it’s not exactly deep blue vs. kasparov

Much like those eminently satisfying reports of wine experts preferring £4 Sainsbury’s plonk in a blind taste test, Gramophone magazine finds that an obscure pianist who built a cult following on the back of an astonishing variety of accomplished performances in the years before her death actually just nicked the music off other people. Entire CDs, with track orders intact, were repackaged with her name. AND NOBODY FIGURED IT OUT until some dude stuck a CD into his iTunes, which automatically connected to the Compact Disk Database. The CDDB is a service which identifies CDs by their “fingerprint” - a crude system that looks at the exact length of tracks down to the millisecond, the number of tracks, and says “well this CD simply MUST be [x]“. In this case, it identified Joyce Hatto’s CD of Liszt’s 12 Transcendental Studies as being, well, Lászlo Simon’s CD of Liszt’s 12 Transcendental Studies. The archives were immediately checked and loads of Hatto’s stuff turned out to be by other people.

All of which immediately provokes two questions.

First, has classical music now reached a saturation point of recording when not even professionals have the ability to recognise identical versions of famous pieces?

Second, should these professionals perhaps stop putting us all on that they can distinguish the subtle nuance they say they can and do something more useful than complain about the “flat colouring” of the soprano section in the second act of a three-champagne opera?

1. Tom on February 21st, 2007

   OMG this is the purest branding story I’ve seen in ages - thankyou Tracer!

   Also there’s a bit in the Manual on this in pop which I am too busy to paraphrase now :(

2. FT's pˆnk s lord sükråt cunctør on February 21st, 2007

   i would honestly have thought saturation was reached in the 70s, in ref. repeat perfs of canon works! (there is FAR more marketing bogosity infesting the classical niche than ANY OTHER SECTOR)

3. FT's CarsmileSteve on February 21st, 2007

   the best game to be played with CDDB i think is to get a pre-release copy of, say, the new travis alBUM and call all the tracks things like “WHY DID YOU BUY THIS RUBBISH, MORAN???”, submit it, and see how long it takes to get corrected…

4. xyzzzz__ on February 21st, 2007

   “The CDDB is a service which identifies CDs by their “fingerprint” - a crude system that looks at the exact length of tracks down to the millisecond, the number of tracks, and says “well this CD simply MUST be [x]”.”

   But in the link: “Joyce Hatto plays Liszt - or rather, no she doesn’t, Laszlo Simon does - for ten of the twelve tracks on this CD.”

   so its 10/12 - Shouldn’t 12/12 tracks from the CD be lifted (as I first thought when I read about it last week) so that it could come up as Lazlo?

5. FT's tracerhand on February 21st, 2007

   you’re right, that part doesn’t make sense. unless simon ripped off nojima!! (i suspect that the CDDB made a “mistake” that was close enough to the truth.)

6. JohnneyB on February 21st, 2007

   I think you’re being a little disingenuous to all the people who were duped by the identical performances. By all accounts a great deal of them had been digitally altered to sound like the gal was playing really hard notes, or time/pitch shifted to sound very different indeed. When a recording is being played at a slightly different speed it’s very easy to think you’re just hearing something out of the ordinary - see the debate about the playback speed of Robert Johnson recordings.

   Although yes, classical music is amazingly snobby and impenetrable unless you’ve got an in somehow, and it does give out the wrong vibes to those who aren’t really up on it. Behaviour like this isn’t helping.

7. tom w on February 21st, 2007

   classical listeners are allowed not to have ears for effects anyone listening to most any other sort of music critically oughta be expected to notice?

8. FT's cis on February 21st, 2007

   Nothing I’ve read suggests that they were digitally altered to sound like she was playing hard notes - a few were speeded up, and so it sounded like she was playing unusually fast, but that seems about the extent of it. If you listen to the Hatto/Nojima mp3 on the pristineclassical site, there’s a speed difference of 0.95% between the two and they sound identical (until after a while when they go out of sync): most aren’t timestretched by much more than 2%. I just checked by speeding up a piece of Richter in Quicktime, and it’s still very recognisably his version, his phrasing. The only ones where the sound was heavily altered and sounded ‘very different indeed’ were, it appears, ones where the speeding up had resulted in its actively sounding bad, like that one at about 22% faster than usual. You can’t mess around that much with the pitch of a recorded piano, the notes start to sound warped.

   The problem here is, I think, the plethora of alternate recordings that means that people can get away with this - if you’ve ever listened to radio 3’s ‘building a library’ (one of my favourite radio shows, and it’s devoted to comparing different people’s recordings of the same piece), it’s staggering how many cds are released of the same thing! Being ‘into’ yer classical music generally means developing that collector’s mindset by which you pick a favourite among rival recordings, but no-one can buy them all, and keep them all in mind when listening to one. And, really, you don’t expect anyone to be so ungentlemanly as to do such a thing as this.

   I don’t really understand how CDDB could have been used to find this out - as said above, the track length would have changed and so CDDB wouldn’t have worked. If I were conspiracy-minded I’d say this implies that whoever discovered the Lazlo Simon connection knew about it beforehand through less coincidental means, and used the CDDB-error story as a way of bringing it to light without incriminating any sources.