Understanding CD Doctor results?

[Pilgrim]

I decided to download/install "CD Doctor" simply out of curiosity to see what all the rage was about.

But, not being familiar with the finer aspects of CD R/RW media, etc., I would like to know what is "acceptable" and what is not when it comes to these C1 and C2 errors. I've attached a screenshot of a test result for one of my Sony CD-RW HS discs, formatted with CD_MRW, so comments can be given on an actual test.



Thanks,

Jeff

[dodecahedron]

anyC2 errors are VERY BAD!
a big no-no!

[MediumRare]

Pilgrim- you can save the graph directly as a PNG file (menu save)- no need to go through a capture program. And then the time scale would show too.

Dodecahedron said it already- C2 errors in any amount are bad. You seem to have them interspersed fairly evenly between the C1's. I haven't seen anything like that before. What kind of a drive do you have?

Incidentally, this disk would show damaged/unreadable sectors on a CD-Speed surface scan too!!

G

[Pilgrim]

What kind of a drive do you have?

Lite-on 52246S

Out of more curiosity, I ran Nero "ScanDisc":



And then ran Nero "CD Quality Test":



So, according to Nero's testing utilities, this disc is 100%.

Jeff

[dodecahedron]

very weird...

[CDRecorder]

The MRW formatting could have confused CD Doctor. I have noticed that when I scan a UDF formatted CD-RW with Nero CDSpeed, the last 1/4 or so of the disc is unreadable if I use a CD-ROM as a reader, but the disc shows up as a shorter disc and is all readable if I use a CD-RW drive as the reader.

[cfitz]

Keep in mind that we still don't know exactly what WSES/CD Doctor/K's Probe report as C2 errors. Erik Deppe has said that in most cases his Disc Scan test reports as damaged (yellow) those sectors that have uncorrectable C2 errors (e.g. E32 and up). On the other hand, as time goes on I am suspecting more and more that WSES/CD Doctor/K's Probe report only correctable C2 errors (e.g. E22 and E12) and don't report the uncorrectable errors at all. Hopefully one day we will get all this cleared up. Anyway, if that is the case, then these different charts could be completely consistent.

cfitz

[CDRecorder]

Interesting... So there are both correctable and uncorrectable types of C2 errors?

[cfitz]

Interesting... So there are both correctable and uncorrectable types of C2 errors?

Yes. Read through these for some background information:

http://www.roxio.com/en/support/cdr/cderrors.html
http://www.cdpage.com/dstuff/BobDana296.html#3
http://www.mscience.com/faq13.html
http://www.cdrlabs.com/phpBB/viewtopic. ... 6820#46820
http://www.cdrlabs.com/phpBB/viewtopic.php?t=8194
http://club.cdfreaks.com/showthread.php ... adid=61943

cfitz

[KCK]

I also get weird results for CD-MRW (Mt. Rainier) discs on my LTR-48125W.

For a Platinum 80min/700MB 4x-12x disc, K's Probe 1.1.7 at max speed yielded C1_Max = 33, C1_Avg = 8.083, C2_Max = 165, C2_Avg = 13.053, whereas for a Datex 80min/700MB 10x disc I got C1_Max = 217, C1_Avg = 35.472, C2_Max = 153, C2_Avg = 4.640.

I can't post pictures, but in both cases they were similar to Pilgrim's CD Doctor picture in the sense that the C2 errors were spread quite regularly, as if following some fuzzy pattern.

Both discs passed CD Speed surface scans with no errors (all green).

My cheap disc are of questionable quality, but they have been doing fine for additional temporary backups, as verified by binary file comparisons.

Frankly, right now I don't know if I can learn anything useful by running CD Doctor or K's Probe on CD-MRW discs. Maybe CD-MRW needs special treatment.

[cfitz]

The MRW formatting could have confused CD Doctor.

Frankly, right now I don't know if I can learn anything useful by running CD Doctor or K's Probe on CD-MRW discs. Maybe CD-
MRW needs special treatment.

That is a possibility. Would any of you be willing to take the exact CD-MRW disc, full erase it, burn an ISO-9660 format disc on it, test it, then return it back to CD-MRW and test again, that might tell us something.

cfitz

[MediumRare]

I left this post here instead of putting it in the K's Probe thread because it was inpired by the happenings here.

