Philip Beck, an attorney representing Google, told a U.S. District court in New York last week that the company’s video filtering technology could be “up, running and effective”, “hopefully” by September.

Beck’s declarations, made during a scheduling hearing in Viacom’s $1 billion suit against Google, appeared to be one of the most concrete indications of when the widely anticipated software will be available.

Thing is, it might not be. Google PR clarified what Beck meant a few hours after the statement:

Today we’re working with some of the major media companies to experiment with video identification tools. We’re excited about the progress so far, and we’re dedicated to making these tests successful. We hope to have the testing completed and technology available by sometime in the Fall, but this is one of the most technologically-complicated tasks that we have ever undertaken, and as always with cutting-edge technologies, it’s difficult to forecast specific launch dates.

So there you go. The software — designed to block video clips from being uploaded without the permission of the companies that own the content — might take a bit longer to reach media executives’ hands.

Beck also offered one of the more colorful definitions of how this “complicated” (in Google’s words) system would work: [from the court transcript]

A studio could say here’s our new Tom Cruise movie and we don’t want any knuckleheads with their video cameras in the theater videoing our movie and then uploading a 10-minute segment onto YouTube, so take our movie. And we would get the movie from them, and then our computers would do their magic, and that’s what’s being worked on now, so there would be key information extracted from the video and stored on a computer. … [A]ny video that gets uploaded basically gets filtered through the fingerprint database, and like the AFIS that the FBI has, and if there’s a hit, then within minutes the computer knows that and pulls it down.