Slashdot Mirror
31 Lawsuits Filed Over Alleged JPEG Patent
dcrouch writes "Compression Labs has initiated a lawsuit in the Eastern District of Texas against 31 major companies for infringement of its 4,698,672 patent. The patent, filed in 1986, includes 46 claims for various embodiments of digital signal compression technology and reportedly covers JPEG compression. From the dates on the face of the patent, it appears that it will expire in October 2004. This looming date may have prompted the suit. Compression Labs will certainly have a fight on its hands. A major question will be why the patentee waited so long to stake its claim. The Eastern District of Texas court has established special patent rules that help speed the progression of litigation."
Looks like we should have seen this coming a long time ago..._ not_free_patent
Published Thursday 18th July 2002 17:12 GMT
http://www.theregister.co.uk/2002/07/18/jpegs_are
oh and here are some more related articles
Some Google results
_JS
Not really. Compress a 640x480 photo image to ~50KB with PNG ;-) [and still look decent].
PNG is meant for *lossless* raster art.
JPEG is meant for *lossy* "photo realistic" art.
TOM is meant for pepsi.... need more pepsi....
Someday, I'll have a real sig.
Well, PNG is lossless, whereas JPEG is lossy. JPEG is superior when it comes to compression of e.g. photographies.
It's not a submarine patent, it was granted and published back in 1987.
Well, this isn't exactly an issue of a submarine patent. The implementation has been available for many many years, and they're only now trying to leverage their patent on it, as opposed to a submarine patent which is designed around technology that doesn't exist yet, and once someone actually invents it, is used to leverage patents from them. So, close, but not quite.
[insert witty comment here]
Yup, while PNG can do 24-bit with no problem, it's lossless compression, which doesn't perform particularly well with photos, and it just can't compete with jpeg in terms of file size.
I was always under the impression that you are supposed to defend violations of your patent or it goes into the public domain? If this is the case then I don't see this as a long court battle.
*best* .. well, that depends on what you're measuring. It's going to give you a smaller file size. A PNG will still more accurately recreate the original. You can convert a BMP into a PNG back and fourth a million times and you won't loose quality unless there is some form of error. Except for the lossless version of the JPEG2000 standard, you loose information every time you compress a JPEG. compressing back and fourth between jpeg and bmp quickly makes something that is unusuable. If space isn't an issue, you would want a lossless format. I would think you generally wouldn't want to convert it into a lossy format until the very last step- distribution, wherein effeciency of communication is more important than perfection of information.
PNG and JPEG are for very different purposes.
PNG is a lossless compression format, aiming at graphical images.
JPEG is a lossy compression format, aiming at photos.
It's like shortening by abbreviations vs. shortening by digesting. With abbreviations, you can restore the exact original, but your compression ratio is limited. With digesting you get much better compression while still getting the important facts, assuming the digester knows enough about the subject of the text he writes a digest on.
The Tao of math: The numbers you can count are not the real numbers.
That is trademarks
Patents do not require that you defend them.
The goal is to exploit a loop hole in patent law where validity of patent is counted from the date when it was granted, and not from date where it was applied for. So, if you've got a patent application, and you know that no competitor is even close to being able to commercially exploit it, you (the application) just slow down the approval process as best as you can (by filing papers as late as possible, by introducing trivial amendments which forces the patent office to restart the procedure from the beginning, etc.). You keep on stalling like this until you see that a competitor is almost ready to infringe: you then let the application proceed at normal speed, and enjoy 25 years of monopoly from that date on.
Say no to software patents.
Do you keep wav files on your hard drive because they are loss-less?
After a certain point, compression isn't hindering the quality that much, and the amount of space that is saved is more important than the small amounts of the quality that are lost. Just like the mp3 format works so well by slimming down the things the human ear can't hear, jpg works by blending and compressing images in way that the human eye can barely tell the difference while saving optimum amounts of space.
If you've known you had a case but did not bring suit and allowed your damages to mount, your suit is barred by the doctrine of laches. They would have a good argument based on this to either dismiss the suit or severely limit their damage award, assuming the validity of the patent, which I'm not.
One line blog. I hear that they're called Twitters now.
They have been pursuing licensing agreements since at least 2002.
