Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Why do different pressings of the same CD have different CRC checksums (Read 4453 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Why do different pressings of the same CD have different CRC checksums

Hi there!

I was just wondering why many (non-remastered) CDs have different CRC checksums. For example: I have two copies of The Fragile by Nine Inch Nails (one from Germany, the other copy shows no information about where it's from) and both have different CRC checksums, however the audio is exactly the same since the content comes from the same digital master. I played around with Replay Gain and found that there are very little, neglectable differences, but these happened most likely due to this CRC 'problem'.

Why aren't these discs exactly the same? Do different plants simply produce different products, like different burners have a different offset? Could one assume that one plant produces better discs with less jitter and errors? This is kind of weird to me, because even people at home can produce exact copies, no matter what burner they have. I am aware that audio cds are pressed, but shouldn't it be possible to produce the same output everywhere?

Any information about this would be much appreciated!

Why do different pressings of the same CD have different CRC checksums

Reply #1 –
Well, from what I've learned, yes, different plants make different pressings, with different offsets. I can only assume that there are various devices from various manufacturers that are used in that process, and they make slightly different results in the meaning of where the starting sector is located on the disc.
On the other hand, I don't think that pressed CD will be more or less "jittery" if coming from different plants, at least not that it can be heard, measured - maybe. Errors reading CD will be compensated in the early stages of DA converter, in the CD player buffer, so you won't notice anything... as every medium has some reading errors, some more, some less - but they are corrected.
I've heard about CD player that shows reading error level, but can't remember the model, I think it was Sony, and it was in the 80's... too bad they don't make them any more.

If I misunderstood something, or explained wrong, please correct me.
Error 404; signature server not available.

Why do different pressings of the same CD have different CRC checksums

Reply #2 –
hlloyge, I think the part regarding offset in your reply is correct. Perhaps the jitter & error correction part is not as exactly correct. I hope someone who really knows how the drive HW works can explain it in detail (I don't think I am qualified to do that).

Based on the knowledge I have learned mostly here at HA, I have recently tried to explain the offset variation at the J. River forum a few times. I hope also my replies have been factually correct. Here are two of them:

A quote from: http://yabb.jriver.com/interact/index.php?...91480#msg391480
Quote
The audio stream in a Red Book Audio CD is one continuous track. Because of the limitations in the used technology the placement of this track is not precise. The exact start position is allowed to vary.

It varies already at the manufacturing stage. Different manufacturing plants may produce CDs with slightly different offsets even from the same master recording. Also the CD players have some tolerance in this and each different player mechanism may vary slightly.

For playback purposes the amount of this skew is a non-issue. A correctly manufactured CD doesn't contain the actual recording in the very beginning or the very end of the CD.

In a Red Book Audio CD the individual tracks are created by using a separate index. Even though the start position of the audio stream may vary slightly the individual track durations are always correct. All tracks will have exactly the same amount of shift.

When Andre Wiethoff developed EAC he wanted to create a system that would allow to compare rips that were done on different drives that may have different offsets. He performed empirical tests and determined a reference offset that is used in EAC. There is no official standard for this kind of offset. Since then the "EAC reference" has become a de facto standard and some other programs have started to use it.

A read offset correction system is only useful for comparing rips that were created with different drives to each other. It has no effect to the audio quality. Similarly a combined read and write offset correction system doesn't make possible to burn a more precise audio CD because the offset is allowed to vary. It is only useful for burning an Audio CD that has the same amount of shift as the source Audio CD.*

Here are two links to external threads. I'd recommend reading them before posting further comments.
http://www.digital-inn.de/exact-audio-copy...ay-offsets.html
http://www.hydrogenaudio.org/forums/index....showtopic=50301


...and another one from: http://yabb.jriver.com/interact/index.php?...96082#msg396082
Quote
The Audio CDs have varied offsets (as I explained, even different pressings of the same CD may have different offsets). The drives have varied offsets. There is no correct offset.

Also a set offset correction may cut same samples from the beginning of the first track or from the end of the last track. Only some drives can overread into lead-in/out data.

In a sanely mastered Audio CD the very first and very last samples are silent or contain only inaudible steady background noise because the offset variation is a known factor.

The only "standard" was introduced by the EAC developer for comparison purposes.


*EDIT: It's probably slightly incorrect to speak about "the same amount of shift as the source audio CD has because there is no standardized correct position for the audio track. Though, if I have understood this correctly, the INDEX 01 points on a pressed CD are likely to always be slightly shifted from the positions in the original master data (as they were before the CD was manufactured).

Why do different pressings of the same CD have different CRC checksums

Reply #3 –
EDIT 2

In my above edit, I don't mean only the INDEX 01 points. I mentioned them because INDEX 01 is normally considered to be at the beginning of each track. Naturally all INDEX entries are "shifted" the same amount.

Why do different pressings of the same CD have different CRC checksums

Reply #4 –
Hello,
Actually, the exact start position is precisely defined, and it is 30 samples away from EAC's reference, as said in the digital-inn thread you mentioned above.