The MRW formatting could have confused CD Doctor.

Frankly, right now I don't know if I can learn anything useful by running CD Doctor or K's Probe on CD-MRW discs. Maybe CD-
MRW needs special treatment.

That is a possibility. Would any of you be willing to take the exact CD-MRW disc, full erase it, burn an ISO-9660 format disc on it, test it, then return it back to CD-MRW and test again, that might tell us something.

cfitz

I was fascinated by the Pilgrim's strange scans and installed InCD just to try out KProbe on a disk filled with a packet writer. cfitz suggested a series of test to do and here they are for a Verbatim Ultraspeed RW (650 MiB / 74 min):
1. Full erase, Format CD-MRW, fill with drag and drop (not shown)
2. Full erase, write as normal RW (ISO)
3. Full erase, Format UDF (not Mount Rainier), fill up
4. Full erase, Format CD-MRW, fill up

Unfortunately I left the autoscale option on for the first case so I reran this as case 4.
CD-Speed surface scan shows no errors for Case 1/4 and also the Quality test is impeccable (as pilgirm showed).

Case 2. normal RW (ISO)

Case 3. UDF (not Mount Rainier)

Case 4. CD-MRW


The distribution of C2 erors is not random: a count of 8 predominates and values close to multiples thereof are next (I checked this by looking at the csv-files). This may be a systematic property of the UDF format.

Code: Select all

           CD-MRW (Case 4)
No. of C2-errs  Sample Count
           0    2998
           4       1
           5       1
           7       2
           8     117   <----------
           9       1
          11       1
          15       1
          16       2
          17       6
          18       3
          38       7
          39       5

           UDF (Case 3)
No. of C2-errs  Sample Count
           0     3340
           2        1
           7        5
           8       14   <----------
          16        1
          30        1
          34        1
          39        1

I deinstalled InCD after the tests- as Inertia mentioned in his excellent FAQ, it does so cleanly. I ad no problems with InCD- I just don't use my RW's in this way (a USB flash drive is a lot handier as a big floppy ). I do prefer the handling with Mount Rainier, though, mainly because I'm a bit impatient.

G

[cfitz]

Interesting. Thanks for running those tests, MediumRare. They would seem to confirm that the packet writing formats (particularly Mt. Rainier) write something to the disc that KProbe and CD Doctor interpret as C2 errors but that actually are perfectly normal and not cause for concern.

The C2 "errors" are much more regularly spaced for the Mt. Rainier case than the UDF case. I wonder, can you correlate the spikes of C2 "errors" on the UDF disc with the number of files, directories, or drag-and-drop operations you stored/performed?

I left this post here instead of putting it in the K's Probe thread because it was inpired by the happenings here.

I hope we aren't overwhelming Karr. Maybe it is just as well to leave this here for now.

cfitz

[MediumRare]

The C2 "errors" are much more regularly spaced for the Mt. Rainier case than the UDF case. I wonder, can you correlate the spikes of C2 "errors" on the UDF disc with the number of files, directories, or drag-and-drop operations you stored/performed?

There were 12 files in 3 operations, as far as I can remember (not being very systematic in this "experiment"). This may correspond to the 24 non-zero C2-counts, 2 per file. The files were mpeg-segments, most of them about 45 MiB each, 1 small. So the spacing is not correlated with the file size. I guess I could try again and fill the disk in stages, checking after each file.

I'm sure this isn't an arcane secret if the proper data sheets or specifications are available. Maybe someone who knows a bit more can say something here. I'm not an expert on file formats or drive internals, so I tend to rely on "black box" observations and try to find systematic trends or inconsistencies in data presented- it's the (theoretical) physicist in me. That's one reason I looked at the scaling in K's Probe, both the funny spacing on the logarithmic axis and the irregular vaules on the "time" axis.

I hope we aren't overwhelming Karr. Maybe it is just as well to leave this here for now.

I think so too. I have a couple of minor points that I'm saving up for later.

G

[cfitz]

I'm sure this isn't an arcane secret if the proper data sheets or specifications are available.

Agreed. But isn't that true for so much of what we post here?

I tend to rely on "black box" observations and try to find systematic trends or inconsistencies in data presented- it's the (theoretical) physicist in me.

