile off should see an optometrist.

The important thing mentioned in there is that CMGI has pulled the plug on Vorbis development. That's a far more important, and ominous development. Corporate backing was allowing progress to be made very quickly on Vorbis, and will be critical when the inevitable patent infringement suit comes. If someone else with deep pockets doesn't step in soon, we can just resign ourselves to paying Frauenhoffer's license fees for the forseeable future.

(And please, let's not delude ourselves that the mythical Open Source Community will magically step in and finish the project: enthusiasm and spirit are no replacement for in-depth knowledge of signal processing.)

It's not very likely at all.

> It's actually quite likely that Vorbis infringes on several of Fraunhoffer/Thompson's patents

Actually, no. Monty, Xiphophorus, and iCast did their homework, and also hired some clueful technology IP lawyers to look things over, some of whom I've met -- these guys are sharp, and they grok psychoacoustics.

Have you read the actual patents? If you work out exactly what is being claimed, FhG doesn't actually own the farm, as I've heard it told.

Yeah, I was skeptical too, but I've been convinced.

Note, also, this: FhG has never actually claimed any infringement by Vorbis. FhG's lawyers could certainly read the source if they wanted to know.

Re:Sheesh

Or maybe they have a patent on "A method of compressing audio to preserve quality but reduce space usage.

Well according to this patent [delphion.com], obtained by Thompson for his "invention", that may be exactly what they claim. The patent would seem to cover any audio compression method that converts from time domain to frequency domain, does quantization, then entropy coding.

The other Fraunhofer patent [delphion.com] is at least a bit more focused, and specifies a breakdown into frequency groups, followed by quantization, then compression. The Ogg Vorbis [xiph.org] scheme avoids the first stage of prefiltering into smaller frequency bands, and does the transform in one feel swoop. This requires more work for the transform, but arguably gives better results.

In short, the first patent I mentioned seems difficult to defend against, unless it can be shown to be overly broad or invalid. The second is exactly what Ogg Vorbis was avoiding.

Info about the Patents

From this url [mp3licensing.com] given elsewhere by another poster, I looked up all the patents that Thomson Multimedia and Fraunhofer have in the US (apparently some weren't approved in US but in other countries). With all the hub-ub about overbroad/silly patents I thought I could go read some in more detail. The list of patten numbers is:

• 5,742,735
• 5,455,833
• 5,579,430
• 5,559,834
• 5,703,999
• 5,706,309
• 5,736,943
• 5,701,346
• 4,942,607
• 5,214,742
• 5,227,990
• 5,384,811
• 5,321,729
• 4,821,260

(You can look up any of them at the patent office [164.195.100.11]. Just enter all the numbers in the search field separated by spaces.)

Some interesting things I noted:

1. I know something about math (and signal processing) and I couldn't decipher what was meant by a lot of them. In other words, how can the patent office people decide if an application is a valid patent if the application is incoherent? I know that specialists are supposed to look at applications for patents in their field but they might not be up to speed about everything in their field. Yeah, they might reject it for that reason but aren't they just as likely to approve it, thinking "maybe I'm just stupid/ignorant about this sub-field."
2. They seem to be fairly specific (although specificity doesn't necessarily mean something is patentable).
3. These "ominous words" were found at the end of patent #5,579,430, titled "Digital encoding process":
   • Although the invention has been described and illustrated in detail, it is to be clearly understood that the same is by way of illustration and example, and is not to be taken by way of limitation. The spirit and scope of the present invention are to be limited only by the terms of the appended claims.
   Since most of the patents I found did not specify that the encoded signal had to be audio, this seems like they have a patent on any use of whatever their algorithms are trying to do (which I found not very clear...) In other words, it is almost like somebody patented a specific hash table function (I'm sure someone has) and then patented it specifically for application X, but didn't rule out the possibility of "owning" it in any application that uses hash functions.

4. All of the above must be taken with at grain of salt because the legal-ese in the patents (especially the beginnings where the claims are listed) is very weird and I had trouble deciphering what kind of math they were getting at. Not to mention one could spend days if not weeks reading them all and all supplemental material. Overall it looks like Ogg would have to include some very specific algorithms to be infringing (unless just the fact that the patents claim to patent one method of doing a certain type/part of encoding signals is enough to claim infringement--i.e. one form of encoding algorithm counts as owning them all...but that doesn't seem very reasonable.)

   Rachael

The horse's mouth: Time to calm down folks

Linde said what he said only because it was an opportunity for some free FUD. Between the lines, it says, "we're worried/scared, they're on our radar, and we need to make some noise. Words are cheaper than lawyers."

Big surprise.

If Slashdotters didn't expect that already, well, shame on you. Sudden worried speculation about Fraunhofer's and Thomson's 'newly ominous tone' is just the snowball they'd like to start (while pressing full-steam ahead with the new webcast and download licensing). I'd be annoyed if they managed to start it with a single public sentence (we've known they didn't like us for quite a while in private). Let's not be a herd of sheep being maneuvered into the chute.

Thomson and FhG both have a reputation of a loud bark, but tend to pursue relatively little litigation in practice and they'll have to work hard to find basis against Xiphophorus. When we did our patent review, we focused on the FhG/Thomson MPEG patents. Our counsel advises us we don't infringe, what we knew already.

In other words, nothing's changed from yesterday except that Linde has decided to bluff before the call.

Monty

Eric Scheirer is on our side

Don't attack Eric for telling what he sees to be the truth. That's his job.

Behind the scenes, he's a friend of the Ogg project and has been for some time. He's doing his job by calling it how he sees it, and we don't ask him to spin the facts toward our favor. He also doesn't have control over which quote a reporter will choose.

Monty

We should catch this earlier next time

These delayed-action patent issues are becoming predictable. The community ought to keep its collective eyes out for this in the future. While not exactly the same, the similarities are striking:

1. gif/unisys/lzw Unisys waits until the file format is literally covering the world, and then threatens lawsuits.
2. Rambus We've been using SDRAM for how long, and they're just waking up to the idea that they have patents on it?
3. CDDB Don't worry, it's "free." Right. Sure.
4. Now, Fraunhofer ...

We should all be experienced enough with this phenomenon to see it coming a mile away. From this perspective, things like Windows Media are not competitors to mp3, they are just different complainants in the patent lawsuits.

I strenuously suggest people use png [libpng.org], .ogg, and anything other technology that isn't trying to strangle open standards.

The Internet wouldn't have existed if they played by these rules at the beginning.

Steve