-- "It was as if the paint factories had decided to deal direct with the art galleries." - Thursday Next
Nope, just the ones covered by their patents that can be upheld in a court.
In fact there is a version of MNG (muti-image PNG)which uses JPEG and is called JNG (JPEG Network Graphics).
This offers the advantages of the PNG/MNG file-format specifcations (transparency, meta-data, &c) along with the JPEG compression algorithm, and is meant to be a replacement for JFIF (the JPEG file image format) which is the commonest JPEG-based file format.
BTW, this story sounds *very* BAAD...
[woops...formatted now]
Joe Llywelyn Griffith Blakesley
[This post is in the public domain (copyright-free) unless otherwise stated]
With the rate bandwidth throughput is increasing, why not make png the standard? Then we can have lossless copies of our images.
When baud was 300, the difference between a JPG and BMP/PNG type was HUGE. Now the compression gains from the jpg standard is decreasing as bandwidth increases (i.e. the 0.01 seconds saved for downloading a jpg compared to a png is negligible.)
Go go png format!
... can turn up some more information about this
This isn't new. It doesn't look like Microsoft was ever in talks with these people. I'd guess they didn't sue them because they didn't want to get their butt handed to them by a company that gives out 1.9 billion like candy.
Yes, that's why I've put "legal definition" between quotes. AFAIK, the practice has been outlawed since a while (or has the law simply been rewritten such that the expiration timer starts on date applied rather than date granted?), and since then the media have started reusing the term "submarine patents" for the practice of not enforcing granted patents until your competitor is big enough that it hurts.
Say no to software patents.
...is if you could design a method to arbitrarily reduce quality before PNG compression, in such a way that it becomes easily compressible. Kinda like the oh-so-much rumored wavelets that I've yet to see materialize.
I'm not sure how exactly PNG works, but I assume it has some "easy" and "hard" pixels to encode. If you accepted some loss to quality, for the bonus of getting lots of extra easy pixels, could you get near-JPG performance out of it?
Obviously, this would be a rather major job to do and I have no idea if the PNG compression algorithm is capable of such a thing (obviously it can compress anything, but will some "leverage" significantly improve compressability?), but would someone with some clue tell me if it is?
Basicly, instead of
Original -> JPG: discard info & encode rest (lossy)
Original -> discard info (lossy) -> PNG: encode (lossless)
Live today, because you never know what tomorrow brings
Nope, that's incorrect. A patent, if granted, is valid from the date it is first applied for. This date is called the filing date (or priority date - it depends whether this is a "first filing") and is THE most important date in a patent's life. This is the date it arrives in the patent office.
:)
There are other important dates such as the date the application is published (usually around 18 months after the filing date) and various deadlines - especially for PCT (WIPO) applications but a patent is retroactively valid from that original date for the normal length of a patent.
This meane, for example, that if I filed an application on 1 Jan 2000, and it was granted tomorrow, I would have protection for 20 years from 1 Jan 2000 assuming I paid all my annual renewal fees.
An interesting aside here is that whilst the patent is pending I still have to pay teh renewal fees (i.e. in the case above I'd have to have paid my fees for 2000-2004 even though nothing had been granted and nothing was guaranteed to be granted)
Troc
Troc's dubious podcast and blog: http://www.trocnet.net
...to their website. here it is.
No, once the patent is issued, it's considered public, and it can't be hidden.
But if you delay the issuing of your patent... then you can do some very evil things...
http://c2.com/cgi/wiki?SubmarinePatent
After perusing the comp.compression faq for an overview of jpeg (see question [75]), I don't think this patent, even if valid, will affect JPEG as we know it. As the parent post points out, the patent covers compression via diff's between images. (This is, as I understand it, a major component of MPEG encoding, so I'm curious why they're not the ones getting sued.) JPEG, as your web browser uses it, likely doesn't use such a technique. From the FAQ:
The hierarchical mode represents an image at multiple resolutions. For example, one could provide 512x512, 1024x1024, and 2048x2048 versions of the image. The higher-resolution images are coded as differences from the next smaller image, and thus require many fewer bits than they would if stored independently. (However, the total number of bits will be greater than that needed to store just the highest-resolution frame in baseline form.) The individual frames in a hierarchical sequence can be coded progressively if desired. Hierarchical mode is not widely supported at present.
