marks to mess up a CD mess it up too much?

[dodecahedron]

instead of tossing away a used CD i decided to use it for error testing.

this is an audio CD with 1 track about 4 minutes.
you can see the "band" of burned area on the CD, about 1/4 inch wide.

i made a small mark with a sharpie-like marker across this "band", got 100% damaged in CD Speed ScanDisc. the CD could still be read with explorer, the tracked played fine in winamp.

made a bigger mark, about 1/3 inch wide (so it covers the whole width of the "band") and 3/4 inch long. also made it darker.
now the CD is not recognized by windows at all (doesn't recognize a CD in the drive), neither do InfoTool and CD Speed. some of the times the light of the CD drive doesn't even light up at all - the drive itself doesn't recognize the disc.

i was very surprised!
is this normal?
could the mark have "hidden" (damaged) the TOC part of the CD, or ATIP info, or whatever, so the CD can't be recognized at all by the drive & OS?

[cfitz]

I'd say everything you described sounds normal. Remember that data is interleaved to spread it out over the physical surface of the disc. It is then de-interleaved when it is read back. This distributes the damage from a defective physical area of the disc to multiple logical sectors, preventing a localized defect on the disc from rendering data in a few sectors unreadable at the expense of creating many damaged, but still correctable, sectors.

When you made your first mark, its extent in the direction parallel to the CD-R's spiraling data track was small enough not to completely defeat the error correction of the drive, but the de-interleaving spread the effect of the mark to many logical sectors. Combine that with the fact the mark's extent in the direction perpendicular to spiral covered every turn of the spiral, and you end up with 100% damaged sectors. Still, the error correction (and possibly error concealment) did its job so your disc remained playable.

But the second mark you made was absolutely huge. I don't know the exact figures, but I think defects on the order of 2-3 mm will lead to uncorrectable errors, so when you wiped out 20 mm of every turn of the spiral, you really did a number on that disc. Yes, you wiped out the TOC, the ATIP wobble, the data... everything! There was nothing left for the drive to even recognize that you had inserted a CD and not a bagel.

cfitz

[dodecahedron]

thanks cfitz.
yeah, i know about the de-interleaving and all that.

the first mark was 4mm wide, and the ink stain wasn't very black, i mean it was thin enough to be semi-translucent (if that's the right word?). i had thought that was bad enough, and was sruprised when the ScanDisc showed 100% damaged, i'd thought there would be unreadeable. and even more surprised when explorer and winamp "saw" the track just fine.

the second mark, apart from being bigger, i also "rewrote" it a couple of times to make sure the ink stain was really black (totally non-translucent). this time i expected 100% damaged in ScanDisc, but didnt' think it would mess up the disc so bad that the system didn't even recognize it's existence!

i didn't think this second mark was so huge, as you put it, remembering the pictures of messed CDs in the Kprobe thread.

where is the ATIP/TOC located on the CD? right at the beginning of the spiral?

[cfitz]

semi-translucent (if that's the right word?)

I think "translucent" by itself is the correct usage, but I know what you mean.

i didn't think this second mark was so huge, as you put it, remembering the pictures of messed CDs in the Kprobe thread.

Well, that disc is certainly messed up pretty badly, but the inner area around the hole where the spiral groove starts is clear, and the really big marks with large extent in the direction parallel to the groove (the toughest types of defects) are at the far edge of the disc. It might be that those areas don't show up on the test since the test cuts out before it reaches them. I'm not sure.

I did look up the burst error specifications for CD's, and the CIRC is supposed to be able to correct error bursts up to 3,500 bits (2.4 mm) and compensate for bursts up to 12,000 bits (8.5 mm). So that would agree in general terms with what you did and saw. The 4 mm initial mark resulted in damaged sectors (uncorrectable C2 errors), while the 20 mm second mark completely overwhelmed the drive. Of course, I'm sure that the actual interaction of the drive with your marks is more complicated than this, but what you saw agrees to the first order.

where is the ATIP/TOC located on the CD? right at the beginning of the spiral?

The TOC is in the lead-in, which for a single session is at the beginning of the disc. The ATIP is spread throughout the disc (it isn't actually pits and lands on the disc but rather a wobble in the spiral groove that encodes, via frequency modulation, information such as the absolute position within the groove (hence Absolute Time In Pregroove). The portion that contains the disc identification code that we call the ATIP code is at the beginning of the groove. Thus, lots of important stuff is at the beginning of the disc around the hole, and if you mess up that portion, you can render the entire disc unreadable.

cfitz

[dodecahedron]

OK thanks for the detailed replys, amigo
from now on i won't mess up CDs near the hub, only further outside!

[cfitz]

from now on i won't mess up CDs near the hub, only further outside!

Always be considerate of your CD's when abusing them...

Just to be clear, the real culprit responsible for rendering your CD unreadable was the damaged TOC. (Well, actually the real culprit was you, but you know what I mean. ) The other important stuff at the beginning of the disc including the ATIP disc identifier information, the PMA (program memory area) and the PCA (power calibration area) is important only for burning, isn't required for reading a burned disc, and isn't even visible to a typical CD-ROM reader.

cfitz

[dodecahedron]

Just to be clear, the real culprit responsible for rendering your CD unreadable was the damaged TOC....The other important stuff at the beginning of the disc including the ATIP disc identifier information, the PMA (program memory area) and the PCA (power calibration area) is important only for burning, isn't required for reading a burned disc, and isn't even visible to a typical CD-ROM reader.

yeah, that's what i had suspected.
thanks for clarifying that up, cfitz.