Album dublette: Understanding AccurateRip vs. SHASUM

[gef_audio]

need your help for the understanding of accuraterip.

i have a dublette of an album. one of those albums was ripped 2 years back, yesterday i ripped the other one (both to flac files). the ripping was done on different machines with different cd-drives.
then i checked both logfiles: both are accurate, accuraterip crc is identical for each track.
the album i ripped yesterday has a higher confidence, which should be normal as other accuraterips have been done.

based on that, i assume that the flac files are also identical !?.

to crosscheck my assumption it did the following: i extracted the pcm-file for both flac-files (by a script based on ffmpeg).
then i generated the hash-sum with sha for the pcm-files >>> result: they are different !?

WHY?
thanks for your ideas.

[Dat Ei]

Re: Album dublette: Understanding AccurateRip vs. SHASUM

Good morning,

if the accuraterip crc are identical, it is most likely / sure, that the audio data in the two files are identical. With dBpa you can check the integrity of the files. If the two pcm files differ, I would guess that the generated pcm files might include different metadata affecting the sha checksum, or that ffmpeg adds additional data to the pcm files.


Dat Ei

[gef_audio]

Re: Album dublette: Understanding AccurateRip vs. SHASUM

good morning,

thanks for your reply.
hmmm ... i agree, i would also assume that the files are identical.

i do not know dBpa as of now (will check that), but i suppose it should be something similar to my script.
my ffmpeg-script is quite easy: ffmpeg -i <input> -f s16le -acodec pcm_s16le <output>
this removes all the metadata stuff and generates a pcm file, which is pure audio. then i use shasum to generate checksums and compare the files.
i use that for a while and it works for me, this is my integrity check.

by the way: the size of the pcm files is the same, but the hashsum is different ...

any impact by different dbpowerAmp version on mac vs. windows?

[simbun]

Re: Album dublette: Understanding AccurateRip vs. SHASUM

my ffmpeg-script is quite easy: ffmpeg -i <input> -f s16le -acodec pcm_s16le <output>
this removes all the metadata stuff and generates a pcm file, which is pure audio. then i use shasum to generate checksums and compare the files.

Assuming the flac isn't corrupt (check with 'flac.exe -t <input.flac>') then you can pull the audio only md5 directly from the file using:

metaflac.exe --show-md5sum <input.flac>

Or if not using flac the simpler command is:

ffmpeg.exe -i <input.ext> -vn -f md5 - 2>NUL

Are you ripping to a single FLAC file? I know there are samples ignored at the beggining and end for AccurateRip verification, but I can't remember whether that effects the CRC checksum too.

I assume both drives are offset corrected?

[Spoon]

Re: Album dublette: Understanding AccurateRip vs. SHASUM

For the start of last track your drive has to overread and be set to overread for it to match another overreading drive, only plextors in general can do that, so those tracks will be out on a few samples on normal drives.

[gef_audio]

Re: Album dublette: Understanding AccurateRip vs. SHASUM

ah, thanks - that helps and is interesting ... offsets could be the topic ... (md5 is different)

this is what the logfiles show:

---- dublette A -----
Track 1: Ripped LBA 0 to 30597 (6:47) in 1:37. Filename: x.flac
AccurateRip: Accurate (confidence 49) [Pass 1]
CRC32: F02879CC AccurateRip CRC: 320B33E1 [DiscID: 005-00075b10-002304f1-38095805-1]
AccurateRip Verified Confidence 49 [CRCv2 320b33e1], Using Pressing Offset +6
AccurateRip Verified Confidence 10 [CRCv2 6b57e75a], Using Pressing Offset -653
AccurateRip Verified Confidence 4 [CRCv2 156aa1f4], Using Pressing Offset -670


---- dublette B -----
Track 1: Ripped LBA 0 to 30597 (6:47) in 0:47. Filename: y.flac
AccurateRip: Accurate (confidence 46) [Pass 1]
CRC32: B1DD6D97 AccurateRip CRC: 320B33E1 (CRCv2) [DiscID: 005-00075b10-002304f1-38095805-1]
AccurateRip Verified Confidence 46 [CRCv2 320b33e1]
AccurateRip Verified Confidence 24 [CRCv1 e9dcbce5]
AccurateRip Verified Confidence 9 [CRCv2 6b57e75a], Using Pressing Offset -659
AccurateRip Verified Confidence 4 [CRCv1 c63c78be], Using Pressing Offset -659

is it an issue with the offsets? (rip was done with different drives)
and what is relevant here? CRC32 or AccurateRip CRC:320B33E1 ?
is there somewhere a document, which describes the output of the logfile?

thx for your support.

[gef_audio]

Re: Album dublette: Understanding AccurateRip vs. SHASUM

For the start of last track your drive has to overread and be set to overread for it to match another overreading drive, only plextors in general can do that, so those tracks will be out on a few samples on normal drives.

thanks.
is there a setting, where i can adjust that? and please see also my additional reply with the logfiles below.

[simbun]

Re: Album dublette: Understanding AccurateRip vs. SHASUM

For the start of last track your drive has to overread and be set to overread for it to match another overreading drive, only plextors in general can do that, so those tracks will be out on a few samples on normal drives.

So a live or dj mixed disc rip will actually be missing audible samples (not that we'd notice them) at the beginning of the last track? So ripping the same disc on a Plextor (with overread enabled) and another drive without will produce different AccurateRip disc ids?

EDIT: I assume I'm reading too much into "For the start of last track", and this only impacts the end of the last track (not where to start reading) which is why some samples at the end of the disc are ignored for the AccurateRip calculation.

[simbun]

Re: Album dublette: Understanding AccurateRip vs. SHASUM

thanks.
is there a setting, where i can adjust that? and please see also my additional reply with the logfiles below.

Have you gone through the CD ripper setup guide?

It looks like the dublette A rip wasn't offset corrected (if I'm reading the logs correctly). It seems to be off by 6 samples.

[gef_audio]

Re: Album dublette: Understanding AccurateRip vs. SHASUM

FIXED - thanks for all support


my drive is an apple superdrive, where the drive offsets cannot be set, because apple uses different drives.

what i did now is:
- identify the real device in superdrive > use disk util
- check up the offset http://accuraterip.com/driveoffsets.htm

i ripped again and voila the hashes etc. are identical.


thanks again, you made my day

[gef_audio]

Re: Album dublette: Understanding AccurateRip vs. SHASUM

FIXED > thanks for all support.

my drive is an apple superdrive, where the offset was 0.
i identified the real device (disk util) and did a lookup of the cd drive offsets ( http://accuraterip.com/driveoffsets.htm )
then changed the offset and did a rerip: now sha, md5 etc. are identical for both files.

thanks again for your support.

[gef_audio]

Re: Album dublette: Understanding AccurateRip vs. SHASUM

thanks for all support, fixed now :-)


i adjusted the offset (it was an apple superdrive with offset = 0, disk util shows real hardware, where i looked up and adjusted the accurate rip offset)
ripped again, then md5, sha is identical.

thanks again :-)

[Spoon]

Re: Album dublette: Understanding AccurateRip vs. SHASUM

So a live or dj mixed disc rip will actually be missing audible samples (not that we'd notice them) at the beginning of the last track? So ripping the same disc on a Plextor (with overread enabled) and another drive without will produce different AccurateRip disc ids?

EDIT: I assume I'm reading too much into "For the start of last track", and this only impacts the end of the last track (not where to start reading) which is why some samples at the end of the disc are ignored for the AccurateRip calculation.

Accuraterip skips the first and last samples of a disc to overcome this.