Quality of an Original CD compared to Copies...???

[vitualis]

The original audio CD has about 30% overhead in Error Correction Bits.
It can replace the complete loss of about 2000 bits in a row.

Where did that come from? Absolutely not.

Please read what Cornucopia and myself wrote before.

If you read a CD with errors , they will be corrected. (up to some limit)
If you then write that to another CD there is some chance of error
again, but they aren't cumulative.
So there should be no generation loss in copying CDs.

Did you even bother reading what I wrote before? Audio CDs cannot be correctly tracked at the bit level which is what is required for a perfect digital copy assuming that all the bit can be correctly read.

Thus, even assuming EVERY BIT is correctly read, there is still the generational shift in the audio samples.

As far as the fuzzy rectangles go... they are read and converted to digital
data before being written again. The fuzziness is not "copied" , it does
not increase.

Agreed. Unless the "fuzziness" leads to an incorrect bit being read. This will then be copied onto all subsequent generations. Error correction is not 100% and indeed for audio CD that has ONLY C1 and C2, it is FAR from perfect. Which is why we need additional ECC and EDC for CD-ROMs.

Anyone who has made VCDs will know this. Burn enough VCDs even with a good burner and media, and you will still get read errors on sectors on the VCD. VCDs MODE2 FORM2 sectors only contain EDC and no ECC. If these were audio CDs they would have lead to incorrect audio samples compared to the original -- but you would never know unless you did a bit comparison to the original.

As for the "CD copying equipment that the record companies" they don't have copying equipment. Professional CDs are pressed. The whole concept is different. Record companies don't have "media" and they don't do any "copying". CD-R discs are created to "masquerade" as real CDs under the laser of a CD drive. I don't know what it is like now, but when I last looked at the subject, CD mastering companies don't even accept the original data on CD-R so there is no "copying" of any kind.

Regards.

[lordsmurf]

Pressed metal has it's own drawbacks.

[Cornucopia]

Actually, here in the US most replicaton plants will accept CDR's as data masters (although it's a good idea to give them 2 or 3 for comparison, and they probably won't accept CDRW's).

Couple things to add:

Vitualis is correct about the big boys just stamping discs--and this means that they continue to use the same "clean room" glass master to stamp out more discs whenever they want more copies. They certainly don't do CDR copies of CDR's.

Believe it or not, on a microscopic level, burned discs have sharper edge definition than stamped discs. Of course, since discs are coded with a modified NRZ (non-return to zero) code, it's not really the physical pit/land as it is the transition between said pit/land. And anyway, as long as there is a way to infer the transition, however rough and distorted, it'll still be decoded correctly--that's the beauty of digital coding.
Even though, on a microscopic physical level, the disc follows ANALOG methods, it's the digital coding that separates the message from the medium.

Excerpt from Ken Pohlman's CD Handbook:

• Raw Error (before CIRC) 1 error for every 0.1--1.0 million bits
  (~4-40 errors/second)
  After CIRC, ~1 uncorrectable error for every 10-100 billion bits (Audio tracks)
  and
  ~1 uncorrectable error for every 10,000-100,000 trillion bits (Mode 1 tracks)
  Raw (as read from laser) undecoded bitrate at 1x 4.3218 Megabits/sec

Scott

[FOO]

it's more prone to micro-errors after each copy

I think I suggested the opposite.