> While this may be possible, there is no guarantee that one CDR of a
> given
> type will perform identically to one that is ostensibly identical
> but
> purchased at a different time/place.

I am seriously interested in the ability to be "reasonably" confident
about the long term archive quality of my optical media. My driving
interest is my collection of over 100,000 digital photos, although
general data integrity is also of interest.
I have concluded that doing your own testing would be not very
difficult. A standardised test at elevated controlled temperature,
and possibly at elevated humidity over a fixed period followed by
error testing. Humidity is messier to handle than temperature control
but still easy enough.

I'd be interested in pursuing this. Is anyone else interested in
participating in an informal project?
What would be useful would be software that allowed easy error
testing. BLER (block error rate) and jitter seem to be the two most
useful although BER can't be bad for you. Mayhaps there is free test
software already available to do this. I'd be happy to look at
temperature and humidity control if others looked at the testing side.
(Standard environmental chamber stuff - but ideally aimed at a level
that could be replicated cheaply and easily by anyone keen). The
problem with BER/BLER testing is that on standard systems you are
usually isolated from the raw data by the drive's error correction
system. Whether this is easily bypassed or equivalent tests done I
don't know. This suggests that some systems allow you to bypass the
error correction http://www.mrichter.com/cdr/primer/errors.htm

Testing need not be against a rigorous absolute standard which would
require standardised drives etc. Comparing a small range of new
CDs/DVDs on a given machine and then the soak tested CD/DVDs should
give a fairly good indication as to how well they had done. In the
NIST report that I cited one brand / batch of CDRs performed far far
better than all others. Whether this is repeatable between batches is
something that the report doesn't tell you and which we could easily
find out ourselves.

If a reasonably fast, cheap & easy test could be achieved then one
could buy a reasonable batch of a selected media (say 100+ quantity)
and then test a small sample for performance. How large a sample was
needed and how indicative this was of batch performance are things
that should be able to be determined quite quickly by testing.

Here's a review which gives some interesting results for new media
from several manufacturers. The selected Sony disks worked superbly at
24X and failed badly at 40X.
http://reviews.cnet.com/4520-3185_7-5020710-3.html?legacy=cnet

Here our new friends at Media Science suggest that any simplistic
analysis won't be useful. AND that formatting errors are involved to a
greater or lesser extent in over 50% of CD failures.
http://www.mscience.com/cdrfail.html



        RM


-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist