CD vs: DVD for archival

I have searched but have not seen any late information.

 

I know CD's hold far less. But is there any evidence that DVD's can

have more problems over the years due to higher density of stored

information?

 

Thanks in advance.

Archival qualities depend on the materials the disk is made with and the conditions they are stored in. After doing lots of research I'm using Delkin e-Film Archival Gold DVD-R and CD-R disks and i'm pretty comfortatable with that choice of brand and disk. But these are also backed up on to multiple hard drives as well

I think you'll find DVDs generally the better option over CDs. They seem more stable to

me. I have friends who use cheap spindle CDs, but they're very careful about backing them

up routinely. A DVD may last for decades and decades, but will there be a DVD drive

around that can read them at that point?

With the next generation of optical media already around the corner, DVD need not last very long. Just long enough.

Emre, what's coming around the corner?

NIST actually studied this question! I don't understand all the

details, but it seems like some brands of DVD-R degrade faster than

the best CDR brands. Odd, because all DVD+-R uses the same dye

technology I think, metalized azo. Here is <A HREF="http://www.itl.nist.gov/div895/gipwg/StabilityStudy.pdf">the

NIST publication</A> if anybody wants to help me understand it.

The trouble is that most errors are measured differently on CDR

versus DVDR.

Thank you for the above link. It certainly points out that the type of ink used in the Delkin Gold CD and DVD is superior. It does not indicate that CD might be any better than DVD so I will use Delkin Gold DVD from now on.

 

Thanks again for the link, I'm actually surprised to find a manufacturer claim of superiority and of a superior material (dye type) to be based in apparent fact.

I've given up with optical media as an archival method. HD's are so cheap these days. I'm in the process of transfering my DVD archive back to a HD system - multiple copies at that. I've already had some problems recovering data on DVD's only a couple of years old.

One of the top two disc makers, Mitsui/MAM advertises their archival gold CDs as having a 300 year life span, but they rate their "archival" DVDs at only 100 years. If they have only 1/3 the confidence in their DVDs, I will stick to their CDs. Also, CD players are ubiquitous to a greater degree than the correct version of DVD player.

 

For my own use, I use multiple hard drives, but when I provide a disc to a customer, it's on a MAM gold CD.

Eric: I am referring to HD-DVD/BR, both of which are planned to be offered as recordable media (in contrast to DVD, which was initially read-only).

 

I have a spindle of CD-Rs containing 35mm film scans. At five pictures a disc, all I can say is no thanks.

As a practical matter, CDs are too small for cameras with high MP sensors, even 6MP and up. I use 4G cards with a D2x, which would require 7 CDs just to backup the NEF files.

 

The thermodynamic stability of CDs and DVDs is established well enough that they can be considered archival. The data will likely outlast the technology needed to read them.

 

The more immediate problem is to get a good recording and to protect the discs from physical damage.

 

I've been recording CDs since 1994 and DVDs since 2001, pursuant to an audio/video recording business. In my experience, discs which are hard to read are those which were recorded with too many errors - bad from day one. This depends on the burner, disc materials and possibly the software. The reflective layer on CDRs is easily damaged, and often makes the disc unreadable by ordinary means (forensic software is readily available which will recover data sector by sector). On CDs, the layer is protected only by a thin layer of lacquer. On DVDs, the reflective layer is sandwiched between two polycarbonate discs, each half the thickness of an ordinary CDR. Thus, DVDs are highly resistant to physical damage. Scratches on the reading surface are not good, but can be polished out if they are bad enough to affect reading.

 

It is essential to do a read after write comparison. If this test fails, the disc is bad and must be redone. However, this test only catches about 1 in 10 discs with high error rates. I use Plextor drives and "Plextools" software to check C1/C2 error rates for CDRs, and PIE/POE rates for DVDs. Because of redundant encoding (Reed-Salomon), discs can tolerate correctable error rates up to about 600/second. Anything over 100 seems to be marginal. A good disc burn will have 1 to 6 errors/second. It's not necessary to perform this test on every disk, but a good idea on a batch by batch basis, or if any disc is rejected on a file-compare test.

 

Some discs perform consistently well, and better than their competitors. I have standardized on MAM CDRs and Taiyo Yuden DVD+R and DVD-R discs based on this experience. There may be better discs available, but are only available in boutique quantities.

One thing is apparently DVD-RAM has more error correction that other formats, something worth considering.