I think so! But then again... isn't empirical observation the job of the experimental physicists? Don't theoretical physicists just sit in their offices and ponder the equations written on the blackboard?

What is your field of study within the broader domain of physics?

cfitz

[MediumRare]

Agreed. But isn't that true for so much of what we post here?

I guess so. It's a kind of game. Maybe it'll be spoiled if we take it too seriously or dig out too many spec sheets.

Don't theoretical physicists just sit in their offices and ponder the equations written on the blackboard?

My all-time favourite Gary Larson cartoon is the one with "Abducted by an alien circus company, Professor xxx is forced to solve calculus equations in the centre ring" (or somethimg like that). Actually we look for deeper meaning and unifying elements in the results shown to us.

What is your field of study within the broader domain of physics?

I studied nuclear physics. Electron scattering at the U. of S. (Sakatoon) and relativistic heavy ion physics in Frankfurt, but that was quite a while ago. I still have a bit to do with physics now, but relativley simple things- my work is mostly programming and writing reports (which I don't like that much).

But if you've done physics, it really influences the way you think. I like to get to the basics, and understand how and why things really work.

[edit]If you gan get hold of the Jan. 1985 issue of Scientific American, there's an article by Walter Greiner and Horst Stöcker entitled "Hot Nuclear Matter" which describes the sort of things I worked on in Frankfurt. Walter Greiner was my Ph.D advisor and my friend Horst has since had a stellar career in academia.[/edit]

G

[dodecahedron]

But if you've done physics, it really influences the way you think. I like to get to the basics, and understand how and why things really work.

yeah!

[KCK]

MediumRare:

Thanks a lot for very interesting tests. As I wrote earlier, I've been happily using crap discs, but now I'll try to get better CD-RW media for testing purposes.

Your picture of Case 4 (CD-MRW) is qualitatively similar to what I got earlier for the Datex 80min/700MB 10x disc, with three differences. My C1 errors are higher (C1_Max = 234, C1_Avg = 34.365, C1_Tot = 105020) and spread more uniformly in space (not suprising for a poor disc). My C2 values are also larger (C2_Max = 146, C2_Avg = 4.486, C2_Tot = 13708). However, I can't see any dominating C2 value (like 8 in your case); in fact thick red bars fall between 10 and 100 on the log scale, whereas thin red spikes almost reach 100 in the first two thirds of the disc, and slightly exceed 100 in the final one third of the disc, where the thick red bars also grow thicker. However, except for this final thickening of red bars, the periodic distribution of red bars in space is quite similar to yours.

In other words, the distribution of C2 values is not as concentrated as in your case, although a fairly narrow range dominates, but the periodic distribution of C2 errors in space is outstanding in both cases.

As for cfitz's querry on the possible correlation of C2 errors with the number of files, directories, or drag & drop operations, my disc has 5032 files (about 400 MiB) in 1188 directories, and it was filled mostly by copying about 20 root directories, with several subsequent file synchronization updates. Of course, any attempt at finding correlation should consider much less complicated cases.

Thus it might be useful to run K's Probe on an empty CD-MRW disc. Don't forget that background formatting may take several minutes (probably 4 or 5 on your drive); you can check via Nero | Recorder | Medium Info.

I'm afraid "black box" observations will have to suffice here. Apparently all CD-MRW specifications are commercial, and drive manufacturers don't disclose anything. Maybe Karr could tame his Probe on CD-MRW discs, but I agree that this should be postponed, since it would be much more useful if he addressed more basic issues first, and we might have already overwhelmed him.

[MediumRare]

Thus it might be useful to run K's Probe on an empty CD-MRW disc. Don't forget that background formatting may take several minutes (probably 4 or 5 on your drive); you can check via Nero | Recorder | Medium Info.

OK I'll see what I can do. I'll do some incremental copying and testing. May be a while before I get around to it. Today is Easter Sunday, and we're expecting company

Happy Easter everybody !!

G

[MediumRare]

KCK- I just reread my post and I mistakenly attributed your excellent FAQ to Inertia. Sorry about that. The tip on using " Nero | Recorder | Medium Info" to determine the formating status was good- thank you for that. Nero only starts on the second try when InCD has mounted a disk, and InCD won't start at all if Nero is still open after erasing the disk. I guess this is a reasonable behaviour.

