I ripped a scratched CD using EAC, and got errors.

Topic: I ripped a scratched CD using EAC, and got errors. (Read 6737 times) previous topic - next topic

0 Members and 1 Guest are viewing this topic.

I ripped a scratched CD using EAC, and got errors.

2006-02-24 18:34:06

Hi.

I have a CD that has tiny areas that are practically transparent. I mean that when you hold it against a light you can actually see light passing through. This of course makes certain parts of it (one track) impossible to play a CD player.

I tried to rip the CD using EAC with Secure mode. When it got to the scratched part, it became really slow ( < 0.1%) and started giving Read/Sync errors (I will paste the log file at the end of the message). I waited for about an hour and not much progress... I saw on an other thread here that in these cases it's good to try "Burst mode" and "Test and Copy...". So I did. It finished the job and produced a file that (to me) sounds perfect, but of course there were errors (Timing problems).

What do these errors mean? I'm using really poor speakers, so maybe the quality is actually bad and I can't notice it. Is it possible that I have a perfect file (as far as I care), and that the errors are things that only the computer can notice? And how is it possible to get a WAV file that sounds ok, when there are parts in the CD that are PHYSICALLY missing?

I'm going to try and start another rip in Secure mode before I go to sleep and let it work through the night.

So here is the log:

Code: [Select]

 Used drive  : LITE-ON DVDRW SOHW-1673S   Adapter: 0  ID: 0
Read mode   : Burst
Combined read/write offset correction : 0
Overread into Lead-In and Lead-Out : No

Used output format : D:\Program Files\Monkey's Audio\wapet.exe   (User Defined Encoder)
                     192 kBit/s
                     Additional command line options : %d -t "Artist=%a" -t "Title=%t" -t "Album=%g" -t "Year=%y" -t "Track=%n" -t "Genre=%m" mac.exe %s %d -c2000

Other options      : 
    Fill up missing offset samples with silence : Yes
    Delete leading and trailing silent blocks : No
    Native Win32 interface for Win NT & 2000


Track  8
     Filename D:\Mp3z\Unknown Artist - Track08.wav

     Timing problem 0:02:09 - 0:02:13
     Timing problem 0:02:15 - 0:02:21
     Timing problem 0:05:01
     Timing problem 0:05:43
     Timing problem 0:05:56
     Timing problem 0:06:01 - 0:06:44
     Timing problem 0:08:10 - 0:08:15
     Timing problem 0:08:31 - 0:08:33
     Timing problem 0:08:41
     Timing problem 0:10:37
     Timing problem 0:10:56
     Timing problem 0:11:00 - 0:11:01
     Timing problem 0:11:17 - 0:11:19
     Timing problem 0:11:23 - 0:12:02
     Timing problem 0:12:11
     Timing problem 0:12:13 - 0:12:14
     Timing problem 0:12:17 - 0:12:19
     Timing problem 0:12:24
     Timing problem 0:12:28 - 0:12:30
     Timing problem 0:12:41 - 0:12:44
     Timing problem 0:12:47 - 0:12:49

     Peak level 100.0 %
     Test CRC FEE94D24
     Copy CRC 357A8222
     Copy finished

No errors occured


End of status report

Cheers!!!

I ripped a scratched CD using EAC, and got errors.

Reply #1 – 2006-02-24 18:45:54

It means some samples are blank ones.

In the event where these scratches would be on the plastic side of the CD, you could buy some brasso, and use it to buff the CD. Then, apply some carwax, and make those tiny scratches disappear. A trick while using the brasso is to use a soft cloth, but buff very quickly. The heat will make the plastic softer and should "melt" the cd back so that scratches are no longer visible.

Unfortunately, since you say light passes through, I presume it's the metal coating of the CD that's damaged. It is not possible for you to have a perfect file, but it should be legal to download an ape/cue (or any lossless) or eac/lossy rips of that cd on a peer-to-peer network.

Good luck,
Shade.

I ripped a scratched CD using EAC, and got errors.

Reply #2 – 2006-02-24 19:40:38

Thank you for the reply.

But is it possible that no human being is able to notice these blank samples (maybe you can gather from the log file exactly how bad it is)?
I know I should only care about if I can notice them, but still for general knowledge. And maybe it would mean that downloading a lossless copy (which I couldn't find one with a reasonable number of sources) is unnecessary...

I ripped a scratched CD using EAC, and got errors.

Reply #3 – 2006-02-24 20:14:25

Personally, I find it's pretty bad; and I've downloaded lossless rips (of cds I had) on the ed2k network with one source only (max 4), in approx 8-12 hours -- It's not that bad. Just depends on what client you use. You may want to PM me for details.

