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: XLD offset correction values and log interpretation (Read 7982 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

XLD offset correction values and log interpretation

Hi folks,

I was searching for an answer to this question quite a while now, but couldn't find anything, so I decided to start a topic here:

I am using XLD on my Macbook Pro Retina with an attached Apple USB Superdrive. I normally use the XLD internal ripper.
Since XLD could detect my drives offset automatically (6) I started to rip my CDs to ALAC for storage purpose and AAC for listening.

When I came to Machine Head - The burning red, something strange happened. It was the first disc that the Accurate Rip Summary showed only a v2 confidence (what is this?) and no v1 confidence (and this?). On top of that the overall confidence was 2 out of 225. So I looked at the list of alternate offset values and decided to rip the disc again, with a manually set offset that should have the best confidence. And there it was a v1+v2 rip with 135 out of 225.

So my questions:

What is the offset and how does it affect the quality of my rips?
What is the v1 and v2 stuff?
How is it possible that you can have a good rip with different offsets set?


Below you can see both logs.

OFFSET 6 -> automatically:

Code: [Select]
Machine Head / The Burning Red

Used drive : HL-DT-ST DVDRW  GX30N (revision RP09)
Media type : Pressed CD

Ripper mode             : XLD Secure Ripper
Disable audio cache     : OK for the drive with a cache less than 1375KiB
Make use of C2 pointers : NO
Read offset correction  : 6
Max retry count         : 20
Gap status              : Analyzed, Appended (except HTOA)

TOC of the extracted CD
     Track |   Start  |  Length  | Start sector | End sector 
    ---------------------------------------------------------
        1  | 00:00:00 | 00:53:12 |         0    |     3986   
        2  | 00:53:12 | 04:48:45 |      3987    |    25631   
        3  | 05:41:57 | 04:04:50 |     25632    |    43981   
        4  | 09:46:32 | 04:11:00 |     43982    |    62806   
        5  | 13:57:32 | 03:51:48 |     62807    |    80179   
        6  | 17:49:05 | 03:56:37 |     80180    |    97916   
        7  | 21:45:42 | 04:56:65 |     97917    |   120181   
        8  | 26:42:32 | 03:31:73 |    120182    |   136079   
        9  | 30:14:30 | 04:04:47 |    136080    |   154426   
       10  | 34:19:02 | 03:41:70 |    154427    |   171071   
       11  | 38:00:72 | 05:17:65 |    171072    |   194911   
       12  | 43:18:62 | 06:43:53 |    194912    |   225189   

List of alternate offset correction values
        #  | Absolute | Relative | Confidence 
    ------------------------------------------
        1  |   -639   |   -645   |     135     
        2  |   -637   |   -643   |     58     
        3  |  -1973   |  -1979   |      8     
        4  |  -1309   |  -1315   |      8     
        5  |   -138   |   -144   |      7     
        6  |  -1318   |  -1324   |      4     
        7  |  -1306   |  -1312   |      2     
        8  |  -1967   |  -1973   |      2     
        9  |   -725   |   -731   |      2     

AccurateRip Summary (DiscID: 00141610-00c090d1-8d0bba0c)
    Track 01 : OK (v2, confidence 2/222)
    Track 02 : OK (v2, confidence 2/226)
    Track 03 : OK (v2, confidence 2/227)
    Track 04 : OK (v2, confidence 2/227)
    Track 05 : OK (v2, confidence 2/226)
    Track 06 : OK (v2, confidence 2/225)
    Track 07 : OK (v2, confidence 2/224)
    Track 08 : OK (v2, confidence 2/227)
    Track 09 : OK (v2, confidence 2/225)
    Track 10 : OK (v2, confidence 2/224)
    Track 11 : OK (v2, confidence 2/225)
    Track 12 : OK (v2, confidence 2/219)
        ->All tracks accurately ripped.

All Tracks
    Album gain               : -6.34 dB
    Peak                     : 1.000000
    Statistics
        Read error                           : 0
        Jitter error (maybe fixed)           : 0
        Retry sector count                   : 0
        Damaged sector count                 : 0



OFFSET -639 -> manual:

Code: [Select]
Machine Head / The Burning Red

Used drive : HL-DT-ST DVDRW  GX30N (revision RP09)
Media type : Pressed CD

Ripper mode             : XLD Secure Ripper
Disable audio cache     : OK for the drive with a cache less than 1375KiB
Make use of C2 pointers : NO
Read offset correction  : -639
Max retry count         : 20
Gap status              : Analyzed, Appended (except HTOA)

TOC of the extracted CD
     Track |   Start  |  Length  | Start sector | End sector 
    ---------------------------------------------------------
        1  | 00:00:00 | 00:53:12 |         0    |     3986   
        2  | 00:53:12 | 04:48:45 |      3987    |    25631   
        3  | 05:41:57 | 04:04:50 |     25632    |    43981   
        4  | 09:46:32 | 04:11:00 |     43982    |    62806   
        5  | 13:57:32 | 03:51:48 |     62807    |    80179   
        6  | 17:49:05 | 03:56:37 |     80180    |    97916   
        7  | 21:45:42 | 04:56:65 |     97917    |   120181   
        8  | 26:42:32 | 03:31:73 |    120182    |   136079   
        9  | 30:14:30 | 04:04:47 |    136080    |   154426   
       10  | 34:19:02 | 03:41:70 |    154427    |   171071   
       11  | 38:00:72 | 05:17:65 |    171072    |   194911   
       12  | 43:18:62 | 06:43:53 |    194912    |   225189   

