From research done by a veteran specialist in the signal processing field it appears that the 900 patents held by MPEG-LA are each rather weak on their own, but together form an impenetrable thicket that makes programming in this area a hazardous exercise and license-payments to the pool unavoidable. Among the patent holders in the MPEG-LA pool are many of the large companies who are currently found lobbying for the Software Patent Directive.
Projects such as !MPlayer and VideoLAN are currently leading a web demo, warning their users that the software patent directive as promoted by the Commission and Council would most likely mean the end of these projects.
The !MPlayer project writes on a protest web page:
- MPlayer is seriously threatened by software patents due to the numerous patented multimedia techniques. Also threatened are the many programs built upon MPlayer and the other free software multimedia players, like xine, VLC, avifile, gstreamer and especially FFmpeg, which provides the framework all of the above players use. Multimedia is a patent minefield. All important techniques and formats are covered by broad and trivial patents that are harming progress and alternative implementations, such as free software multimedia players.
Below are summarising comments on the study draft.
Summary of That Draft Study
All patents in the list of the MPEG licence association in regard to the MPEG-4 standard were examined and analysed. After intensive study of relevant literature and more than 100 patents of the relevant companies we can say now: Upon careful examination, we can not find any advances over the prior art in said list that could justify the granting of a patent. Most of these patents should be attackable in court, but who would take the burden of litigation against 900 patents owned by dozens of large companies ?
See the complete report at
Apple Computer, Inc.
Electronics and Telecommunications Research Institute
- KR 248,404
France Télécom, S.A.
Koninklijke Philips Electronics N.V.
Mitsubishi Electric Corporation
Samsung Electronics Co., Ltd.
- CN ZL 97112956.8
Sharp Kabushiki Kaisha
Sun Microsystems, Inc.
SUMMARY of the MPEG-LA-Analysis:
After studying relevant literature and more than 100 patents intensively (including all EP- and WIPO-patent applications), one definitely can say:
The patent claims do not fulfil any prerequisites of article No. 52 of the European Patent Convention : they unexceptionally describe programs for computers (rules and algorithms for operating known data processing equipment). Computer-implemented concepts; nothing more.
But even if for a moment we accept the EPO's strange interpretations of Article 52, we find that, compared with the prior art, there is no invention in said MPEG-4 list with an "inventive step" that could be sufficient to justify the granting of even one of these patents.
Most of all "inventive steps" in said patent claims are well-known time-domain, but packed into nebulous or indefinable semantics, such as:
"*Parameter*", associated to a "code-block" that represents and indicates, whatever one could imagine, (PHILIPS, see claim 1 of EP0460751A2)
"*control signal*", which is "indicative" for all sorts of something.. (PHILIPS, see claim 1 of parallel German patent DE69127504C0)
"*portions of signal*", made up by an "algorithm" to a code (PHILIPS, see claim 1 of US 5,844,867)
Some "inventions" are re-patenting matter. Technologies, that have been described already 20 years ago; simply implemented and/or added into new patent-claims to pretend novelty:
- "*adressing*" of data packets for identification (Philips)
- "*time stamp*" or "*elapse time relation*" to data bytes or data packets (SUN Microsystems)
- "*multiplexing*" (MUXING) of time-relevant data oder other data (Samsung)
Other cases are odd and peculiar:
- "*graceful degradation*" (ETRI)
Related to such technologies - known since decades - and to nebulously and undefinably formulated semantic monsters, dozens of additional patent claims got based upon. PYRAMIDS of new patents and patent-families became founded thereupon:
Finally, let us analyse patent applications published before 1997 regarding those types of "prior art": Inputing the term "*multiplex*", a professional patent search database shows 52 690 matches until May 1, 1997.
On inputing "*data block addressing*": 944 matches; on "*frame addressing*": 130 matches ! On inputing the word "*time stamp*" 2 962 matches, on "*time address*" 813 matches; on "*elapse time*" 14 373 matches!
When inputing the term "*data packetising*" you get 6,354 matches.