Now I've done the incremental writes with InCD (Mount Rainier), followed each time by a scan with KProbe (took a while to evaluate it ). There's too much information to reasonably present completely, so I picked 5 stages: the scans with 0,1,2,7 and 11 files. The C2-errors fill the entire scan starting at 8 files.

0 files, freshly formatted

1 file 47.1 MiB

2 files 94.1 MiB

7 files 330 MiB

11 files 500 MiB


This time, the numbers 7 and 15-17 play a big role, as shown in the table:

Code: Select all

No. of files   0           1           2           7          11
Size (MiB)     0        47.1        94.1        330.        500.

Samples     3169        3101        3128        3124        3119
C2 > 0         7          61          74         161         187
C2 = 0      3162        3040        3054        2963        2932

max C2        17          61          40          83          46

C2 = 7         0          19          24          57          61
     8         5           2           2           4           3
    14         0           3           3           4           4
    15         1           8          10          22          25
    16         0           7           7          24          29
    17         1          11          14          26          38
    18         0           3           3          14          14
    24         0           3           4           3           3

I left out a lot of C2-counts that were only represented once in the scan.

The dense cluster of C2-counts at ca. 8 min. shows up after the first file and does not change therafter. The same applies to further non-zero counts: in most (but not all) cases, they stay put after they first show up.

It's interesting to look at a postion with a 7-count for the various scans (12 in total). There are occasional cases where this is missing in one scan, but shows up again in later ones, so with a bit of detective work, it should be possible to localize the LBA involved to several sectors, i.e. much better than one sample (75 sectors). I thought of using this as a clue to the sampling strategy that KProbe uses, but that question has pretty well been settled by Karr Wang's information in the main thread and I won't "break my head" (to translate a saying) with that anymore.

There is an incredible amount of data in these scans (more than would interest most people), so I'll stop for now.

G

[KCK]

MediumRare:

Again, thanks for your hard work. As for the FAQ attribution, I hope Inertia won't mind (I don't!). Concerning the interaction between Nero and InCD, in general one should start Nero before inserting an InCD disc, but sometimes InCD kicks in anyway.

Your pictures are certainly interesting, especially the dynamic fill-in of C2 errors with the growing number of files. They do confirm our earlier observations about the quasi-periodic spread of C2 errors along the space axis, and the occurence of "special" dominating values of C2 errors (now 7 and 15-17).

I guess most of what we are seeing in these pictures is systematic rather than random in the sense that the C2 patterns are due to data structures (links between blocks, etc.), with relatively small random distortions introduced by the usual "physical" causes of C2 errors (if any).

Well, if we could inveigle Karr Wang into looking at these nice pictures...

[Inertia]

MediumRare:

Again, thanks for your hard work. As for the FAQ attribution, I hope Inertia won't mind (I don't!).

On the contrary. What's to mind? (except that I can't take credit for your first class comprehensive InCD troubleshooting guides)

[KCK]

Well, you should certainly take credit for having started my InCD troubleshooting last November!

[alchip80]

MediumRare and KCK

guess most of what we are seeing in these pictures is systematic rather than random in the sense that the C2 patterns are due to data structures (links between blocks, etc.), with relatively small random distortions introduced by the usual "physical" causes of C2 errors (if any).

I remember reading this somewhere on a web site, that the question was brought up

"If you are testing a Track at Once or Packet-written disc, the test results may indicate CU (uncorrectable) errors. These occur because there is Session Information & Link-Blocks contained in these formats. However, the data on the disc is not actually in error. These tests only reports this information as CU error data".

It is possible that this is what we are seeing when it comes to cdrw and packet writings. Maybe there is a way to separate this information from the true C2 errors. I am not sure as this is not my area of expertise.

[KCK]

alchip80:

Thanks for your input. Well, so far we have been discussing possibly "false" C2 errors reported for CD-MRW discs by CD Doctor and KProbe, whereas you are talking about CU. I can't be sure, but I guess both CD Doctor and KProbe distinguish C2 from CU.

BTW, concerning the (un)availability of CD-MRW documentation, I have just located the draft "Multi-Media Command (MMC) Set for the MRW Formats":

ftp://ftp.t10.org/t10/document.03/03-098r0.pdf

Of course, it relies on the initial Philips specification, which is still not publicly available, but it's better than nothing.