On 7/13/05, Raul Miller <[EMAIL PROTECTED]> wrote: > It's difficult to create JPEG image rendering software without using > technologies described by MP3 patents.
So do tell us where MP3 patents fit -- what patents, which claims, and what part of JPEG compression? None of the patents discussed below is part of the Fraunhofer / Thomson or Philips / Sisvel / Audio MPEG suite. Let's start with http://www.w3.org/Graphics/JPEG/AnnexL.html , which was long generally considered an exhaustive list of the patents remotely applicable to JPEG, none of which has been claimed against baseline JFIF compression in fourteen years of commercial use (at least ten of ubiquity). Seven of the ten patents originally listed in Annex L were assigned to IBM. (Note that #4,725,884 is actually an irrelevant patent entitled "Optical Metrology"; the Gonzales, Mitchell, Pennebaker patent is #4,725,885.) Two were assigned to AT&T Bell Labs. The last was originally granted in Japan and assigned to Mitsubishi; its US equivalent is #5,311,177. Subsequent disputes have arisen based on other patents -- see http://www.jpeg.org/newsrel1.html -- but I for one am not impressed. The principal aggressor in current JPEG litigation is the successor to Compression Labs, Inc. (among others) -- now a blatant IP litigation shell company called Forgent Networks, with a textbook pump-and-dump stock swindle attached. (That's my own unqualified judgment; but check out their web site and Google "forgent networks stock". See also http://www.aspnetresources.com/blog/forgent_jpeg_saga.aspx and references therein.) IANAL, TINLA, etc. -- and it's also not investment advice. Forgent's suits have been consolidated (perhaps not including a later one against Microsoft) and seem to be headed for a swift death (well, swift as these things go) at the hands of the Honorable Phyllis J. Hamilton of the Northern District of California. (I prophesy that Godwin Gruber LLP is going to regret taking this client.) That litigation is principally about #4,698,672 -- not in the Annex L list -- which a quick review persuades me is a comically trivial tweak to #4,302,775 (expired). There are actually two more patents mentioned in the current edition of Annex L: one assigned to IBM and replacing #4,725,885, plus #4,665,436. I would humbly put the latter in a category with my own patent: part algorithm and part circuit description, clever in its adaptation to painfully tight resource constraints, funded by an incorrigible cowboy operator, impossible to evaluate its applicability to MPEG without litigation. Doubtless, if it had not been permitted to expire for lack of maintenance payment in 1999, it would also be in the portfolio of a similar make-money-fast-with-lawsuits company, as would mine. (While you're at it, check out #5,533,051, which appears to contain the only other reference within Google's reach to that "Optical Metrology" patent. Not only is it a variant of perpetual motion, the applicant cut and pasted his prior art references from places like Annex L, and the examiner didn't check them. I would not trade places with Mr. Scott A. Rogers for all the tea in China. Did I mention that, given the potential cost of litigating an obviously stupid patent, the USPTO is hard to defend for more reasons than just corporate 0wn3rship?) Cheers, - Michael P. S. See also http://www.faqs.org/faqs/compression-faq/part1/section-7.html .
