Nov 22, 2004, 12:47 PM
Link: iSweep Away Duplicate Tunes More Efficiently Than iTunes 4.7 (http://www.macbytes.com/link.php?sid=20041122124708)
Posted on MacBytes.com (http://www.macbytes.com)
Approved by Mudbug
Nov 23, 2004, 03:48 PM
Unfortunately, detetcting duplicates is far from simple.
Apple's method of looking for tracks with identical title and artist is OK. Using some other metadata may also work.
But the problem is uglier than that. For instance, you may find the same title/artist song on multiple albums. Some are really duplicates - like the original album and a "greatest hits" compilation album. Some are not - like the original album song and a live performance.
There's no way software can automatically know the difference. Even when the tracks are the identical performance (original vs. compilation), the lengths may be different by a few seconds and the equalization/mixing may be different. So simply comparing the files won't work. You actually have to listen to both and use other off-line information (like liner notes) to reliably determine if the two tracks are really the same. And (in the case of alternate mixing/EQ) you need to listen to determine which one sounds better.
It can get even worse when you have different edits of the same material. For instance, Jethro Tull's A Passion Play (http://www.collecting-tull.com/Albums/APassionPlay.html). This is a single 45-minute composition. The first release on CD has two tracks (representing the two sides of the LP and cassette versions) while later releases of the same mix have the entire song in one track. There's also a Mobile Fidelity (http://www.collecting-tull.com/Albums/APassionPlay_MobileFidelity.html) gold-disc edition that has it in 16 tracks (the movements that comprise this single composition.) There is no possible way software can determine that the one-track, two-track and 16-track releases all contain the same content from the same original master recording.