You may not notice it, of course; it depends if many samples are consecutive.

Good luck in any case,
Shade.

I ripped a scratched CD using EAC, and got errors.

Reply #4 – 2006-02-24 20:42:42

Quote

Unfortunately, since you say light passes through, I presume it's the metal coating of the CD that's damaged. It is not possible for you to have a perfect file

It is perfectly possible to obtain a perfect rip when the metal coating
has disappeared, that's also what error correction is for. It all depends
on the size of the missing data (the theoretical limit of correctable defects
is around 2.4mm). As for "blank" samples if you mean they are null then
they are very easy to spot in a listening test, so it is almost certainly not
what he got.

I ripped a scratched CD using EAC, and got errors.

Reply #5 – 2006-02-24 21:10:47

Quote

It is perfectly possible to obtain a perfect rip when the metal coating
has disappeared, that's also what error correction is for. It all depends
on the size of the missing data (the theoretical limit of correctable defects
is around 2.4mm). As for "blank" samples if you mean they are null then
they are very easy to spot in a listening test, so it is almost certainly not
what he got.

... and analog sound better than digital, right?

You can't recover lost information. You can't perform exact entropy analysies on a piece of music; You can't spot a single null sample in a file, nor will you be able to if you have 2 or 3 distanced a bit, because they simply don't last long enough for you to notice...

Get your facts straight...

I ripped a scratched CD using EAC, and got errors.

Reply #6 – 2006-02-24 21:23:32

Hahaha, I love getting lectured by newbies

If you don't have the first clue about CIRC and audio extraction,
why do you waste people's time arguing on messageboards ?
First buy a book or read some technical info on RS error
correction, then come back to hold a serious discussion.

I ripped a scratched CD using EAC, and got errors.

Reply #7 – 2006-02-24 21:34:18

Quote

Hahaha, I love getting lectured by newbies

If you don't have the first clue about CIRC and audio extraction,
why do you waste people's time arguing on messageboards ?
First buy a book or read some technical info on RS error
correction, then come back to hold a serious discussion.
[a href="index.php?act=findpost&pid=367036"][{POST_SNAPBACK}][/a]

I'm here to learn. If you noticed misinformation, please tell us what it was and what it should be. Thank you.

Woody

I ripped a scratched CD using EAC, and got errors.

Reply #8 – 2006-02-24 21:42:16

Quote

The CIRC code used on audio CDs can correct burst errors of up to 3500 bits (2.4mm) and can interpolate error bursts of up to 12,000 bits (8.5 mm). Advances in technology in the past 20 years have lead to even more applications for CD technology including DVDs. The error correction on a CD guarantees that high quality music can be enjoyed consistently and reliably.

--2.4mm linear... Which is why it's suggested to clean CDs towards the outside of the disc.

Note that the article says INTERPOLATION : it remains a smart guessing algorithm that makes sure playback isn't interrupted by clicks; It seems CIRC encoding implies a type of parity checking system which is encoded in approx 4% of the CD data. (4 errors on a same frame, max / error record on 109 frames + interleaving). This includes scratches and holes in the Metallic surface, and is therefore a fairly weak protection.

Quote

After interleaving, the data is encoded in a RS(32,28) code. This is also a 2 error correcting code. The data is once again interleaved with a new pattern. Interleaving works well for a CD because 1 frame is spread over 109 frames. Each frame is 32 bytes long. Each frame corrects up to 4 symbol errors. 109 frames can correct 436 errors. With interleaving, 13.625 frames can be corrected, where it would be impossible to do this without the interleaving.

The first code, RS(28,24) is called the C2 level of encoding. It corrects errors due to the physical condition of the CD and the way that the CD was recorded. The RS(32,28) code is the C1 level and it corrects errors due to fingerprints and scratches. This completes the encoding of the audio information, but control information must be added to the CD.

http://web.usna.navy.mil/~wdj/reed-sol.htm for details.

I was wrong after all (somewhat), but in any case, spath has shown himself an ass by not sharing information which could have helped other in any case.

I ripped a scratched CD using EAC, and got errors.

Reply #9 – 2006-02-24 22:02:46

Hmm, how to say that nicely... basic analogies can't make up for
your lack of technical knowledge. You don't know how a drive works
and it's ok, but please don't try to explain others who know how
it works what is possible and what is not. What I wrote is not only
proved by theory but I also verified it during the 10 years I spent
as a drive designer. Missing data can be perfectly recovered (not
interpolated!) until the limits of CIRC error correction.

