How to deal with copy controlled CDs (and understanding error correction in EAC)?

simbun · December 29, 2020, 11:30pm

Regarding the Jamiroquai rip, track 4 and 11 are inaccurate and can’t be repaired by the CTDB database. We know this because the log doesn’t tell you the alternative rips it has, unlike it did with Art of Noise:

2 | ( 1/617) Accurately ripped, or (545/617) differs in 1135 samples @02:21:61,02:22:60-02:22:61,02:23:06-02:23:07,02:23:17-02:23:18,02:23:28-02:23:29,02:23:49-02:23:50,02:24:06-02:24:07,02:24:38-02:24:39,02:24:49-02:24:50,02:24:70-02:24:71,02:25:06-02:25:07,02:25:16-02:25:17,02:25:48,02:26:27,02:26:69-02:26:70,02:27:48-02:27:49,02:27:58,02:27:69-02:27:70, or (1/617) differs in 1135 samples

The above showed that you could repair the track to match 545 other rips.

The reason in this instance that there weren’t any errors in the log is because it was a burst mode rip. If he ripped it again in secure mode and it still didn’t match with CTDB/AccurateRip then I’m sure they’ll be read errors in the log.

Regarding the “Che Sudaka” CD rip in burst mode, it’s differently inaccurate versus the secure mode rip, but still wrong, and still can’t be repaired.

If you had a new CD and a decent drive (which I think you do) then a burst mode rip should be accurate and-dependant on the drive itself-take less than 5 minutes to rip, I know mine do.

I know you’ve said the CDs are in good order, but the ones you’re having problems with appear to be old, and could be suffering from dye or the reflective surface degradation; this makes it much harder for the laser to read, making it take longer and causing more errors.

Assuming there are a decent number of matches, if CTDB or AccurateRip tells you it’s accurate, then you can be sure it’s accurate, if it tells you it’s not, then it’s not, it’s up to you what to do with that knowledge.

In summary the process should be:
Try a burst mode rip (hopefully not all your CDs are old and this’ll save you a lot of time)
If it’s not accurate and can’t be repaired then try a secure mode rip.
If it’s still not accurate and can’t be repaired then listen to it and save it if you can’t hear any anomalies.

By the way, I bought Kelis Tasty and yes, it’s unreadable, defective by design! (I buy interesting sample discs to test with)

FlavioB · December 29, 2020, 11:48pm

I’ll be replying with quotation later, I first want to ask you guys: is it still OK proceeding in this thread (called “How to deal with copy controlled CDs?”) with something that IMO is drifting OT?
I’m available to migrate over to some more appropriate place, if you mind.

Thanks for letting me know.

… and of course: THANK YOU all in this thread for your precious help, particularly the technical expertise from @simbun

FlavioB · December 30, 2020, 12:04am

@simbun so… Jamiroquai CD: as I did not hear anything wrong while listening to the CD track #4, but it can’t be ripped (not even in burst mode), would I be better buying another copy of it? Or should I first listen to the FLAC files of tracks #4 and #11 first? Or should I re-rip only tracks #4 and #11 in secure mode and check the logs afterwards for any errors?

Che Sudaka: same questions as for Jamiroquai CD. Buy another copy? First listen to FLAC files, or re-rip the defective tracks?

CD status
I can for sure testify that the “Che Sudaka” CD is in very good (near mint) order, as it might’ve been played like less than half a dozen times.
Jamiroquai is of course older - and maybe the quality wasn’t that good?
As of now, I’ve declared “to be thrown in the waste bin” the following (older) CDs:
Depeche Mode - Violator
Pet Shop Boys - Behaviour
Matt Bianco - Another Time, Another Place
All these CDs are “quite old” (I guess the oldest being Depeche Mode) and they’ve been used in a radio broadcast station in the mid-nineties, therefore not super clean. I will buy them again, hoping to have luck and get some good readable copies!

I also got Howard Jones’ “Cross That Line” CD, which has some issues with the last track. With “secure mode” ripping method, it just took more than 1 day and it had not finished. So I aborted and did a burst mode rip - here’s the output for track #10:
Track 10

 Filename C:\Users\Administrator\Desktop\FLAC\Howard Jones\Cross That Line\10 Those Who Move Clouds.wav

 Timing problem 0:03:24
 Timing problem 0:03:26 - 0:03:38
 Timing problem 0:03:40 - 0:05:08

 Peak level 98.3 %
 Extraction speed 3.5 X
 Copy CRC B284BFE2
 Cannot be verified as accurate (confidence 29)  [86662FA6], AccurateRip returned [857BFB48]  (AR v2)
 Copy finished

Listening to the WAV file produced, I cannot hear any issue where the timing problems are mentioned. Do I get it right that this means I’m good to go and the rip is OK?

