Slashdot Mirror
CD-R Lifespan - Is It The Label?
sysadmn writes "Slashdot has discussed archival lifespan of CD-R media before. Fred Langa revisits the issue with a new twist: Are glue-on labels causing premature failure? Much more common than rain forest fungus! From Fred's informative LangaList newsletter."
...but I've stopped doing that. For one thing, humidity causes the clue to come off, leaving you with a bubbly cd that doesn't fit in many slot cd players (esp. car players). Dragging the label off at this point will damage the cd.
Commercial manufacturers silkscreen their CDs, they don't use adhesive labels.
All media degrades. The trick is to use redundant data, and re-copy it before the media is expected to fail.
I upgraded (if you call it that) from a plextor 16x to a sony 52x (really 48x). Compusa had 52x Sony media on sale for 30 bux for 100, figured it was a good deal, and would burn well on the Sony drive. Worst media I ever had, 48/32/16 burns would create coasters, and the media also would fail at 16x on the plextor. The media Did work at 4x-1x, so I used 4x and tossed about 40 bad ones in the garbage.
I also noticed the CD's had a static on them, when stacked, they would stick to each other, not a good coating on them either, and Memorex 48x that I bought was a little thicker, did not have any static on them. Seemed more like pressed CD's.
Not very techinical, I just know that the Sony brand was not top quality, and it felt cheap. So I wouldnt be surprised that some media dies after awhile due to air, etc. And slower burning did work on the media, so I guess burning at 1x on junk media would make it last longer, but who knows how much longer that is.
I just wish there was some technical reviews on media, thickness, quality of materials, burn speed test, tested with multiple media. Now I'm having the same problem with DVDR's, they are now 99 bux for a 4x drive, and half the media I buy only burns at 1x.
Bulk Media for DVDR's seems to also offer the "cheapness" now. I just wish there was standards, or some way of know my data was safe, other than making copies every 2 years. And trying to save a buck per disk doesnt seem to be helping the quality I purchase.
Kodak used to subject their CDs to "torture tests" to see how they'd stand up. Their "Gold Ultima" CDs were reported to have a 100-year archival life. Now, they're saying the same thing about the Ultima brand (now that the Gold brand is discontinued). Study results are here.
:-)
All I can say is that Kodak seems to have done a lot worse to their CDs than your friend did with just his tongue.
D.
here is an easy solution that has worked for me for years. 1. get a perm. marker 2. label your cd-r "applications cd1" or "music cd1" 3. make a text file(html format works the best) list contents of each cd-r 4. done simple as that no wasted time trying to make labels.
All media degrades. The trick is to use redundant data, and re-copy it before the media is expected to fail.
Is there a way to detect when a CD is about to fail? The CD drive will auto-correct minor errors without informing the PC - so by the time a drive returns an error code, at least one block of data is lost (if you're lucky, you can copy everything else off the CD).
I'm aware of commercial testing hardware that can report statistics like the Bit Error Rate (BER), Block Error Rate (BLER), etc. But is there any way to do this cheaply, possibly using software?
I have noticed that on some of my CDR's, burning slower results in a different color than a faster burn. I guess the dye is being changed more when it is slower. So now, for CD's that I want to keep for a few years, I write them at 4x even though I have a 48x burner. As for the labels, I don't use them.
... is the major reason for my failed CDs.
I've only used brand name CD-Rs, my all time favourites are Kodak Golds. Unfortunately they're not making them anymore, because the process was far more expensive than the regular blue or green dye one. But with the Golds I never had a coaster, and all the ones I kept since 4 years ago still work. Other brands I had luck with were Maxwells and TDKs, but they don't look as nice as the Kodaks.
I also keep them protected in some way, either in sleeves, jewel cases or spindles.
The only CDs that ever failed were Hi-Vals. They were the first spindle I ever bought, and even out of those, only one died completely. The paint just peeled off the CD, and it became unreadable. But that CD saw a lot of abuse, sitting in my car for 3 years through very hot, humid summers and frigid winters in Toronto.
Overall, I agree with the article, a bit of prevention will preserve CDs for a long time, probably untill the data on them has become useless.
It's only anecdotal evidence, but the only CDs I have that have lasted more than 19 months or so are my Kodak Golds, the oldest of which were burned sometime in 1995. In particular, I've found any CDs that use a blue dye to be the most fragile/short-lived. However, I do live in a location with fairly high humidity (Brisbane, Oz).
