French Archives DVD longevity study from 2008
 

How long do DVDs last? Well, that's a question being asked all over the world -- not just in English speaking countries! Attached is another interesting aging study from Europe, this time performed by the French National Archives back in mid-2007. (To no surprise, it wasn't until late 2008 that the report was released. That's government speed, for you!)

What's unique about this one is that the test has specifically tested known-good (or thought-to-be-good, ala MAM-A/MAM-E) media. Rather than testing crap, they've turned their attention on respected media like Sony, Taiyo Yuden, Mitsubishi and Falcon. You won't find RITEK and CMC in this test -- time wasn't wasted on second-tier and other lower-performing inferior media.

What I found most interesting was the much lower quality returns on MCC "Archival Gold" media as compared to normal MCC discs. I'm left wondering if the different Azo formula is to blame (either as wholly inferior, or simply an inferior drive/dye match), or maybe it's simply a case of bad/damaged discs having been used in the tests.

Both Audiodev CATS and Clover testing systems were used, although I wonder about the calibration methods. That could have have offset something -- and when you get a weird return on analysis data (like Verbatim Archival Gold being crap), you really need to re-run the tests after re-calibration.

Another issue of concern is the use of Nero as the burning application. Not all burning software necessarily operates as it should, and Nero is infamous for its various quirks and issues. In my mind, that almost taints some of the test results. I'd be far more inclined to burn test discs with ImgBurn, Prassi ONES or RecordNow -- never Roxio or Nero software.

All pertinent values, from PI/sum8 to jitter % were given in the various charts that make up the 112-page document. Note that this is from a reference drive, not a home drive, so values can be more carefully analyzed (although not necessarily immune from error, or even valuable in terms of discussing absolute values of the media).

Aside from the Verbatim Gold blip, I don't see anything out of the ordinary here.

Notice how MAM-A/Mitsui media started off with more errors and slowly turned to crap for reading. That's about par for the course, in terms of testing their lame claim that the media last 100 years. No, I don't think I agree with their assertion. Unlike the Verbatim Gold, I don't think this is a blip -- I've seen this quite a bit from various tests, including our own. MAM-A media starts off harder to read, and with more errors. While the discs never seem to entirely "die", the data is hard to retrieve and may be impartial at best --the "loss curve" is more flat compared to others that start out great and go entirely unreadable.

Don't speak French? No problem. Between my command of English and Spanish, my limited knowledge of French, and the use of translation software, I can make my way through this document without a lot of trouble. You probably can, too.

Read it here: http://www.archivesdefrance.culture.gouv.fr/static/2140
Or download the attached PDF.