Page 2 of 2

PostPosted: Mon Apr 14, 2008 3:24 pm
by Halc
Sorry to reply late, but perhaps to the benefit of others....

Our testing of accelerated aging (twice in a humidity test chamber) along with UV exposure to natural UV light for 2 weeks revealed the following:

image corrected
Image
Longer bars mean better quality. This is a summary score for the three aging tests scores.

As you can see, the discs are quite old and there are no guarantees that the discs with same ID codes are still same quality today.

What we did notice however, is that TY 16x was not up to snuff, neither was Verbatim 16x. These were all genuine authentic discs (except for the Gigatain TY fake, which was tested as an example of fake quality).

Now what conclusions can we draw from these for todays' discs?

Very little, I'm afraid.

We do not have exact information on dyes used and factories that made the discs (or even the country of origin for all discs).

Also, manufacturers tweak their formulations and things change a lot in two years.

What is obvious however is that the gold archival MAM-A discs do last long, even if their initial error counts are not as low as for other brands. However, they stay practically unchanged throughout the whole test.

Another thing that is obvious and probably holds for discs of today as well is that 4x is better than 8x is better than 16x.

And I don't necessarily only mean the speed grading, but the actual burn speed. So, it really does pay to wait in this case. 4x does last longer on the average.

What would I buy today?

I might try the Verbatim UltraLife archival grade gold, which seems to be decently priced for an archival grade disc. However, I have NOT tested it, nor seen good stress tests by any third party.

The other alternative is the MAM-A gold archival discs.

All other options (Delkin, Imation, Kodak, etc) are way too overpriced and propably made by MAM-A or Prodisc anyway.

As for BD, I'm not holding my breath. The drives are slow, discs hugely expensive and it'll take quite a long time till these babies take off in volume. I'm expecting 2-4 years minimum. And we have absolutely no idea how they age (never trust mfg data on this, they've been shown to be so misleading/wrong so many times, it's not even funny anymore).

regards,
halc

PostPosted: Mon Apr 14, 2008 5:56 pm
by dolphinius_rex
Sadly Halc, none of this data surprises me. I have faith in Blu-Ray if only because it's NOT based on the DVD standard, which has proven to me time and time again to be heavily flawed.

And yeah, MAM-A Gold Archival media 2 years ago was pretty awesome. Sadly it's not the same disc it was, I would not consider it an option anymore. The Verbatim Ultralife is not a bad choice. Really... there's not a lot out there for most of us anymore. Emtec Gold 4x DVD+Rs are probably the best if you can get them though.

PostPosted: Tue Apr 15, 2008 2:46 pm
by Halc
Thanks Dolphinius.

Have you seen any aging tests for UltraLife?

Also, what can you say about changes for MAM-A gold archival grade DVD? Has it gone down the hill?

I've found very little on that either and not being a subscriber to c't anymore, I'm not even up-to-date on what they have tested.

PostPosted: Sun Apr 20, 2008 8:08 am
by VEFF
Halc,

Thanks a lot.
Even if the test data (which you said is 2 years old, I believe) may perhaps not be applicable to newer discs, it can still be useful to get some idea of disc longevity.


One thing I am trying to understand though is why some discs (e.g. FujiFilm, Maxell, LG etc.) that are mentioned in the section below the graph do not appear in the graph itself?

PostPosted: Tue Apr 22, 2008 2:28 am
by Halc
Why do some discs not appear in the graph?

Because I made an error grabbing the graph off Excel and truncating the graph :)

Thanks for noticing. I've now corrected the error.

PostPosted: Thu May 01, 2008 3:52 pm
by RJW
Suprising results.
But they say so little if you take into account that some of the media is very drive/strategy sensitive. A long with the other production related problems.

For this reason I would prefer to see the individual bars and the used drives. To little time to dig all the info up.

Some interesting stuff.
The RitekF1's use FUJI oxonol dye. A dye which is according to some people in the industry, the most stable out there.
So how come they ended low ? Did Ritek screw up the production or did the drive actually burned them with a to low laser power.
FUJI oxonol dye is known to behave unstable when the laser power is to low.

The difference in speed is clearly seen. Your 8x Mitsubishi azo dye based disc's perform beter as the 16x media. Now I wonder what would happen if you burn the 16x azo disc at 8x, would it perform like the 8x stuff ?

Additional question, did you look for any bonding problems ? Because some results are not in line with other results that I've seen but bonding would explain it.

PostPosted: Tue May 06, 2008 3:27 am
by Halc
Good questions RJW.

Some more info:

Burned drives / discs selected
All of the discs were burned with three different burners (BenQ DW1640, LG GSA-4167B, Nec ND-3551A). Discs were burned full to also stress the outer limit of the discs. All discs were burned at specified speeds. I had a long wrestle about the last one, but I wanted both representative data (people do burn at 'max' speed) and variance (in order to show that lower speed does in fact most often result in a better burn and longevity).

Each three specimens of each disc type were scanned with four different drives.

The best specimen (lowest read error count, lowest read error variance) out of each disc type was selected for accelerated aging in the oven.

This was done to try and limit the "single bad burn" or "incompatible burner" issue, although it cannot of course guarantee total avoidance of either issue.

However, it does reflect real world results, imho, esp. considering what types of drives are/were used here at that time.

Each disc specimen was aged in three separate tests and it was again scanned with four different drives after each aging test to see how it handled the aging process.

I wish we would have had the chance to age three specimens of each disc type, but already the amount of scanning and aging tests with this amount of discs was so horrible that I never want to do this again :)

Ritek F1
Philips DVD-R 16x (Ritek F1) burned at 16x had a not so good starting level of read errors in all drives. Those then doubled during the first aging test and again almost doubled during the second aging test. It handled UV aging fairly well (that is, results didn't change change too much over statistical error margin).

The specimen used for Ritek F1 was a burn with LG GSA-4167B.


16x vs 8x
Based on my sporadic testing all my 16x media have indeed faired better @ 8x burns for initial read compatibility, but I haven't aged them, so cannot I can only guess about longevity.

bonding
We didn't test bonding in any systematic manner. However, trying to pry the layers off discs manually produced no results. I did not bend the discs, but I could of course now... They've been sitting in my storage for a couple of years now :)

What did the test teach me?

1. Don't trust manufacturer data
2. Burn at low speeds
3. Aim for low read error rates/variance in a multitude of drives (different chipset/puh/read strategy optimization)
4. However, aim for stable read error rates (after aging) rather than hyper low starting error rates (which shoot up in aging)
5. Regardless of tolerance, aim for total protection from noxious gases and UV radiation.
6. Other than that, store well, check for errors every now and then and handle carefully