Ah, so you actually suggest to first do a burst rip and if not repairable (which I can see according to how many matches are suggested/detected, right?), then I should try “secure mode”?
And finally, listen to both the CD itself and the FLAC/WAV file for any audible errors - correct?

FlavioB · December 30, 2020, 12:27am

That’s nice - thanks for confirming!

simbun · December 30, 2020, 11:12am

Burst mode doesn’t attempt any error correction at all, it’s the “least secure”, so saying that it can’t even be ripped in burst mode is not right; of all the modes, burst mode is going to be the mode least likely to get it right, although having said that, if you hover over Burst mode in EAC it states “On some badly scratched CDs, this more could result in better copies than secure mode”, so it has its place. The reason I still use burst mode is because I have CTDB and AccurateRip to tell me if it’s correct.

This would be a perfectly valid thing to do, although if you do rip individual tracks you won’t get any CTDB results in the log as it needs a full rip to perform a verification. You could even attempt multiple burst mode extracts on those tracks, you’d then be performing in a similar way as to what secure mode does, rips multiple times and chooses the result that has occurred most often (or matches with AccurateRip in your case).

FlavioB:

I also got Howard Jones’ “Cross That Line” CD, which has some issues with the last track. With “secure mode” ripping method, it just took more than 1 day and it had not finished. So I aborted and did a burst mode rip - here’s the output for track #10:
Track 10
 Filename C:\Users\Administrator\Desktop\FLAC\Howard Jones\Cross That Line\10 Those Who Move Clouds.wav

 Timing problem 0:03:24
 Timing problem 0:03:26 - 0:03:38
 Timing problem 0:03:40 - 0:05:08

 Peak level 98.3 %
 Extraction speed 3.5 X
 Copy CRC B284BFE2
 Cannot be verified as accurate (confidence 29)  [86662FA6], AccurateRip returned [857BFB48]  (AR v2)
 Copy finished
Listening to the WAV file produced, I cannot hear any issue where the timing problems are mentioned. Do I get it right that this means I’m good to go and the rip is OK?

1 day! “Timing problems” just mean that it took a little longer than normal to read the data, which by itself doesn’t mean that there’s an error, however, these only occur in Burst mode (as secure mode would reattempt) so it’s an indication that something could be wrong, and given that it didn’t verify you know that it’s not accurate.

I’ve just performed a rip on an old CD (1990) I’ve recently bought, Enigma MCMXC a.D. (just because it has multiple index points on various tracks), and in burst mode it took 1:32 with 1 correctable error, and in secure mode (with medium error recovery) it took 1:55, so it goes to show there’s really no hard and fast rule around what’s “best”; it’s massively dependent upon the quality of the CD and the quality of the laser, so you just need to tailor the process to your individual setup.
If most of your discs are old, and you’re getting limited success with Burst mode then maybe it’s best to stick with secure mode (but with at most medium error recovery I would suggest), but personally, if it’s taking more than an hour to read it, and the disc is available cheaply, then I’d just buy another.
If I’m after a particular pressing I usually go with Discogs, otherwise I tend to go with MusicMagpie as they’re cheap and when I have had a problem ripping the disc they’ve sent me another, or if they haven’t got one they’ve refunded it.

If the rip hasn’t been verified it means that either you have a unique pressing (which if there are lots of CTDB/AccurateRip submissions is unlikely) or that the rip isn’t 100% accurate; at this point it depends how OCD you are as to what you do next, listen to it for audible errors, or buy another copy and try again.

Out of interest, how many successful rips have you made so far, given we’re only hearing about the problem ones it paints a bleak picture

simbun · December 30, 2020, 11:25am

I just looked through EAC options and under “EAC options > Extraction” there’s an option to “Skip track extraction after duration longer than ? times realtime”.
If you hover over it, it states “This is handy if error correction takes too long, making error-free extraction nearly unfeasible”.
Given that “EAC” admits that a track this broken is unlikely to be error-free, I’d suggest selecting this to save yourself some time whilst still being able to use secure mode and error recovery (for times when there’s just one or two errors), unless of course there’s no possibility of sourcing another copy.

IvanDobsky · December 30, 2020, 2:59pm

You could tweak the title of the thread to include “and understanding error correction in EAC”.

The thing I see is that errors can literally just be a small maths error. Not actually a musical error. The error could even be on the master and buying another CD may hold the same error.

