Slashdot Mirror
The Myth Of The 100-Year CD-Rom
Toshito writes "Are we putting too much faith in the ubiquitous "recordable CD", or CD-R? A lot of manufacturer claims 100 years of shelf life for a CD-R. But in real life, it can be much less. Expect failure after only 5 years... Personnaly I just discovered 6 audio cassettes with the voice of my late grandfather, talking about old times. These tapes are copies of reel to reel recorded in 1971, and they are still in excellent shape.
I was thinking about digitizing everything, do a little noise reduction, and burning this on CD's, for my childrens and great grand-childrens enjoyment, but it seems that old analog tech from the '70 is more reliable than digital. The full story at Rense. Other links about the subject: Practical PC, Mscience, and an excellent reasearch by the Library of Congress (warning! PDF): Study of CD longevity, html version (google):Study html."
I know lots of people that have "worn out" cd's. The first time I heard that, I thought they were kidding, but no... even if you take super great care of say, an audio cd, it will eventually wear out. It's especially bad if you keep it in the original plastic jewel case, and take it out each time -- my friend's rare Pearl Jam CD's are nearly scratched beyond playability, but he was able to extract the digital information before it got lost. What makes CD's better than tapes is that the 0's and 1's will always "be the same" logically, unfortunately the physical media wears out quickly with use. I prefer to think of CD's as a temporary storage mechanism for a permanent idea, like a sketch on newsprint. Once the newsprint disintegrates, you'd better hope you made something good with the idea... it doesn't mean the idea is gone, but the medium isn't like stone.
stuff |
Well, the recordings *will* go through decay, but that's what the constant backing-up process is about. Your basic point is right on the money, though.
The only way to keep bits in any kind of order and in good condition over a long period of time with the kind of technology available to the average consumer is to keep making multiple fresh copies before each individual storage media begins to suffer loss of data.
It's not exactly a fair comparison between CD-R and analog tape for audio. The audio tape isn't "more reliable." It just degrades differently.
As the tape ages, the quality of the audio signal degrades dramatically, but because it is an analogue signal, it can still be deciphered by or ears.
With digital medium, the audio never gets worse. As the media degrades, it just reaches a point where it isn't able to be deciphered as audio data.
If you want to compare the mediums (magnetic tape vs. CR-R), data is probably a better place to do so. You can easily measure the amount of readable/unreadable data in bytes and make a fair, quantifiable comparison.
The first link is to rense.com, a website that promulgates the theory that the US government is experimenting on us with "chem trails" emitted by otherwise innocuous-looking aircraft flying overhead. The webmaster at that site obviously has a very low threshold for rubbish, and no critical thinking ability!
It's the glue. It can corrode the top layer. I've seen some stories about it, still a few floating around, seems to depend largely on which one you use. The problem may be mitigated by this point, now that they know, but I still don't trust labels over cases/sleeves. You have to get that label on really well, air bubbles being your enemy. :D
Auto-reply to ACs: "Truly, you have a dizzying intellect."
The BBC Library still uses vinyl records for long term audio storage. For some items they cut a lacquer master, plate the metal stampers on the lacquer and leave the metal stampers attached to the lacquer.
They believe that this will preserve the audio for about 300 years and they say that vinyl is the only storage medium with a real and predictable life span.
Audio processing technology will get better. Don't ruin your grandkids' heirloom recordings by using today's technology to permanently alter them.
Make working copies and filter those as much as you want, but keep those masters pristine! Maybe somewhere in the background you can hear your grandma yelling at dear ol' grandpa to put that thing away and paint the house, and a clumsy run with an agressive low-pass filter will throw that data away forever. You have something really valuable; please take care of it for the future.
Dewey, what part of this looks like authorities should be involved?
I'm sure what I'll say has already been said, but I can certainly attest to the shorter-than-advertised longevity of CDR media. I recently had to pull some long lost files off of CD's I burned back in the college days, probably 5 years ago or so. These consisted of several types of media, both cheap and expensive, green and blue dye, sticker and no sticker. Basically the dye color has little effect, and stickers really do call for the early death of the media. But most of all, I think it was the early CD burning software or the actual CD-Rec drive that I used. Some earlier CD's, that I know I burned at work (using the latest software at the time) were near flawless. But a batch burned later, on a friend's computer using some lesser known software, was completely corrupt (TOC and CRC errors abound). I now make sure I get decent CDR's like TDK's (not the cheap CompUSA stuff), don't use stickers, always keep them in a multi-CD case, and run a bit-for-bit check on the archive after burning with Nero. I have yet to have a problem since I started this practice at least 2 years ago...although time will certainly tell.
Sometimes I doubt your commitment to Sparkle Motion.
