Tuesday, February 20, 2007

iTunes: Infringement Detection Tool?

For my birthday, I was given the mega iPod video. I don't have any videos on it though because it is not yet possible to load a lawfully purchased DVD on to the iTunes store on my MacBook Pro and then import it to my iPod. So, I have been importing lots and lots of lawfully purchased CDs. When you insert a CD into your iTunes library, many things happen before you import it. A story in the New York Times on February 17th by Alan Riding, reporting from Paris, reveals an unexpected one.

The story, entitled "A Pianist’s Recordings Draw Praise, but Were They All Hers?," concerns the late British pianist Joyce Hatto, who Mr. Riding reports, "was rediscovered by a small group of musicians and critics who contended that her recordings of Chopin, Rachmaninoff, Liszt and others ranked alongside those of the 20th century’s most exceptional virtuosos." He also observes, that "she may have been less prolific than she seemed. ... Ms. Hatto’s reputation for excellence and originality has been shaken by a charge of plagiarism. Gramophone, the London music monthly, has presented evidence that several of the recordings issued under her name were in fact copied from recordings of the same music by other pianists."

Here is where the iTunes angle comes in:

"Jed Distler, a composer and music critic who was among Ms. Hatto’s admirers, contacted Mr. Inverne [of Gramophone] with a strange story. When he put the Hatto CD of the Liszt ├ętudes into his computer, Mr. Inverne recounted, “his iTunes player identified the disc as, yes, the Liszts, but not a Hatto recording.” Instead, it identified Mr. Simon as the performer. "

"Looking for scientific evidence of a hoax, Mr. Inverne then sent the Hatto and Simon recordings to Mr. Rose, a former sound engineer for the BBC. Mr. Rose said that the Liszt recordings were easy to identify as those made by Mr. Simon, but that the Nojima recording had been 'manipulated' to disguise its origin."

“'If all this is true,'” Mr. Inverne said, “'what strikes me is that this sort of piracy was made possible by technology, and later advances in technology uncovered it.


Anonymous said...

I'm afraid this doesn't make sense to me. iTunes uses the Gracenote CDDB database to supply metadata when a CD is inserted. This database is built around the unique combination of track numbers, track lengths, and a couple other simple bits of data that are read when a CD is spun in a PC drive. There is no "acoustic fingerprinting" or other sophisticated "recognition" of individual recordings.

So the only way the match with another CD could have been triggered is if the combination of track lengths, track number, etc, happened to match (these collisions happen relatively frequently by coincidence).

This would almost certainly be a coincidence, rather than evidence of plagiarism. Anyone intent on passing off a recording as one's own would be foolish to leave all the tracks in the same order, and the lengths, as the original. And if steps had been taken to obfuscate the recordings (speeding them up slightly, etc), then the same thing would have resulted in no CDDB match.

In short, this sounds extremely fishy to me.

Jordan said...

More on the iTunes angle. ITunes uses (correct me if I'm wrong) Gracenote's CD Database to determine the track listings of discs inserted and imported to the player. Gracenote's database of tracks was originally built by its users -- who put in by hand the information for each track on an albu, (think wikipedia). Gracenote, it is true, now gets track info straight from the music industry, but still accepts user contributions.

So it sounds to me that some user of Gracenote out there put in the other artist's name for a track, and it just so happened that someone important enough to get media attention looked into it. Maybe it is an example of open-source style collaboration (enabled through technology) rather than a computer program being able to autonomously evaluate the author -- as the story makes it sound.

William Patry said...

Thanks, Fred, I am in DC today attending the AT&T v. Microsfot argument, but will research more when I get back

Christopher Fulmer said...

According to one of the Gracenote bigwigs, Gracenote actually goes beyond track numbers and lengths. Apparently, the music industry will often change track lengths slightly when it reprints a CD. So, Gracenote claims to actually play with the audio to identify the track.

Some details here: http://www.wired.com/news/columns/0,72105-2.html?tw=wn_story_page_next2

y-intercept said...

Great story. It seems to me that the best way to protect the rights of artists is large repositories of known compositions (rather than snooping on individuals).

When some one's plagiarized work becomes popular, something somewhere will eventually match the copy to the source and uncover the masquerade.

Anonymous said...

"When some one's plagiarized work becomes popular, something somewhere will eventually match the copy to the source and uncover the masquerade.
Well, at least dozens of CDs later...maybe.

Anonymous said...

Two fascinating posts from the sleuths -

Andrew Rose - visual waveform match and side-by-side listening (left Hatto, right the other pianist)

Centre for the History and Analysis of Recorded Music (CHARM) - timescape visualisation

In conclusion, Rose states "We have yet to investigate a Hatto recording that has not proved to be a hoax."

The husband, who produced the "recordings" is in deer-in-the-headlights mode, and has yet to be able to produce any substantial information supporting independent creation of the CDs.

Anonymous said...

Further updates to this story on Wired's Listening Post blog.

Turns out Apple is NOT using Gracenote's sophisticated acoustic recognition tool, so this suggests Hatto simply made a track for track copy of the existing CD and released under her own name. Somewhat embarrassing that it took the classical music establishment so long to catch on.

William Patry said...

The New York Times of February 27th, page E3 of the Arts section has a story by Alan Riding in which the pianist's husband admitted to the fraudulent recordings. Here is a link to Andres rose's site with the acoustical evidence: http://pristineclassical.com/HattoHoax.html

Anonymous said...

With the recent article in the New Yorker and an interview I heard on NPR, I am wondering if anyone has filed a copyright infringement claim for these acts of infringement?