The fact is that <B>NO ONE KNOWS</B> and anyone who says they know are lying.<P>

 

Your speculation that CD's might be more reliable because of bigger spot sizes is plausible, but still, it's pure speculation. Likewise speculation that DVD's might be better because they are a technology that was developed with the experience of earlier optical technology.<P>

 

The NIST data is not relevant to the question of archiving because their tests were too short in duration and were done under artificially harsh conditions. The argument that this may accelerate natural degradation is also speculative because different degradation modes have different time scales. (An example might be longevity testing on cars - the NIST test is equivalent to running cars low on oil or under very hot or dusty conditions. Yes, that would test for short-term failure. But it wouldn't tell you anything about which cars are most likely to rust-to-death if they were simply left out for years in the rain.) <P>

 

People have this idea that scientists and engineers have some magical way of doing "accelerated aging" tests on media. But that's not true - accelerated aging tests can only be done for failure modes that you can already anticipate (e.g.'s, UV exposure or oxidation), and those are often <B>NOT</B> the ones that get you in the long run.<P>

 

Bottom line - these discussions about archival longevity of storage media are like a bunch of guys sitting around in a bar B.S.'ing over beers. Fun (maybe) but not a good basis for making any important decisions. Your safest strategy is to do <B>EVERYTHING</B> - HD's (re-copied every few years) CD's, DVD,s hard-copy, and anything else you can get your hands on. Also don't forget to store them in multiple places (e.g., office AND safe-deposit box). Multiple file formats are also a good idea - I store everything in TIFF and Raw, CD + DVD + HD, office and safe-deposit box.

Edward Ingold's writeup might be the best summary of best practices

for optical disc ever published on photo.net!

 

Armando Conti, I don't see why you mention Delkin gold DVD. There is

no information in the NIST article to indicate the brand name or dye

type of D2. Based on Edward's post, it is likely to be Taiyo-Yuden.

 

Peter Nelson, accelerated ageing tests are accepted as reliable by

most scientists, but I also believe in the Easter Bunny.

Plextools also has a Beta/Jitter test. Beta is the ratio between the lengths of "dots" and "lands", and jitter is the short-term variability of the the "dot" spacing. These are functions of the burner quality, and doesn't have to be checked often. If they get off, the read and PIE/POE tests will detect the problem. I've never found any problems in this regard.

 

Peter, while you were killing brain cells with your buddies, discussing whatever over beer, some of us were studying physical chemistry and chemical kinetics;-) The NIST test was SOP all the way, as well as their analysis - those guys are the gold standard for test and measurement.

Your safest strategy is to have FILM as your master.

The NIST was founded in 1901. before the Titanic, Eastland Disaster in Chicago, Firestone 500 tire, or Shuttle(s) failed. A media batch of CDr's or DVD's can have coating problems such that the media is not as well protected. There are other failure modes too. Your roof may leak, a tornado may hit, you might just be a kluz and buy the cheapest non name CDR's, and use them for beer place mats, and the peanut grease will protect them for another 100 years. Having duplicate copies, checking if they are still readable at times always on to move forward a disc that has "issues" and has a compromised date field. If you disc fails you get another free disc! With some older CD's that have gone south, often the small file sizes are ok, but larger ones are gone or line wrapped. Here I also keep smaller copies, copies of modifications, curves that I used, plus some small guys that load real quick, but have at least a VGA size super thumb that doesnt bog Bridge or a Browse program. A real good possiblity is alot of your images will be readable 100 years from now, but there is no "tip of the iceburg" to say /see what is on the disc. Maybe then they will be just be good skeet shooting in 2106.

<I>Peter, while you were killing brain cells with your buddies, discussing whatever over beer, some of us were studying physical chemistry and chemical kinetics;-) The NIST test was SOP all the way, as well as their analysis - those guys are the gold standard for test and measurement,</I><P>

My background is ALSO in science and engineering and whether they are SOP is <B>irrelevant</B>. Accelerated aging tests are based on being able to guess ahead of time what the long-term failure modes are for a given material or fabrication technique. The history of engineering is <B>FILLED</B> with examples of things failing prematurely (and sometimes catastrophically) because the tests that were done did not anticipate the particular failure mode that was actually encountered in the real world. Examples include the de Havilland Comet, and more recently the extensive failures of certain brands of CDs in the 1980's that used a solvent-based lacquer, because no one anticipated that the use of that lacquer would have that consequence.<P>

 

Bottom line is STILL that you can only accelerate for failure modes you can THINK of, and complex, modern technologies always have far more things that can go wrong than anyone can think of. Or, to use the old engineers' adage: "The difference between theory and practice is greater in practice than it is in theory". Also, Murphy's law: if something CAN go wrong, it will."