It seems that the actual variation in offsets was some kind of manufacturer agreement. I didn't find anything about it in the Red Book.
The Red Book doesn't explicitely says "the first track starts here", but given the encoding method, it would be illogical not to assume that the first byte of the first sample of the first track is the first byte of the first block flagged as track 01 index 01.

Why do different pressings of the same CD have different CRC checksums

Reply #5 –
The "30 samples away..." value is only a drive offset value. Is it possible that there actually is a standard also for the CD manufacturing process and the manufacturing plants are intentionally using offsets that more or less differ from the standard?

(I am still trying to re-read and fully understand the digital-inn thread and especially the linked thread at Club Myce (formerly CD Club Freaks): http://club.myce.com/f52/offsets-handling-...-channel-111913.)

EDIT

A quote from Andre's final post in the digital-inn thread:

Quote
- There is no "correct" read offset, even the second proposed offset is only proposed, there is no real "proof" that it is correct. I just believe it could be correct. Does it make that offset better than the first - I don't think so.

Why do different pressings of the same CD have different CRC checksums

Reply #6 –
Quote from: hlloyge on
they make slightly different results in the meaning of where the starting sector is located on the disc.
The starting sector on the disc is the starting sector on the disc.  The discrepancy is that the data is scewed within the starting sector (and all subsequent sectors).


Quote from: Alex B on
The "30 samples away..." value is only a drive offset value.
If you read what was written by the person who performed the measurement, you will see that this is not correct.

Quote from: Alex B on
A quote from Andre's final post in the digital-inn thread:

Quote
- There is no "correct" read offset, even the second proposed offset is only proposed, there is no real "proof" that it is correct. I just believe it could be correct. Does it make that offset better than the first - I don't think so.
Actually the evidence is quite compelling, especially when you compare the way in which the measurements were conducted.  Andre could have easily obtained a different result had he chosen a different set of CDs.


BTW, I'm extremely happy to see that Pio is still around!

Why do different pressings of the same CD have different CRC checksums

Reply #7 –
Quote from: Alex B on
(I am still trying to re-read and fully understand the digital-inn thread and especially the linked thread at Club Myce


I've just re-read it too.

First, the margin left by the specifications is small. The problem comes from the interleaving. Audio data is not written from first to last on the CD. Bytes from a given block (one block = 6 samples) are spread around a small interval, about one sector wide (one sector = 588 samples). But track numbers are not spread at all. They are associated with the sectors written on the disc. Since these sectors contain audio data coming from different parts of the initial stream, and thus going into different parts of the extracted wav file, the track number does the same, and if we associate each sample in a wave file with the track number from which it comes, we end up with samples from track 1 alterning with samples from track 2. The alternance goes on for about one sector until all samples are eventually flagged "track 2".

Here lies the ambiguity in the specifications : does track 2 begin as soon as we see one byte flagged as track 2 ? Does it begin just after the last byte flagged as track 1 ? Or somewhere in between, where samples from track 1 alternate with samples from track 2 ?

The second answer is the most logical one. The first solution would imply that the drive should retrieve samples from the lead-in, before track 1, in order to fill the beginning of the audio stream. With the second solution, the drive can begin the reading from the first sector of the CD. At the beginning, the de-interleaving leaves holes in the audio stream. This part, where samples that are supposed to come from the lead-in and samples from track 1 alternate, can then be discarded, and the true output begins when there are no more holes in the output stream.

In the discussion at Club Myce, the first hint in this direction comes from post 22 : http://club.myce.com/f52/offsets-handling-...913/#post763683
Reddish, looking at the pits and lands of a CD with special hardware, and comparing with the data outputted by two Plextor drives found that Plextor respects this choice :

Quote
PLEXTOR RULE: the MSB of the first left-channel sample of the time-interval corresponding to a sector is contained in the first frame of that sector. This uniquely determines the Q-channel to audio mapping.


Second hint comes from RichMan, who is "involved with the mastering process at huge CD/DVD factories all over the world", in post 35 : http://club.myce.com/f52/offsets-handling-...html#post845283

Quote
When you decode the Q bit of subcode and find it to be 00:02:00, then rest assured that the first sample of audio that goes with that starting frame (that forms 00:02:00) is the first audio sample for that particular A-Time.


But he is actually unsure about it.

The final evidence is given by IpseDixit in post 55 : http://club.myce.com/f52/offsets-handling-...tml#post1608412

He reports the result of the checking done by Richard Wall, from Doug Carson and Associates, who manufactures encoders for CD pressing plants. His encoder, that drives the Laser Beam Recorder that writes the pits and lands on the master, follows the same rule :

Quote
encoder places main channel's 1st byte aligned into channel frame 00: 02.00


This rule is followed by drives that have a read offset correction of +30 according to EAC.

The only other logical, but less convenient choice, to start as soon as a sample flagged "track 1" is found, is respected by drives with a read offset correction of +601, if my calculus are correct.


@Greynol : thanks