Considering how cheap hard drives are now (about $100 for a 150G disk) it's better to buy multiple hard disks to back things up.
I've decided that if I want to back up my important media for my business, I am going to make a storage machine. Basically it will be a low end machine with a fast network connection and a lot of hard disk space. Each night when I want to back things up I'll send all the files over to the backup machine. To be extra safe I would double the drives in the machine and setup a chron job to copy files from one disk to the other and mirror the drives. That way if one physical disk fails, the files are backed up on another. Also, I'd give it a wireless connection and put it on a very well surge protected power strip. That way lightening is less likely to get to it. Yes, call me paranoid, but I've lost so much data due to media failure that I don't take the risk anymore.
Outdoor digital photography, mostly in New Engl
We need to worry. REALLY. I am burning a cd right now. The data is music. Music from about 25 years ago that was on a reel to reel (older tape analogue) that had to be baked (put in an oven to recover the lost footage more here: http://www.soundsaver.com/squealingtape.htm)
If I am burning this to CD and then the CD becomes faulty, which is likely from my experience, the CD becomes useless, there is no turning back. This has happened to me, a skip becomes more and more of a skip until unplayable.
CD's are not a good way to archive anything.
Yeah, now I know to stay the fuck away from Third World shitholes where the savages will break into boats and toss shiny things into the water. Thanks for the heads up!
I was given some cheap disks that were labelled with some kind of permanent marker. They worked fine then. A couple of years later I was looking for a file that might have been on them and found them largely unreadable. Looking at the actual disk from the bottom I can see the metal layer is completely gone directly below the ink (which has itself faded, or maybe sunk into the disk.
I had a similar experience (or so I thought at the time). About 50% of a spindle of cheap 10x CD-RW's simply failed even after repeated attempts to blank and reburn them. Fortunately, I threw all the CD-RW's that failed into my coaster pile rather than the trash, since about a year later after I had replaced my CD burner I discovered that every single one of them worked fine.
The failing burner was a Yamaha SCSI unit that I paid about $200 for four years ago; it was (and probably still would be) quite reliable with other media, but it was only a 24x8x8x unit, which is why I replaced it. Its replacement, a Sony 48x24x48x IDE unit (rumored to be a rebadged Liteon), cost $50 on sale. It, too, seems to work with everything I throw at it -- including those CD-RW's the Yamaha couldn't deal with.
I suspect that drive/media incompatibility is more common than most people think. Some reports of failing media may be due to media that was marginally recorded to begin with due to such incompatibility.
but i have learnt a few things!
- don't let cdrs touch each other ever. they *will* scratch each others data off. never put written cdrs on spindels or two to a slot in a case. i was too cheap to buy another case and immediately b*gg*r*d some of my discs stickin em two to a slot w/ a grain of sand in between!
- if you have to put a cd down not in a case, put it play side down. i've seen people put them label suide down thinking its 'protecting the disc' - worst thing you can do. a scratch on the plastic side ain't so serious, you could maybe polish it out in a pinch
- overprints are good, if you can get them. these are discs where the original label was printed wrong and has been.. overprinted. = cheaper AND tougher in my experience :)
- writing on a disc with soft tip permanent marker (you dont have to buy "cd markers", thats a joke) has never caused me a problem. i think labels are asking for trouble, but ive always been too cheap to try them :p don't use a hard tipped thing eh in case anyone is stupid to try a biro or sommat
- some cds are just sh*t and fall apart no matter what you do! if data is really vital i'd get two stacks of diff. makes of media and burn two sets (of course, im never sensible enough to do this)
- use the verify funtion of your cdr app if you're not in a hurry. then you know at least the disc started out alright.
there's probably loads more things i'll think of after posting this eh.
the biggest problem i have is with some cds developing mysterious "spots". there is a little speck with a discoloured ring around it, that can fall off, but even if it doesnt the disc is ruined. anyone have any idea what it might be that's touching my discs and doing this to them? i'd love to know, 'cos im stumped. id post a picture of one of these if i was somewhere with a scanner.
oh yeah, did i mention i was cheap! :p maybe buying more expensive blanks would help some of my problems, but i suspect you don't get what you pay for necessarily.
This is my Sig, this is my Gun. One is for Slashdot and one is for Fun.