Looks like you discovered the existence of CIRC so I'll stop there
and let you read further.

I ripped a scratched CD using EAC, and got errors.

Reply #10 – 2006-02-24 22:32:59

Quote

Note that the article says INTERPOLATION : it remains a smart guessing algorithm that makes sure playback isn't interrupted by clicks;

*sigh* It says correction (that is, 100% original data recovery) until 2.4mm and
interpolation (what you call guessing) for longer defects. Which matches perfectly
what I have said in my first reply.

Quote

I was wrong after all (somewhat), but in any case, spath has shown himself an ass by not sharing information which could have helped other in any case.

When I see a question that I know the answer for, I answer. When I don't know
the answer I either give my opinion or I don't reply, because for me stating my
guesses as facts would be misleading and dishonest towards the person who
asked. Not only you presented your misconceptions as facts, but you even
started arguing on them. How did that help the original poster ?

As for the fact that there is actual error correction on audio CDs, I didn't know
that anyone here ignored this fact. It has been mentioned hundreds of times on
this board and I can't honestly see how you have missed this in your 10 months
here. If you want some reading, we maintain a list of technical info at
http://club.cdfreaks.com/showthread.php?t=56662

I ripped a scratched CD using EAC, and got errors.

Reply #11 – 2006-02-24 23:03:43

Thanks very much guys for the help.

I have a question:

Quote

It is perfectly possible to obtain a perfect rip when the metal coating
has disappeared, that's also what error correction is for. It all depends
on the size of the missing data (the theoretical limit of correctable defects
is around 2.4mm).
[a href="index.php?act=findpost&pid=367021"][{POST_SNAPBACK}][/a]

So what you're basically saying is that if my CD meets the 2.4mm criteria and the file sounds ok to me, I probably have a perfect rip, exactly as if my CD was completely unscratched? I it possible a home ripping program will help you "fix" a scratched CD? And if that's what error correction is for, then why couldn't "Secure mode" do the job? Please answer this.

If the next question is too complicated then nevermind: How is this recovery done? Is part of the media written in the "neighborhood" of the actual place where it should be and then some logical tests are done on the disc (XOR etc) when an error is discovered?

Cheers

I ripped a scratched CD using EAC, and got errors.

Reply #12 – 2006-02-24 23:11:33

1. Yes, yes if it reads C1/C2 information, and your ripping drive is a perfect one (plextors are pretty damn good). In this case, EAC reported an error, so an error you have.
2. I believe the parity / error correction is quite complicated, but you can certainly find documents about it on google (CD error correction)

On a side note, not to start a rant, but some people on this forum think it's only good to give information if you know it in the first place... How do these people think information was discovered? If people don't share all their knowledge and ideas, no one will figure out anything; as such, I try to participate in this forum as much as possible, and give my opinion whenever I think it may be useful. If you can also bring something useful (whether by explaining why an aforementionned point of view is wrong or by giving new leads to follow), then all the better. However, if all you seem to be able to do is criticize people on their information, insult them by calling them names (noob) and go on about how people are wrong without explaining in what ways, you are nothing but a hamperer of thoughts. It may do you well in life to have a good carreer, but if you only use it to your own advancement, you benefit nobody.

For a better influence in life, (vote)
Tristan.

I ripped a scratched CD using EAC, and got errors.

Reply #13 – 2006-02-25 00:23:56

Quote

So what you're basically saying is that if my CD meets the 2.4mm criteria and the file sounds ok to me, I probably have a perfect rip, exactly as if my CD was completely unscratched? I it possible a home ripping program will help you "fix" a scratched CD? And if that's what error correction is for, then why couldn't "Secure mode" do the job? Please answer this.
[a href="index.php?act=findpost&pid=367056"][{POST_SNAPBACK}][/a]

If EAC reports errors in secure mode (or not matching CRCs in Test&Copy burst mode), you have no perfect (in the sense of "identical to original") rip. But it may well seem "perfect" to you because you normally don't hear it if only a few samples are wrong. "Secure" mode can try to recover the errors by re-reading the faulty sector until it gets an error-free reading, but if this isn't achieved by the drive, you will have errors in your rip (which may or may not be audible).
In my experience, burst mode often achieves a "good enough" rip with problematic CDs, i.e. one that sounds error-free even with heavily scratched CDs that would give errors and take hours with secure mode. It usually helps if you choose a slow read speed for burst mode extraction.

I ripped a scratched CD using EAC, and got errors.

