Pinholes in the metallic layer of a CD no longer contain music, Split from Topic ID #96812 (TOS #5) Options

[jayess]

CD's!!!!!!!!!! Obviously the way to go.

You're absolutely correct, CD's are the way to go, but they aren't the perfect solution either. Most of my collection was bought used and I'm always shopping for new (used) stuff, but it's crazy to see how bad CD's deteriorate. That old stuff from the 80's and 90's, especially the discs with the chrome tops, hold them up to a light source and it's crazy to see how many of them have pinholes, which means there's no longer any music in that hole. And I'm talking about discs that in many cases don't have a scratch on them.

[Porcus]

Greynol, the first step in having bit perfect playback is having all your bits.

Suppose you have a RAID-5 with precisely one disk broken. Do you then have all your bits, or don't you?

[mjb2006]

jayess, I will try to explain, although you really could just read about CIRC elsewhere, as has been suggested multiple times.

The audio samples on the CD aren't written "raw" and sequentially into the pits and lands. If they were, then you'd be right to assume any unreadable spot, be it due to a pinhole or not, would likely be an unrecoverable loss. But in reality, before being turned into pits and lands, the samples are interleaved with other info (timecode, secondary data streams, sync data), so it's not even guaranteed that every pinhole is necessarily "on a sample".

On top of that, and more importantly, all this data is further subjected to CIRC coding, which bloats the data slightly to add some checksum-like codes (to oversimplify a little), and shuffles everything kind of like a deck of cards. The bits for a given sample are thus spread out over a wide area; they're not concentrated in a single spot. The result is that there can be random, occasional tiny errors, as well as completely unreadable bursts of pits and lands, even a straight-up gap of over 2 millimeters, and yet the original info encoded therein can be fully recovered through some virtual sudoku, 100% accurately.

Sure, larger or more frequent holes may exceed the error correction capability of CIRC and result in some undecodable samples. But usually, a significant scratch won't affect a continuous span of samples; rather, it will affect single samples spaced somewhat far apart. In a real CD player, these samples can usually be easily interpolated (e.g. as being halfway between adjacent, undamaged samples) and the result may not be 100% correct, but will be audibly indistinguishable from error-free data. Only when samples can't be interpolated, in the most severe cases, do you get audible glitches during playback. In a CD drive, during DAE (ripping), interpolation isn't normally attempted, so it's quite possible for a lightly damaged disc to play just fine but not rip error-free. "Secure" ripping software can compensate for this, to a degree, with various re-reading strategies.

There's more to know, but hopefully this helps you understand why you shouldn't consider pinholes to be a major threat to the integrity of the music encoded on the disc.

This post has been edited by mjb2006: Sep 9 2012, 06:59

[jayess]

jayess, I will try to explain, although you really could just read about CIRC elsewhere, as has been suggested multiple times.

The audio samples on the CD aren't written "raw" and sequentially into the pits and lands. If they were, then you'd be right to assume any unreadable spot, be it due to a pinhole or not, would likely be an unrecoverable loss. But in reality, before being turned into pits and lands, the samples are interleaved with other info (timecode, secondary data streams, sync data), so it's not even guaranteed that every pinhole is necessarily "on a sample".

On top of that, and more importantly, all this data is further subjected to CIRC coding, which bloats the data slightly to add some checksum-like codes (to oversimplify a little), and shuffles everything kind of like a deck of cards. The bits for a given sample are thus spread out over a wide area; they're not concentrated in a single spot. The result is that there can be random, occasional tiny errors, as well as completely unreadable bursts of pits and lands, even a straight-up gap of over 2 millimeters, and yet the original info encoded therein can be fully recovered through some virtual sudoku, 100% accurately.

Sure, larger or more frequent holes may exceed the error correction capability of CIRC and result in some undecodable samples. But usually, a significant scratch won't affect a continuous span of samples; rather, it will affect single samples spaced somewhat far apart. In a real CD player, these samples can usually be easily interpolated (e.g. as being halfway between adjacent, undamaged samples) and the result may not be 100% correct, but will be audibly indistinguishable from error-free data. Only when samples can't be interpolated, in the most severe cases, do you get audible glitches during playback. In a CD drive, during DAE (ripping), interpolation isn't normally attempted, so it's quite possible for a lightly damaged disc to play just fine but not rip error-free. "Secure" ripping software can compensate for this, to a degree, with various re-reading strategies.

There's more to know, but hopefully this helps you understand why you shouldn't consider pinholes to be a major threat to the integrity of the music encoded on the disc.

so it's quite possible for a lightly damaged disc to play just fine but not rip error-free.

That's what I've been saying all along and it seems to keep drawing disagreement for some reason. If you don't buy used discs with scratches or pinholes, there is no problem ripping them.

There is a couple of good vids on CD error correction:

http://www.youtube.com/watch?v=gHRR-8Q2DHE

http://www.youtube.com/watch?v=gHRR-8Q2DHE

I still standby my point that while a player can correct (quiet) these errors during playback of music, it frequently can't correct the data to make perfect rips that match Accuraterip.