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: Exact Audio Copy - Same CD ripped on 2 drives yields different Copy CRCs (Read 5005 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Exact Audio Copy - Same CD ripped on 2 drives yields different Copy CRCs

I've had this problem occur a few times and thought I would post a description to see if someone could shed some light.

I ripped a CD using EAC with a Plextor drive and according to the log file all tracks matched Accurate Rip and there were no errors.

The same CD was ripped with a Samsung drive and again the log file reported all tracks matched Accurate Rip and there were no errors.

I used a CRC comparison tool and the last track from each rip was found not to match.

The next step was to open the 2 non matching files in a wave editor and a null comparison of the files showed a difference by 6 samples at the very end of the track.

I took a closer look at the log files from each rip and sure enough the Copy CRCs between the Plextor and Samsung were different.

I tried a second Plextor drive and all Accurate Rip results and Copy CRCs were identical to the first Plextor drive.


This happens every once in a while and always on the last track of the CD with the differences at the very end of the track.

The Plextor drives are capable of over reading in lead-in and lead-out but for many years had not set them to do so and found the problem described above occurred more frequently.  After applying the over reading in lead-in and lead-out setting this problem seemed to be greatly reduced for some reason.  If I recall correctly at one time I confirmed that the over reading setting did alter the copy CRC results for the last track for some reason on the Plextor drive.

I thought the problem was eliminated with the over read setting but now it is showing up again.

So why does Accurate Rip report identical values but the Copy CRCs differ and thus the files are indeed different?
Can someone please explain what might be happening?
Which rip is correct?
Does the over reading in lead in and lead out setting or any other setting for that matter have an impact on this?

Thanks,

T

Re: Exact Audio Copy - Same CD ripped on 2 drives yields different Copy CRCs

Reply #1
First, the most important: you should not worry, at least not if you use a fairly recent EAC (I think I recall there was a last-track bug at some pre-beta stage): there should be no music there, and the peak difference in those six samples should be very, very close to zero.

The "over-read" means you get some - possibly non-zero - signal at the very end, that would otherwise be truncated. Unless the CD is very poorly laid out in the mastering, there should be no music there. Think of LP as an analogy (and it isn't too bad an analogy here, in fact): when the last track is done, the stylus would be led to the inner locking groove; now, does it matter whether you lift it the moment it reaches the inner groove, or a split second before or after? Of course you get a split second more or less signal, but the music is over. (If I remember correctly, physical CD players could do the same thing and read a few samples more or less at the very end before they stop.)
AccurateRip was designed to accommodate this, and discards the very end before calculating the checksum.

A longer part of the story is the "offset" - CD rippers do start some samples to the left or right of what they "should" (it is a bit arbitrary because it does not matter). EAC knows this and shifts the bitstream accordingly, but what do you do with the rightmost end if you shift to the left?




If you want to do further testing, then try a different ripping application and bit-compare. You will likely get some differences at the very end, but I bet the peak difference is very close to zero.
- foobar2000 has a somewhat secure ripper, and also http://www.foobar2000.org/components/view/foo_bitcompare can tell you how many samples differ and the peak difference.
- dBpoweramp
- The CUETools ripper


Re: Exact Audio Copy - Same CD ripped on 2 drives yields different Copy CRCs

Reply #2
Porcus, thank you for the reply.  You are spot on about the difference being very close to zero, in the case of the Plextors 6 samples in a row are elevated by 1 increment in the y scale.

I'm using version 1.0 beta 4 and as you say nothing to be worried about but just trying to understand why things are the way they are.

What you say about Accurate Rip discarding the very end makes sense as to why the result is different even though the checksum is the same for both.  So I wonder if something similar takes place at the beginning of track 1, not that I have ever had an issue at the beginning of discs?

EAC is my go to ripper but it's good advice to try another ripper to see if the result corresponds with EAC.

Cheers,

T

Re: Exact Audio Copy - Same CD ripped on 2 drives yields different Copy CRCs

Reply #3
Porcus,

I tried out the same CD ripped with dbPowerAmp.  Interestingly, the results were completely identical to EAC.  The rip from the Plextor drive was always the same regardless of software and the rip from the Samsung drive was also identical regardless of software used.

Cheers,

T

Re: Exact Audio Copy - Same CD ripped on 2 drives yields different Copy CRCs

Reply #4
So I wonder if something similar takes place at the beginning of track 1, not that I have ever had an issue at the beginning of discs?

Kind of, but "even rarer" in reality. Each track starts with the pregap (the index 0), but when you start playing at track N (as opposed to letting track N-1 run to the end), a physical player starts at index 1. Upon ripping to one-file-per-track, one will usually append track N index 0 to the "track N-1" file - except track 1 index 0, which is commonly discarded, as it was supposed to hold silence. In this situation, any differences at the beginning of the CD (a few samples at the beginning of track 1 index 0) would be deleted anyway (and AccurateRip will by design not see it). And pregaps are usually two seconds, which is way enough.

Then there are exceptions, the so-called "hidden track one audio" (HTOA), sometimes referred to as track zero. Then index 0 is longer and contains audio.  On a physical CD player, you would have to start at track 1 (index 1) and then rewind to get it. Some drives are incapable of extracting that part at all, which is effectively the difference you ask for.

Software solutions to HTOA extraction do differ, and I am not sure of the details. I think CUETools will rip track 1 index 0, but the cuesheet will tell the player to start at track 1 index 1. dBpoweramp detects if track 1 index 0 is substantially longer than the usual two seconds, and will offer the user to rip that part to a separate file and tag it with track number 0 and title "Hidden First Track (ability to rip depends on CD drive)". I do not remember what EAC does.

