Thanks spath for the explanation!
I've been thinking about the reading accuracy of drives and relative performance of the servo, optics, laser and signal pick up head.
My current working hypothesis is that calibrated drives that read at 1-4x (such as CATS) who often use very high tolerance built Philips / Teac and the like mechanisms, would at least be on the level of $49 USD cdrw mechanism and probably surpass that in most cases.
Of course, I might be totally wrong, but if I am then the results might show this as well.
That is, if I get lots of errors with CATS on a disc that reads consistently the same way (always same CRC in EAC) without errors on a LiteOn and Plextor drive, then I would be inclined to believe that indeed the CATS is not able to read the disc as error free as other drives.
However, my naive faith in precision engineering / hand calibration compared to bulk hi-speed mass manufacturing is so high, that I don't consider the above scenarion very likely.
However, if it happens, I'll report it.
Pio2001 has already calculated some theoretical error figures for a disc that reads error free and consistently with same CRC. The likelihood of such discs (or reads of such discs) containing lots of errors that have gone unnoticed seems to be diminishingly low.
So, if Pio2001's calculations are correct, then I think my above deductions seem reasonably safe assumptions.
The problem with your test is that C1 and C2 errors always depend on the
drive which measures them, since not only the correction strategy, but
about all hardware and software in a drive can contribute to the C1/C2
errors it reports.
Yes, I just read yours and BobHere's descriptions about these and I'm beginning slowly to understand some of these issues.
So, as chosen correction strategies influence how different symbol errors get flagged (and consecutively, how they get classified as c1 or c2 level errors), then even same kinds of read errors (i.e. frames with same erroneous bits in various symbols) will get flagged as different E-errors by different drives (due to their varying strategies).
Now, assuming this and that each drive's detection, correction and flagging strategies are deterministic (leaving out the statistical reading part for now), then I should at least be able to get comparative results for each drive/software combo.
Question remains, is it possible to deduct from these comparative results, if a drive is omitting to report errors (e.g. not reading every frame, not flagging each symbol, etc.)? Or will we just have to accept the fact that differences are most likely (only) due to different CIRC strategies?
I'm not sure. I'll have to think more about this. You bring up such good points, that I'm really grateful for your input, spath!
Therefore there's no absolute reference of C1/C2 errors for a given disc
that you can compare your Kprobe/Plextools results with to know which
one is the most accurate.
Yes, it remains problematic.
It would be interesting to know how AudioDev CATS calculates the E symbol errors. Are they raw symbol errors based on the read data OR are they symbol errors calculated after each level of error correction (with a chosen strategy) has been used?
To be honest, I don't know how CATS calculates E errors, so I'll have to ask more about this from the company that does the testing for me.
If anybody has information on this, I'd like to hear from you.
You can use a CATS for various things (e.g. writing quality testing), but
not to prove that Plextools+Premium is accurate or not.
I agree (based on my current knowledge), but I'm fortunate that I don't need CATS for that.
I've already shown elsewhere that Plextor Premium + PlexTools Pro do not show any C2 or CU errors on disc that remain totally unreadable on that particular Plextor Premium and on any other drive I've tested that same disc on.
I think this shows that on some special cases at least, Plextor Premium and Plextools will not report read errors that happen with that drive doing the testing and on a large portion of other drives.
To me, for all practical purposes, that is a grave measurement error that throws a shadow of doubt over Plextor Premium + Plextools scan results.
I welcome all corrections and comments on the issues, I'm learning more all the time and hopefully will be able to contribute something useful to back to the community.