Although it seems like burning at a slower speed means that your data lasts longer, for some newer CDs burning at 2x might actually cause your data to be less secure. Most CDs sold nowadays are optimized for faster burns, say at 48x. The "fast" media doesn't handle slow burn speeds quite as well as older media optimized for 2x would.
Actually, 5.25" double-density disks have been shown to theoretically last 90 years, and many of these disks have lasted 20+ years IN PRACTICE (I have some 25 year old Apple II disks that STILL work without errors to this day).
Have you ever head of Reed-Solomon? There is redundancy built into CDs.
Pressed CDs shouldn't be as vulnerable to bit rot as burned CD-Rs. But I can't understand how the discs would lose quality. One either gets a valid frame of redbook audio or not. I can understand that some of the frames might go bad (even to the point where the built-in error correction can't help) and lead to audio defects, but I don't see how the whole disc would sound different than before.
Ita erat quando hic adveni.
The National Institute of Standards and Technology also has an extensive guide, Care and Handling Guide for the Preservation of CDs and DVDs, including a one-page Do-and-Don't Quick Reference.
This is all true. You may not know the vast difference in materials used for CD's.
If the CD feels sticky around the edges, it may (may) mean a low quality glue was used. It provides a potential path for fungus to migrate into your CD.
Gold reflecting layers (very rare to find anymore) are the absolute best. Gold generally doesn't react with the stuff in the atmosphere.
High quality archival stabilized dye layers are also hard to find anymore. Phthalocyanine was the absolute best last I looked (a few years ago) with an estimated stable lifetime of 200 years.
A CD that you want to hold data for 100 years should have a quality glue job, gold reflective layer and Pthalocyanine dye. I know of only two brands that have ever been made to this quality. One was Kodak Gold (some marketing suffix here), but it went out of production several years ago. The other is Mitsui Gold, which cost about $1 each in 100 packs.
And no matter how nice the CD manufacture is, it will not last unless properly stored. The three tenets of archival storage are: Cool, Dry and Dark. Don't leave your CD-R's on the shores of a tropical beach.
Education is a better safeguard of liberty than a standing army.
Edward Everett (1794 - 1865)
You're better off storing audio as FLAC or so; the format's open, lossless, streamable, error checking, robust, and has a proper metadata standard. Use the space it saves to make a bunch of PAR2's, and you're laughing.
Support for this stuff's not going to disappear overnight; you can keep specifications and reference implementations about if need be.
This means that if you have a 2X recorder, writing at 2X is *much* better than 1X. If you have a 32X recorder, writing at 32X will produce measurably better discs than writing at 4X, 2X or 1X. This has been true since around 1998 or so. It is quite true that you could get better results with some early 4X recorders when writing at 1X than 4X. However, none of those devices are current any longer.
The "writing slower is better" story is a myth. Please don't spread it further. And yes, if you want more information about disc testing Media Sciences is a company that is dedicated to disc quality and testing. I do not work for them.
... if you know what you're doing. First of all, there are specialty CD-Rs intended for archival purposes. These will inherently last longer than normal CDs for numerous reasons, assuming the manufacturers are not full of crap. To find these CD-Rs, check a photography store, as photographers tend to have a need for both archival and mass storage thanks to digital cameras. You will likely find some there.
Second, the biggest mistake most people make in CD archival is to write on the CDs with magic marker -- DO NOT DO THIS. The ink will, given several years, leach through the extremely thin plastic on the labelled side of the CD and pollute the optical layer, resulting in a ruined CD. Adhesive stickers, I'm told, are not much better. There are special CD-labelling markers out there, I don't know if they work well as I haven't tried them, but I doubt they're worse than a magic marker. I have found that writing very lightly with a soft, dark graphite pencil works well. If you're very paranoid, you might consider not labelling the CD at all and just be meticulous in returing it to its (properly labelled) case when you're done.
Additionally, store the CDs properly. Somewhere reasonable. Not in direct sunlight. Safely stowed in their jewel cases.
Of course, even doing all this, no one can tell you that your CDs will still work in 100 years. It hasn't even been 100 years since we invented the damn things, how do we know how long they will last? Still, these are steps that should allow your CDs to last for at least as long as a magnetic tape, and with perfect accuracy, as opposed to the slow degradation of audio tapes.
What we really need is something similar to the S.M.A.R.T. technology in harddrives nowadays, to alert you that "Listen, I'm getting close to reaching the limit of my error-correction techniques here. This media probably isn't going to last a whole lot longer. You may want to do something about that." Currently, there's really no way to tell until it's too late.
Random and weird software I've written.
Really, what do you expect when most people pick up spindles that all some from the crappy Ritek or Princo plants in Taiwan because they can get them for $9 a spindle? I've had those go blank on my shelf too, and now I know better.
Want a long lasting CD-R? Search the spindles to find the ones that are made in Japan. Sometimes these will be on the same shelf with the Taiwan ones, wearing the same packaging, and for the same price (if you're lucky). Usually these are made by Taiyo-Yuden, a high-quality CD-R manufacturer (and one of the co-developers of CD-R technology). Look for a frosted hub for positive ID.
For archival quality, you'll need to spend a couple of bucks a disc on media that has a gold reflective layer. The standard here has always been Mitsui (now branded as MAM-A). Even their silver discs are a cut above in quality.
Oh, while I'm here. In 1996 I scribbled all over a burned CD-R with various colored Sharpies, then last year cleaned it all off with carb cleaner. It hadn't migrated into the disc at all, and cleaned off without a trace. The data was fine. Anyway, I mention this because I hear people claim Sharpies kill CD-Rs all the time, and think it's nonsense. These people probably bought the cheap-o discs and are looking for something other than their own cheapness to blame it on. Oh, BTW, the scribble disc was a Sony, made by Taiyo-Yuden.
I have to burn them at less than max speed, apparently the more stable dye requires more laser power. Otherwise no surprises so far. (knock on wood)
(2004/02/17) in CD-Recordable FAQ:
CD-RWs are expected to last about 25 years under ideal conditions (i.e. you write it once and then leave it alone). Repeated rewrites will ccelerate
this. In general, CD-RW media isn't recommended for long-term backups or archives of valuable data.
The rest of this section applies to CD-R.
The manufacturers claim 75 years (cyanine dye, used in "green" discs), 100 years (phthalocyanine dye, used in "gold" discs), or even 200 years
("advanced" phthalocyanine dye, used in "platinum" discs) once the disc has been written. The shelf life of an unrecorded disc has been estimated at
between 5 and 10 years. There is no standard agreed-upon way to test discs for lifetime viability. Accelerated aging tests have been done, but they may not provide a meaningful analogue to real-world aging.
Exposing the disc to excessive heat, humidity, or to direct sunlight will greatly reduce the lifetime. In general, CD-Rs are far less tolerant of environmental conditions than pressed CDs, and should be treated with greater care. The easiest way to make a CD-R unusable is to scratch the
top surface. Find a CD-R you don't want anymore, and try to scratch the top (label side) with your fingernail, a ballpoint pen, a paper clip, and
anything else you have handy. The results may surprise you.
Keep them in a cool, dark, dry place, and they will probably live longer than you do (emphasis on "probably"). Some newsgroup reports have complained of discs becoming unreadable in as little as three years, but without knowing how the discs were handled and stored such anecdotes are
useless. Try to keep a little perspective on the situation: a disc that degrades very little over 100 years is useless if it can't be read in your
CD-ROM drive today.
One user reported that very inexpensive CD-Rs deteriorated in a mere six weeks, despite careful storage. Some discs are better than others.
An interesting article by Fred Langa (of http://www.langa.com/) on http://www.informationweek.com/story/showArticle.j html?articleID=15800263&pgno=1
describes how to detect bad discs, and discusses whether putting an adhesive label on the disc causes them to fail more quickly.
By some estimates, pressed CD-ROMs may only last for 10 to 25 years, because the aluminum reflective layer starts to corrode after a while.
One user was told by Blaupunkt that CD-R discs shouldn't be left in car CD players, because if it gets too hot in the car the CD-R will emit a gas that can blind the laser optics. However, CD-Rs are constructed much the same way and with mostly the same materials as pressed CDs, and the temperatures required to cause such an emission from the materials that are exposed would
melt much of the car's interior. The dye layer is sealed into the disc, and should not present any danger to drive optics even if overheated.
Even so, leaving a CD-R in a hot car isn't good for the disc, and will probably shorten its useful life.
See also http://www.cd-info.com/CDIC/Technology/CD-R/Media/ Longevity.html,
especially http://www.cd-info.com/CDIC/Industry/news/media-ch ronology.html about some inaccurate reporting in the news media.
See "Do gold CD-R discs have better longevity than green discs?" on http://www.mscience.com/faq53.html.
Ant(Dude) @ Quality Foraged Links (AQFL.net) & The Ant Farm (antfarm.ma.cx / antfarm.home.dhs.org).
AFAICT, reading all the available literature from Mitsui on their gold MAM-A discs, the reflective metallized layer _is not actually gold_. It's aluminum.
Mitsui is claiming their _special dye_ is what makes their MAM-A discs last so long, and the dye is what gives their discs their gold color. Not the metallized layer.
And really when you think about it, it doesnt matter how long-lived the reflective layer is, if your dye deteriorates. Since you're recording your data onto the dye layer -- not the reflective layer.
Umm... what literature would that be? There is a white paper which explicitly states that the reflective later is 24k gold.