Only you can answer this. What do you have the music for? If you rip the track to FLAC, and cannot hear any problems, then do you really need another copy just to get a tick in a box of “mathematically perfect”?

simbun · December 30, 2020, 4:56pm

By master I assume you mean the glass master, although this would result in lots more submissions with the same rip as you. If you mean that it could be an error on your individual replication, then I guess it’s possible, but it’s still an error.

One of the problems we have is that we don’t have a scale for how much of a problem an error in EAC is, and that’s what CTDB/AccurateRip were designed for, to give you confidence in your rip. If it matches with those databases then you know it’s right, but if it doesn’t, you still don’t know how bad it is. My understanding is that it must be an error in the audio stream, as it’s the Program Area that’s being extracted during the rip itself (the TOC is read from lead-in, and the pre-gap and index information is read from subchannel data before the actual rip starts I think), but if you know you can’t get a bit perfect copy, then it’s down to your ears (and your OCD) to say whether it’s still acceptable.

FlavioB · January 3, 2021, 2:22pm

I’ve checked the CD (played it on my CD player) and the following track plays without any issues.
Same applies to the FLAC copy of this track.
So, what should I do with the following output?
Track 13

 Filename C:\Users\Administrator\Desktop\FLAC\13 - Kinobe - Summer in the Studio.wav

 Suspicious position 0:02:05 - 0:02:11

 Peak level 86.2 %
 Extraction speed 0.1 X
 Track quality 95.1 %
 Copy CRC 8ECC8D02
 Cannot be verified as accurate (confidence 36)  [FFF9E142], AccurateRip returned [B8D03E85]  (AR v2)
 Copy finished

FlavioB · January 3, 2021, 4:40pm

Here I have an odd one (to me). Listening to the audio CD all is perfect. But listening to the WAV (extracted in burst mode), I got this:
Track 4

 Filename C:\Users\Administrator\Desktop\FLAC\Jamiroquai\Jamiroquai   Return Of The Space Cowboy\04 Light Years.wav

 Peak level 98.9 %
 Extraction speed 15.5 X
 Copy CRC E59265EE
 Cannot be verified as accurate (confidence 200)  [97768B0C], AccurateRip returned [CA401C14]  (AR v2)
 Copy OK

And I do hear a distinct “pop” ad around 2:00…
What would you suggest? Trying a rip of track #4 only, as the audio CD is fine?

simbun · January 4, 2021, 1:23pm

FlavioB:

I’ve checked the CD (played it on my CD player) and the following track plays without any issues.
Same applies to the FLAC copy of this track.
So, what should I do with the following output?
Track 13
 Filename C:\Users\Administrator\Desktop\FLAC\13 - Kinobe - Summer in the Studio.wav

 Suspicious position 0:02:05 - 0:02:11

 Peak level 86.2 %
 Extraction speed 0.1 X
 Track quality 95.1 %
 Copy CRC 8ECC8D02
 Cannot be verified as accurate (confidence 36)  [FFF9E142], AccurateRip returned [B8D03E85]  (AR v2)
 Copy finished

FlavioB:

Here I have an odd one (to me). Listening to the audio CD all is perfect. But listening to the WAV (extracted in burst mode), I got this:
Track 4
 Filename C:\Users\Administrator\Desktop\FLAC\Jamiroquai\Jamiroquai   Return Of The Space Cowboy\04 Light Years.wav

 Peak level 98.9 %
 Extraction speed 15.5 X
 Copy CRC E59265EE
 Cannot be verified as accurate (confidence 200)  [97768B0C], AccurateRip returned [CA401C14]  (AR v2)
 Copy OK
And I do hear a distinct “pop” ad around 2:00…
What would you suggest? Trying a rip of track #4 only, as the audio CD is fine?

Did you follow what I said before?

I should have added to the last point “if you can sleep at night knowing it’s not bit-perfect”

FlavioB · January 4, 2021, 9:57pm

Yes… sorry, I just was trying to “go on”

So the Jamiroquai track:
burst mode reported
Track 4
Peak level 98.9 %
Cannot be verified as accurate (confidence 200) [60B8F8D3], AccurateRip returned [CA401C14] (AR v2)
Copy OK
I did not hear any issue in the resulting FLAC, so I will ignore the verification above and leave it as it is

simbun · January 5, 2021, 9:22am

It looks like you’ve tried burst and secure mode (you said it was going to take ages) and it still isn’t verifying, so as we said before, the last thing you could do (before listening to it) is to extract the problem tracks individually in burst mode (or in secure mode with “Error recovery quality” set to low and “Skip track extraction after duration longer than” enabled) just to see if the drive manages to stumble upon a correct extract.
If the CRC of a subsequent track extract matches what you’ve seen before (in the above instance 60B8F8D3) then the drive is reading the disc consistently and you’re unlikely to get a more accurate rip.