My take is that this "hierarchical mode" extension is the part covered by the patent. The problem is that, assuming the FAQ article is correct, most implementations of JPEG decoding we encounter won't support it, meaning most people aren't going to be encoding their JPEG's this way since it would be incompatible with mass-market JPEG implementations. It sounds more like something a company would use as part of a proprietary format built on JPEG.
Bottom line: don't expect this to have any ramifications for Mozilla, IE, etc.
I am a man of const int sorrows
Granted patents are PUBLISHED documents. This means they are public and available to be read by anyone. So the company with the patent isn't hiding their patent waiting to spring it on an unsuspecting world, the world is failing to do some simple book research before relying on some technology.
If I want to invent something or research something or whatever, the first thing I do (and this is standard practice in research) is do a literature search for bachground info - no use reinventing the wheel etc. Any decent research will quickly reveal both non-patent literature and patents (pending or granted) which revolve around the work I am researching. If I then create something that infringes a patent it is either:
a. my stupidity
b. that I don't care and want to fight the patent
c. cross licensing
d. gonna buy that company
etc etc
Troc
Troc's dubious podcast and blog: http://www.trocnet.net
A little clarification: PNG is actually a very nice substitute for JPEG's (especially the way JPEG's are currently being abused). It's single shortcoming in comparison to JPEG is that it does not compress to as small a size. JPEG is still a better choice for the web, for this reason, but PNG beats it hands down in other roles.
I've been working as a designer for over ten years (I started back when it used to require a degree, not just a computer). It's been my experience that JPEG is one of the most abused graphic file formats in general. It is good for the web...it's intended purpose...but it is awful for everything else. Unfortunately, everyone insists on using it for everything else...printing, digital cameras, stock art...all apparently blissfully unaware that this LOSSY algorithm is slowly but surely leeching the color data from their pictures every single time they save them. The result: Precious memories that print with muddy colors, photos with ugly artifacts in them, and unhappy designers who have to explain to their clients why there is no Photoshop cure for being a moron.
PNG's are great. They support multiple levels of alpha transparency, retain all their data, and compress even photos very well. They are a much better option for a multi-purpose format. (They would be even better if M$ would get off their collective asses and implement them properly in IE. Currently, IE treats them as if they have only one level of transparency)
JPEG's can still be smaller, at the expense of quality, but broadband may eventually make that moot. I fear we will all still be using JPEG's even then, though.
I've seen video games still using the PCX format, which is a crap format if ever their was one. Old formats die hard.
Do yourselves a favor: Use JPEG's on the web if you want, but archive your pictures with another format, like PNG, TIFF or Photoshop (PSD). CMYK images need to be saved as TIFF's or PSD's...ironically, the proprietary PSD is probably more universal.
What are you smoking? I'm talking practical here. Of course data is lost in a recording. That's the nature of most any kind of recording. The point is to cut out loss wherever possible. To record an image one has to use a lens (or series of lenses), and a medium to record the light onto (either film or CCD). These are unavoidable. However, a lossy compression scheme is avoidable. Furthermore, if one is going to save a compressed image, edit it, then recompress it using a lossy scheme -- it adds up. By your logic, there's no reason to clean your optics, because you're losing data one way or the other -- this is nonsensical.
You're also talking about the still-semi-obscure JPEG 2000 standard/codec, which you fail to mention isn't nearly as easy to use or widely avaialble as JFIF or PNG. I think this is a case of "my bike's made out of lead and I like it!". PNG is out there, it's free, full featured and it works very well.
-Turkey
Strictly speaking, a WAV file can contain any of dozens of representations of the audio data, many of which are lossy compressions. The WAV format just defines the file structure. There is an entry in the header that specifies the audio data representation. Much of the time WAV files contain linear PCM data, which is uncompressed, but they don't have to. In fact, theres no reason in principle that a WAV file couldn't contain FLAC compressed data, though I believe that no FLAC identifier code has been registered.
Wow...nothing like people quoting bad law on /. and getting modded up on it. Here is how it really works:
For patents filed prior to June 8, 1995, the length of the term is 17 years from date of issuance, regardless of the length of prosecution. Thus, you can keep the patent sitting in the patent office using continuing applications and other tricks, and then change the claims of the patent to closely match emerging technology. Also, because patent application publication did not automatically happen in apps filed 1995 and before, people would not have an opportunity to see that there was a pending patent that could cover their technology.