I mentioned Delkin as they came up most often in the search I did for the type of dye and gold/gold mentioned in the tests. Instead I'll look for the brand Edward mentions. Thanks for the posts.

The name brand "gold archival" discs I have that are failing were burned at a low write speed, with good software, and double checked for readability, and stored in a temp and humidity controlled room away from dust. These 5 to 7 dollar apiece discs did rot away due to poor coatings. I did loose some data. The NIST folks have their government pensions, they live in a bubble and are far away from the reality of disc quality that folks really buy to save their images on. It is wise to keep duplicate copies of valuable work in two locations, using two different brands of discs, to hedge ones bet. <BR><BR>Edward and Peter's abit opposite takes on this matter are both healthy. In my experience accelerated life tests in engineering are very usefull, but often unknown failures really kill off stuff in the field, stuff not dreamed of by experts in their bubble/lab.<BR><BR>One chap I know had his images on theis big raid bank of HDA's, and the name brand power supply fried ALL the HDA's, and none of the data was recoverable, even by the recovery HDA places. Once while at a HDA company I saw a field failure of an entire Server rackmount that got zapped by lightning in Florida. The building had a hardened "lightning proof" design, by a PE that does just lightning stuff.<BR><BR>In the summer home the CD's and DVD's CD's that went under salt water in a few cases were still readable, alot one could look right thru them. One ammo back tub of Cd's was found a block away, and all the CD's are good. In other cases CD's way above the flood line got enough salt vapor to start the rotting process. With color negatives that got salt water dunked, they have this weird horrible smell. With B&W negatives, a salt water dunked negative usually becomes clear. Strangly the non-archival Luminos RC prints from about 1970 are msot all ok that went under salt water, and the archival non RC stuff just a mess with no images left. <BR><BR>In troublesome rotted CD's and DVD's, often the readback is just real marginal, and one can try many different readers. Once we tried this and only 2 out of say 25 readers could pull out the data, a dumb LG CDW/DVD reader from Walmart that cost 30 bucks. This 48X CD reader is on an old Pentium Pro; that has a slow bussmaster ide interface. Maybe the reader working in a slower mode has a far better robustness in reading, since it clearing reads CD's and DVD's that most all my other 20 plus computers wont read.

What if this was tire.net and acme roadrunner 3000 tires are voted to be the best longest mileage low defect tire?<BR><BR> A pro taxi driver might still check the tire pressure, walk around and check for bubbles on the sidewalls, check for nails, check for possible future problems. Being old school and tired of BS whussy marketing chaps claims, his memory of changing a flat in the cold rain with other past tires still lingers.<BR><BR> Just because you buy the gold acme discs doesnt mean one should burn, store and assume they will last a century, Do as a taxi driver does and check your tires and discs, to "catch" potential failures early on. The rare batch of discs that have a poorer compromised coating then can be reread and placed on hopefully better media. Here I keep track of what media was used and what batch, after loosing some data on "archival gold" CD's burned many years ago. This allows one to isolate and recover ones data quicker, since the defects are with one batch that was not up to snuff, that died an early death.<BR><BR>The same thing should be done with slides, fungus just doesnt happen at once, it slowly invades and mucks up the slides surfaces.

I don't know if this is useful to the discussion, but I just had a Maxell DVD-R (supposedly high-quality brand) returned to me which used to play 3 years ago when I made it, and now does not play. I took a photo with a macro lens showing tiny black pinholes all over the disc:

http://www.bealecorner.com/D30/misc/Maxell-DVDR-spots1.jpg

(Normal DVD-R media does not show any such spots.)

 

The DVD had been played a few times and then stored in a case on a bookshelf in a house with typical indoor living conditions. No extreme conditions, no direct sunlight, no dunking in water.

 

I don't know what causes this. BUT if the pinholes are due to corrosion of the reflective layer (silver alloy?) then the 24 kt. gold layer used in "gold archival" DVDs ought to avoid this specific problem. MAM-A advertises such a disc, I haven't tried it though.

http://www.mam-a-store.com/dvd-r-media.html