October 16, 2007
News: After much delay, YouTube finally deployed its filtering type system for copyright owners. The system apparently is based on a database of copyrighted videos submitted to YouTube by the content owners that can be used to automatically identify copies of the same video uploaded to YouTube without authorization. YouTube makes a digital fingerprint of the submitted videos, which then can be used to track copies of the same videos. The system even reportedly is able to identify close copies of a copyrighted show, such as when someone tapes a show by pointing a video camera at it. Google CEO Eric Schmidt says no system can be 100% effective; realistically, they’re shooting for 80 to 90% effectiveness. Google and YouTube developed the system with participation from Time Warner, CBS, Disney, NBC Universal, and Viacom, the latter of which is, of course, suing YouTube for copyright infringement. (More and here.)
Analysis: The filtering system sounds pretty sophisticated. The catch, of course, is that YouTube needs copyright owners to submit their content to YouTube’s database. Or at least that’s how it’s being described in the press reports. Some copyright holders probably won’t think this filtering system is enough. As Rob Gould, marketing VP for Broadcaster.com, said, “If there has been a clip from ‘American Idol’ posted to the site by Joe Schmoe in Oklahoma instead of Fox, you can be pretty sure it’s not supposed to be there.”
April 25, 2007
News: NBC, owned by Universal, will be testing out YouTube’s deployment of content-identication software (reportedly from Audible Magic out of Los Gatos, CA). According to WSJ:
“Fingerprinting technology, by analyzing the audio or video tracks of a clip, could alert YouTube to the presence of material that a media company has registered as its own — regardless of who uploads it or what they title the clip. General Electric Co.’s NBC-Universal says it plans to participate in a test of fingerprinting on YouTube that it expects to start shortly. Technical staffs from the two companies are working together and they hope to have results by this summer, according to NBC.”
Analysis: Glad to see the two sides working together to deal with copyright infringement. According to WSJ, the process will work something like this: “Google is expected to use fingerprinting to flag pirated clips to the content owners, which then have to request they be removed.” Of course, fingerprinting is a little bit short of automated filtering. WSJ: ‘It sounds like some kind of crazy lost and found,’ a senior executive at one of the big media companies says. ‘It’s not going to be enough,’ says another.” I guess I think everyone needs to be patient and give this technology a chance to develop.
Here’s how Audible Magic describes its own system: “Patented CopySense identification techniques recognize media content based on digital “fingerprints” derived from perceptual characteristics of the content itself. The approach is highly accurate and requires no dependence on metadata, watermarks or file hashes. Best of all the technology is highly immune to compression or distortion, and it is indifferent to file or streaming format. It’s the best approach for recognizing content “in the wild.” Integration of CopySense technology is made simple with an efficient and compact API library.”
April 19, 2007
News: YouTube’s Julie Supan today described the forthcoming filtering technology from YouTube, known as “Claim Your Content.” According to one account:
“The technology does not identify copyright material,” Supan said, “because we can not distinguish between copyrighted material on YouTube that the owner wants on the site from material that is put up without the owners’ permission.”"What we are testing,” Supan went on to say, “is identification technology that will help content owners identify and locate their content on YouTube. This will facilitate their ability to ‘claim their content’ and leave it up for promotional purposes or, if they wish, seek to remove it from the site.”
Analysis: Hmmm. More should be said about this filtering. Given the anticipation for its deployment, I’m starting to fear it may not meet the growing expectations.
April 17, 2007
News: At the National Association of Broadcasters, Google CEO Eric Schmidt said that YouTube will soon launch a filtering system called “Claim Your Content,” which reportedly “will automatically identify copyright material so that it can be removed.” (More here)
Analysis: Not yet clear whether this is the Audible Magic filtering technology for audio that had been announced back in February, or something more.
February 24, 2007
News: After partnership deals with Viacom and then CBS broke down in the past week (with Viacom actually going to startup Joost to make a deal), CEO of Google Eric Schmidt announced yesterday that fighting copyright infringement is one of the company’s highest priorities.”We are definitely committed to (offering copyright protection technologies). It is one of the company’s highest priorities. We just reviewed that (issue) about an hour ago. [Anti-piracy filtering] is going to roll out very soon … It is not far away,” Schmidt said. Correcting earlier reports (or changing them) that YouTube would only offer filtering to partners, Schmidt said the filtering would be available to all. “We have to do that. But it takes a while to roll this stuff out.” (More here)
In a separate report, it was also announced that Google-YouTube will be adopting Audible Magic’s filtering, a popular digital fingerprint recognition system. According to the report, “[t]he system works by comparing the audio fingerprint of a video to a large database of copyrighted material. ” (More here)
Analysis: It looks like adult supervision is stepping in. Google CEO Eric Schmidt probably felt compelled to speak out publicly against copyright infringement because all of the bad press YouTube received in the past week, not to mention the two deals with Viacom and CBS that reportedly broke down. But no one should expect that Audible Magic will make unauthorized clips all miraculously disappear. Its filtering, first of all, is limited to audio fingerprinting (as I understand it), and, in any event, no system of filtering will come close to being 100% effective. But that’s not a bad thing. Copyright has always been “leaky,” going back to the founding of the U.S. in 1789. Copyright holders deserve a “fair return,” said the Supreme Court, but no more than that: “The immediate effect of our copyright law is to secure a fair return for an ‘author’s’ creative labor. But the ultimate aim is, by this incentive, to stimulate artistic creativity for the general public good.” Twentieth Century Music Corp. v. Aiken, 422 U.S. 151, 156 (1975).
February 19, 2007
News: YouTube is late in rolling out anti-copyright-infringement tools, and now it reportedly has said or suggested that it will extend the filtering only to businesses that have secured deals with YouTube. One media source reportedly described this tactic by YouTube as close to a “mafia shakedown.” Said a Viacom rep, the “proposition that they will only protect copyrighted content if there’s a business deal in place is unacceptable.” (More here)
Analysis: If true, this maneuver by YouTube seems very questionable to me. YouTube should try to set a corporate policy of working with copyright holders to stop infringement, as far as practicable. Dangling filtering software like a bargaining chip makes YouTube sound opportunistic. This tactic could just agitate copyright holders into suing YouTube.
February 13, 2007
News: MySpace beat YouTube to the punch today, announcing the deployment of software tools to help copyright holders identify audio files that have been incorporated into video without authorization of the copyright holders. The copyright holders can apparently send digital fingerprints contained in audio files to MySpace for its database; MySpace’s filtering presumably then attempts to identify video and audio files uploaded onto MySpace that contain the digital fingerprint. Any file that does so (without authorization) will be blocked. (More here)
Analysis: I can’t help but think this is a positive development for all interested parties. We don’t know how well this filtering technology works yet, but it’s a start.