Apps filed between June 8, 1995 and May 28, 2000 have a term of 20 years from date of filing. However, this term of 20 years can be extended based upon delays in prosecution regarding secrecy orders, interferences, and/or successful appeals. (The URAA changed this law).
Apps filed after May 28, 2000 are given 20 years after filing, but they can be adjusted based upon on the number of days of delay caused by the PTO minus the number of days delay caused by the applicant. These adjustments are rather complicated, and usually not worth getting into (unless it is a pharm patent where keeping generics off the market mean huge money per day).
While I might be a patent lawyer, the above is not legal advice in any way.
The patent will actually expire on October 27, 2006 -- U.S. patents that were issued or pending as of June 1995 expire either 17 years from issuance or 20 years from earliest claimed priority date, whichever is longer.
Secondly, they are only entitled to damages for the 6 years preceeding the filing of the complaint (see e.e. 35 USC 286).
Thirdly, according to other stories and press releases, they did try to negotiate licensing agreements with these companies but couldn't reach an agreement. Patent lawsuits are expensive (2-5 million) and you try to license/settle and only resort to filing a lawsuit as last resort. Perhaps their motives were not so evil -- patentees typically don't want damages - they want royalties (easier to deal with and less costly to get) so maybe they just did everything they could to try and license/settle.
Finally, what is y'all obsession with submarine patents -- they don't exist anymore. If you have something since 1995 -- it is only entitled to a 20 year term from its first filing date/priority date. No reason to submarine anymore -- plus recent court decisions have applied a "unclean hands" type of rejection of lawsuits based on so-called submarine patents (see Lemelson's bar code patent fiasco in Nevada).
Just thought I should correct some of the erroneous assumptions.
For software-developers, the best thing to do about patents is to figth politically against them, and otherwise ignore them as best as possible. The *only* way to be certain you won't infringe some software-patent is never to release any software at all.
The person filing the patent does his best to draw out the process and make modifications to the patent in order to claim prior ownership of other people's recent work.
http://www.algovision-luratech.de/company/news/pat entquarrel.jsp?OnlineShopId=1075701030541733420#4
(Remove the space after "...pat" - the slashdot formatter mangled the URL)
I would imagine that a similar standard applies to patents. But you would be imagining wrong. As opposed to other areas of the law, laches in patent cases only limits the amount of damages (past 6 years) as opposed to prevent the entire suit. If you engage in conduct that specifically endorses its free use, you could be in trouble. However, that does not appear to be the case here.
Going to have to call you on that one. Having implemented the JPEG2000 standard myself in the course of my job, and having worked next to the guy who implemented JPEG for us, I can tell you that they are completely different beasts (except for the name).
JPEG2000 uses a wavelet transform on the entire image normally, not on 8x8 blocks, and it has about a gajillion more options you can use in it. As such you could set up tiling to operate the wavelet on 8x8 blocks, but that would be amazingly innefficient. Further, they ditched the huffman encoder and got an arithmetic enocder instead. In sum, JPEG2000 was not a rehash of the old, it was in fact a complete start from scratch implementation which should not in any way be encumbered by the original JPEG patent baggage.
You're a bit off-base... PNG is a lossless format. It fits in with the RAW and TIFF.
My blog. Good stuff (when I remember to update it). Read it.
As to the patent encumberedness, yes there are some patents governing JPEG2000, however it would appear the ISO/IEC have done their homework in this regard. From the last page of the standard, after listing people who hold patents related to the standard:
"The holders of these patent rights have assured the ISO and IEC that they are willing to negotiate licences under reasonable and non-discriminatory terms and conditions with applicants throughout the world. In this respect, the statements of the holders of these patents right are registered with the ISO and IEC."
I am not aware that any patent holder has sued over patent rights related issues in JPEG2000, and some of them (look at the JJ2000 copyright notice) agree not to mess with you over IP issues if you use their stuff in a JPEG2000 implementation. I'd like to think that they've become a little more wise about this sort of thing in the wake of the JPEG/GIF fiascoes.
Ever heard of Motion-JPEG?