List of alternate offset correction values
        #  | Absolute | Relative | Confidence 
    ------------------------------------------
        1  |   -637   |      2   |     58     
        2  |  -1309   |   -670   |      8     
        3  |  -1973   |  -1334   |      8     
        4  |   -138   |    501   |      7     
        5  |  -1318   |   -679   |      4     
        6  |   -725   |    -86   |      2     
        7  |      6   |    645   |      2     
        8  |  -1306   |   -667   |      2     
        9  |  -1967   |  -1328   |      2     

AccurateRip Summary (DiscID: 00141610-00c090d1-8d0bba0c)
    Track 01 : OK (v1+v2, confidence 133/222)
    Track 02 : OK (v1+v2, confidence 134/226)
    Track 03 : OK (v1+v2, confidence 135/227)
    Track 04 : OK (v1+v2, confidence 134/227)
    Track 05 : OK (v1+v2, confidence 133/226)
    Track 06 : OK (v1+v2, confidence 133/225)
    Track 07 : OK (v1+v2, confidence 132/224)
    Track 08 : OK (v1+v2, confidence 134/227)
    Track 09 : OK (v1+v2, confidence 134/225)
    Track 10 : OK (v1+v2, confidence 132/224)
    Track 11 : OK (v1+v2, confidence 133/225)
    Track 12 : OK (v1+v2, confidence 127/219)
        ->All tracks accurately ripped.

All Tracks
    Album gain               : -6.34 dB
    Peak                     : 1.000000
    Statistics
        Read error                           : 0
        Jitter error (maybe fixed)           : 0
        Retry sector count                   : 0
        Damaged sector count                 : 0

XLD offset correction values and log interpretation

Reply #1 –
http://wiki.hydrogenaud.io/index.php?title=AccurateRip
Should answer all 3 questions.
Edit: Almost, I'll add
AccurateRip version 2 (v2) refers to the newer improved algorithm
AccurateRip version 1 (v1) refers to the older original algorithm
korth

XLD offset correction values and log interpretation

Reply #2 –
>What is the offset and how does it affect the quality of my rips?

They are different pressings of the same disc, these pressings have an inherent offset built in

>How is it possible that you can have a good rip with different offsets set?

The program can detect the offset, then correct for it when calculating the AccurateRip match.

XLD offset correction values and log interpretation

Reply #3 –
Ok i read the link above a few times now thought about it and there are a few more questions now. I'll try to answer my questions by myself but I'm not sure.


So there a basically 2 offsets my software has to take care of. The offset from the disc or more the pressing of it and the offset the drive has that I am using.
The offset from the drive is known by the software so it adjusts it right in the preferences.

So now I rip a CD and get - according to the log-file - a good rip but bad confidence (3 out of 500 for instance). This can only be due to the fact that 497 people had a disc submitted to the database that is from another pressing and just 3 more had submitted one from the same pressing. Correct?

But then again, how can it be that if I manually set my offset to the most confidential one in the log-file, rip my disc again and get a 497/500 confidence, as I completely ignored my own drives offset here and set it to something that should take care of the discs pressings offset?


I understand that the offset can be detected and change by XLD but I warn the software that it changes the offset without giving a notice in the log. I don't think it'll do that.


And something else:
Do I really need to care about offsets? Because it only affects the song in the very beginning and the very end... and since the possible loss is a fraction of a second (which is at these positions silence for 99% off all songs), why should I care?

XLD offset correction values and log interpretation

Reply #4 –
A good rip but bad confidence? There's really no such thing. A match is a match. Don't worry so much about the numbers

Knowing about offsets or other pressings can be useful info when you don't have a match but the software needs to have the proper settings when using AccurateRip.
korth
 

XLD offset correction values and log interpretation

Reply #5 –
The only offset you  can actually "take care of" is the drive offset, which exists because  there's a well-defined correct behavior that most drives deviate from: when you  ask for Block X on the disc you should get back the samples contained in Block  X on the disc, and only the samples  contained in Block X on the disc. If every time you ask a drive for Block X it  skips the first twenty samples from that block but tacks on the first twenty  from Block X+1, that drive has an offset of 20 samples. So reads using that  drive will always return the wrong data, but because it's always wrong in the same way you can derive the right data from  it: just read the whole disc and shift everything over by the number of samples  in the offset. (Overrun and underrun come into play here but let's just assume  the most common case, which is that--as you noted--most discs start and end  with a number of zero samples, so you usually won't lose any meaningful data  when you do the shift.)

For the discs themselves, there's no known "correct" position for the audio data on a given title, so you can't compare that "correct" position to the position of another disc and calculate an offset value. If two  pressings* of a disc have sample-identical audio content  but the first nonzero sample on Version A is twenty samples later than the  first nonzero sample on Version B, they're both equally "correct",  and there's no way to "take care of" the difference between the two.  It's like having two editions of a book with the exact same text but different  pagination; one is not a more "correct" pagination than the other,  they're just different.

Finally, the "X of  Y" values in an AccurateRip report aren't really "confidence"  values, and frankly it's misleading for apps like EAC and XLD to be labeling  them that way, because it implies that you should be less "confident" in the accuracy of a 3/62 match than a 6/62 match. In fact the real-world chances of two  rips made by different people on different drives being both bit-identical and wrong is vanishingly small, so even a 1/62 match is very very likely to indicate an accurate rip.

*Physical masterings,  really, not pressings; different pressings from the same physical mastering are  bit-identical by definition.