Parchment can last longer than that if you take proper care of it. You might need to design a special printer to use it. My suggestion would be to have thousands of monks make copies of your data.
My only political goal is to see to it that no political party achieves its goals.
Large batch produced CDs (such as music CDs, AOL CDs etc) are manufactured by stamping the CD pits into a thin layer of aluminium. This is then covered in plastic to protect it on the bottom and a layer of ink on the top. There is no dye layer to degrade with time.
CDRs have a blank die layer into which the CD writer burns the pits. This dye layer is what is causing the problems for long lifetime.
wot no sig
What is an acceptable digital archival media?
All media degrades. The trick is to use redundant data, and re-copy it before the media is expected to fail.
And don't forget to store archival media in an archive. That's right; cool (doesn't have to be freezing, but a consistent not-warm temperature), dark, low humidity.
Non-damp basements spring to mind (as long as you don't have heating equipment in there).
You might consider sealing off CD-R media in an airtight container, such as a (zip-lock) plastic baggy of suitable quality. If it's an opaque one that locks out light as well, all the better.
If you're in the archive (basement) and need light, take care not to hit your archive media with direct light, and take care that UV light is kept to a minimum (i.e. fluorescent lighting is usually better than incandescent lighting).
It also helps to use standard 650MB CD-R media with the ISO 9660 filesystem, in stead of the latest and greatest 818 MB overburned FooFS combination.
Don't burn at 52 speed. Use media that is specced for the lowest speed you can find, and burn it at that speed or lower. You need to drink coffee anyway.
Consider tape backups. Consider a regular archiving/back-up cycle, so that if a piece of media is a dud, you'll have a duplicate from the week before or after.
Store media before use under the same conditions as you would the written archival piece. Don't buy from stores that have humid, warm warehouses.
Index and catalogue your archive. Not only will it make it easier to find things, it will also help to find or sort out duplicates - handy if you switch to a different media and re-archive your old CD-Rs on DVD for example. Archive the index as well.
Use integrity checks, such as checksums. Use compression formats that are easy to fix if a few bits go bad (e.g. no spanning ZIP archives which are useless if the last disk goes missing).
Mix lots of media, unless you're sure that what you've got is a high quality. That way, if wednesday's backup is a dud because of crappy media, the backup from tuesday might at least be better.
Once in a time, randomly check media to see that it's still readable.
Another poster asked about CD-R based RAID; you can just use PAR files (of USENET fame).
SCO employee? Check out the bounty
You are correct on the method of manufacture, but the aluminum layer is still subject to oxidation. If the polycarbonate layer is cracked, split, perforated, or otherwise damaged or defective, moisture can reach the surface and corrode it.
As a reliability engineer, I can tell you that the long-term longevity of manufactured materials is driven (a) by process characteristics (i.e. is there a manufacturing flaw) and (b) thermodynamics. Diffusion processes and chemical reaction rates are all driven by temperature.
If you want your CDs or CD-Rs to last forever, store them below 5 degrees Kelvin or so. You can immerse them safely in liquid helium, it's inert. Make sure it doesn't go superfluid on you, and climb out of the Dewar and away from your 'Britney Unplugged' sessions.
Check out the freeware win32 tool Nero CDSpeed which has many excellent functions for measuring the quality of optical discs (CD, CD-R, CD-Rw, DVD, etc) and tools for verifying the integrity of stored data. Its scandisc function test the drive's ability to read each file at the filesystem level and read each sector at the physical level, telling you which sectors are good, which are failing and which are dead.
An interesting side effect of using this tool is that I've noticed that the manufacture of pressed DVDs is highly variable! Some discs are excellent, some are crap. And it seems to be pretty consistent with the company that distributed them too. Some discs read very smoothly while others require all kinds of speed adjustments by the drive to get data out of them.
Use 10% of your CD space for PAR2 files. Then as long as you have semi-intact files and enought 'blocks' in your PAR2 files to repair them, its like having RAID 5 on a cd. What you could also do is for every 3-5 cds you burn, burn a cd that contains JUST redundancy data, a-la PAR2. This is like raid 3 on cd.
It depends on howmuch you care about your data, but this way, you could lose any one of those 3-5 cds and replace all the data, good as new.
Here is the PAR2 Spec for the many software nerds out there.
And here is Quickpar, a good PAR2 makeing tool.
It's easier to fight for one's principles than to live up to them.