FlavioB · January 5, 2021, 10:53am

Did another burst rip, only for track #4. This is the result:
Track 4
Timing problem 0:00:36
Timing problem 0:00:38 - 0:00:39
Timing problem 0:01:41
Timing problem 0:01:49 - 0:01:51
Timing problem 0:01:53
Timing problem 0:01:59
Timing problem 0:02:01 - 0:02:17
Timing problem 0:02:20
Timing problem 0:02:24 - 0:02:25
Timing problem 0:02:27
Timing problem 0:02:31
Timing problem 0:02:37 - 0:02:38
Timing problem 0:02:44
Timing problem 0:02:47
Timing problem 0:02:55
Timing problem 0:02:59
Timing problem 0:03:03
Timing problem 0:03:06 - 0:03:08
Timing problem 0:03:20 - 0:03:25
Timing problem 0:03:27 - 0:03:34
Timing problem 0:03:36
Timing problem 0:03:38 - 0:03:45
Timing problem 0:03:49 - 0:04:26

     Peak level 98.9 %
     Cannot be verified as accurate (confidence 200)  [22A73A1E], AccurateRip returned [CA401C14]  (AR v2)
     Copy finished
 
 
No tracks could be verified as accurate
You may have a different pressing from the one(s) in the database
 
No errors occurred
 
End of status report

The CRC does not match, so does that mean that my drive is not really “ok”?

simbun · January 5, 2021, 12:19pm

If you’d ripped a clean new disc multiple times and you couldn’t get an exact repeatable extract then yes, you’d have a bad drive.

This track has errors (unfortunately it doesn’t tell us what errors, but the fact they’re not recoverable must mean they’re bad C2 errors), and that means that the samples the drive is extracting are uncorrectable. I guess at the lowest level the drive can’t determine whether it’s reading a 1 or a 0, so on subsequent reads it might read the opposite. This is why extracting it again somtimes leads to different results (as shown by the conflicting CRC’s).
If it were a single fault in isolation then it could correct it, but when you have many problems it can’t.
This isn’t a reflection on your drive, just on your disc.

When playing on a CD player, if the drive encounters errors that can’t be corrected it performs some sort of interpolation with adjacent samples to mask the error so it’s less noticeable, which is why even badly scratched CDs can play “ok”.

FlavioB · January 5, 2021, 12:42pm

Thanks for the explanation.
So in this case, as my ears don’t hear any issues neither from the CD, nor from the FLAC file, I can put my OCD at rest and let it be, correct?

simbun · January 5, 2021, 1:55pm

If you can’t hear the errors, and knowing that it’s not bit-perfect doesn’t keep you awake at night, then tag it and file it away

FlavioB · January 5, 2021, 4:10pm

I just ripped Placebo’s “Once more with feeling” CD in burst mode. EAC reported timing errors on track 1, but no audible problems in the resulting FLAC. This CD is copy protected, so could the timing errors be due to this protection?
The CD itself plays perfectly and no audio errors can be heard.
No track matched, though: is this due to the fact that I’ve used burst mode?

simbun · January 8, 2021, 9:43am

I think we’ve already covered Timing problems in this thead and also copy controlled CD’s (Kelis Tasty).

No, it’s due to the fact that it wasn’t ripped accurately and therefore didn’t match CTDB or AcccurateRip. At this point you’d perform a secure mode rip to see if it performed any better, however, as you know with Copy Controlled CD’s, errors have been inserted into the CD which make it unreadable by CD Roms (at least accurately), so you’re unlikely to get an accurate rip, although maybe one where you can’t hear any problems. Given that this CD does have submissions in CTDB though, I’d guess that it was also subsequently released without copy protection (or there’s a way to defeat it).

I don’t think we’re covering any new ground with this; I think everything you need has already been covered (to excess already ) so I won’t be keeping an eye on this anymore.

Enjoy the rest of the process!

IvanDobsky · January 8, 2021, 10:40am

So, out of the heap of CDs you have ripped, how many are actually playing back badly now? Are you getting a 99% rip success rate? I think you are focusing on numbers too much and forgetting the point of the enjoyment of the music.

Just for larks I have had my ripping PC working on a Road Kill CD for the last two days. By Road Kill I mean a CD that was found in a hedge at the side of the road very scratched. I just want to see how much I can recover from a really bad case. I know I am being evil to EAC…