Re: Exact Audio Copy - Same CD ripped on 2 drives yields different Copy CRCs

Reply #5
Porcus,

Fascinating info!  I threw out the question as a curiosity and you provided full details.  Have you had direct involvement with Accurate Rip or the CD manufacturing business?

http://wiki.hydrogenaud.io/index.php?title=HTOA

The link above mentions some EAC methods to extract HTOA.  I have used the range method a few times.

T



Re: Exact Audio Copy - Same CD ripped on 2 drives yields different Copy CRCs

Reply #6
I am not affiliated with anyone in the business, and there are many others here who know this much better than I do.
Back in the day I set up a 200-disc changer to rip my collection (some 7k items) and spent quite some time reading myself up on it. Turns out I still overlooked quite a few things. 

Possibly what annoys me most is that I decoded HDCDs. Assuming you intend to store your CDs in a lossless format (and these days there is hardly any reason not to keep a lossless copy), then do not use a HDCD processor on these - HDCD is flagged in-signal, it can be decoded on-the-fly during playback with foobar2000, it need not be done upon ripping, and it is "potentially unsafe" if stupid things were done in the mastering process.

Also, since you are using EAC: even if you rip to separate tracks, keep cuesheets. I used dBpoweramp back when it did not support cuesheets, and I have on occation missed the index information. (If I were to do the rip job again - not bloody likely! - I would have considered using CUETools instead. It was not around back then.)

Edit: HTOA with EAC: https://hydrogenaud.io/index.php/topic,112937.0.html and the links given therein.

Re: Exact Audio Copy - Same CD ripped on 2 drives yields different Copy CRCs

Reply #7
A 200 disc changer?  Like those old Sony or Pioneer units?

If you don't get it right the first time it's definitely a lot of trouble to go back and redo, especially with 7000 discs!!  Quite the collection!

I'm 1/7th of that and just ripping here and there.  No rush really but it would be nice to get it all done at once.  The benefit of the slow pace means I have altered things along the way but I think at this point all my settings in EAC are alright.

As for HDCD, I read up on it and most of what I found if I recall correctly is to just rip like a normal CD.  I'll have to revisit that one, maybe I'm doing it all wrong, although my collection may have about 20-50 HDCDs.

How many HDCDs are in your collection?  It may not be too daunting to just re-rip those?

As for cuesheets, back when hard drivers were expensive and low capacity, I remember paying $430 for a 6GB hard drive, CD-R was the storage medium of choice.  I would burn without cuesheets as I had no clue.  Nero was one of the first burning software I tried then moved on to CloneCD, Plextools and somehow I ran across EAC and that's when the whole cuesheet importance was understood.  I discovered dbPowerAmp as well but never moved to it because it lacked cuesheet creation (I did not know dbPowerAmp now supports cuesheets creation).

As for CUETools, it's included in EAC for checking rips but again I did not know it was a ripper as well.  You provided me with much new info and I will have to look into CUETools.

Your talk about setting up a 200 disc changer prompted me to have a look on youtube to see what home brew automation for ripping might be out there.  There are some interesting units out there for automated ripping supported by dbPowerAmp.  Some guys even rigged automated systems using components from printers, electronics, and even automating EAC.  Cool stuff!  One over the top solution employed an industrial robot and even plucked CDs, liner notes, and tray cards right out of the jewel cases.

https://www.youtube.com/watch?v=YUccT6rNGVY

The automation is nice but I don't trust the automated tags, often there are typos and formatting inconsistencies.  It would be nice to have a solution out there that could give accurate tagging results with consistent formatting.

Have you heard of Roon at all?

Re: Exact Audio Copy - Same CD ripped on 2 drives yields different Copy CRCs

Reply #8
Never heard of Roon, and that video was news to me. I automated dBpoweramp for a Sony changer.
There is an EAC automation application, REACT. Not sure if the current EAC is supported, check http://wiki.hydrogenaud.io/index.php?title=REACT:Mods

Metadata is an issue yes. dBpoweramp can combine metadata from different sources and "outvote" spelling errors, and that saved me a lot of work. Unfortunately, its main provider Allmusic often gave the wrong year (first year the release was available in the US or something). And it does not support Discogs. Since you have settled with EAC: http://cue.tools/wiki/CTDB_EAC_Plugin 
My impression is that Discogs has the most albums and MusicBrainz the best formatting.

HDCD, you are right: rip as normal CD. I have about a hundred HDCDs still encoded, and it would not be so much job to re-do, except that the collection is packed down in boxes. Whenever I need to move anything around in that room and repack anything, I catch the ones I see and re-rip. No worries over metadata then, I just copy from the ones I already have and rescan ReplayGain.

Re: Exact Audio Copy - Same CD ripped on 2 drives yields different Copy CRCs

Reply #9
Roon is expensive but everyone who uses it pretty much agrees it's the best way to view your music library.  It uses it's own tagging database and pulls data from many places and layers new tags over your tags, thus you can revert back to your own tags at any time.  It integrates with streaming services.  You need one 'server' version to run on the network and other various players can feed from that server.  Many products are now marketed as Roon Ready.  It integrates All Music write ups on artists and bios and discographies and as you browse you can select music to play that you own or from steaming services within that info.  It's been a while since I have read about it so there are likely many new features.  Like I said, it is expensive, but it's almost unanimous that it is the best interface for a music based player.

It does what dbPowerAmp and others do where metadata is concerned, but it does it much better.

You just reminded me of MusicBrainz so I will have to go back for a look.

Well, with 7000 CDs boxed up, hopefully you have a filing system. :^)