Reply #14 – 2006-02-25 03:46:41

Well, I'm no expert on all of this stuff, but I do know a bit about how digital sampling works. If you have holes in a cd where you can see light through, that means that data is physically missing. Therefore, unless the so-called error correction on an audio cd is extremely robust, you can never be absolutely sure that you have a perfect copy of the data on the cd, because the original data no longer exists to make the comparison. So the best you can do is interpolate, and maybe perform some sort of statistical analysis on your interpolation procedure to find the probability that your ripped version is exactly the same as the original data was. So it's my point of view, from experience with used cd's and Exact Audio Copy, that discs with pinholes in them can never be fixed. I find this problem to be most common among older cd's or ones which were record-club pressings and have no printed label on them (ie: maybe the title of the disc, and list of tracks is on there, but the top of the cd looks silver and is reflective).

But while we're on this topic, I should mention that some strange things do happen. For example, I have 2 copies of The Eagles Hell Freezes Over, both with tiny pinholes and both are loaded with errors, according to EAC and Plextools CD read test. But I also have Kid Rock Devil Without A Cause, that has a relatively large "hole" in it through which I can see light. If I play that disc on a standalone cd player (in my case, a really cheap one), it gets really messed up and almost stops reading the disc on track 4. But I tested that cd with Plextools, EAC, and Nero, and none of those tests reported any errors at all with the disc. I confirmed that by listening to the rip and watching the waveform in a wave editor. So I don't know what happened there, but I'm glad my Kid Rock cd works. I was not expecting that it would, but it does.

Just my 2 cents.

I ripped a scratched CD using EAC, and got errors.

Reply #15 – 2006-02-25 04:26:27

Quote

...but of course there were errors (Timing problems).

What do these errors mean?

In burst mode there are no software error detection(rereading), but as a small help then Andre implemented the display of "Timming Problems" whenever a read command sent to the drive needed a lot of time to execute, which could have several reasons like the drive loosing sync or trying to recover from a read error. This is not a reliable form of error detection, and should not be taken as an indicatrion of an error has occured, but only as a small hint of that there is the possibility of a sync or read error has occured during the rip.

I ripped a scratched CD using EAC, and got errors.

Reply #16 – 2006-02-25 09:29:21

Quote

So what you're basically saying is that if my CD meets the 2.4mm criteria and the file sounds ok to me, I probably have a perfect rip, exactly as if my CD was completely unscratched?

It's extremely unlikely. 2.4mm is the theoretical limit, in practice it is smaller
due to various lengthening efects. Now if your disc was within this limit, during
your secure mode rip the drive would have corrected the errors without EAC
noticing. And since your burst rips gave different CRCs, your file most probably
contains plain wrong bytes or bytes from he drive's error concealment mechanism.

Quote

Is it possible a home ripping program will help you "fix" a scratched CD?

Hmm, I'm not sure of what you mean here, but if you have a Plextor drive,
Plextools is worth a try.

Quote

And if that's what error correction is for, then why couldn't "Secure mode" do the job?

Secure mode has nothing to do with the error correction mechanism of the drive.

Quote

If the next question is too complicated then nevermind: How is this recovery done?

Next to user data (audio) some additional data is written on the disc, which allows
the drive to detect and correct errors. To understand the algorithm you need some
knowledge of abstract algebra, and in case you're interested you can find a gentle
introduction to CIRC in Pohlmann's CD Handbook.

I ripped a scratched CD using EAC, and got errors.

Reply #17 – 2006-03-15 21:50:29

Just want to bring this topic back again.

I recently ripped a CD's tracks. Tracks 2 to 11 ripped perfectly with Test & Copy + Secure (CRCs weren't always matching if ripped in Burst ). But I have been unable to rip Track 1 perfectly. The CD does not appear to be scratched, but I have noticed several very small spots on different parts of the CD, I bought another copy of the CD to see and there were the same very small (white?) spots around the CD. I tried removing them using a soft cloth but they appear to be stuck there.

It appears the first few tracks (Track 1, 2, 3) have the most problems but I managed to rip Tracks 2 and 3 with Test & Copy (matching CRCs). But Track 1 (both discs) has errors, each time I rip the CD, the Test and Copy CRCs are different every single time! Either the CDs were manufactured with very dodgy machinery, or something because I have never seen these unremovable spots on other CDs before

Is there any way to clean these CDs? Will Brasso do the trick?

I ripped a scratched CD using EAC, and got errors.

Reply #18 – 2006-03-17 16:43:09

I have seen similar spots, except they were dark, IIRC - I had long thrown the discs out. That was the Arabesque - IX+X twofer release by the Russian label CD-Maximum.

The first copy (won on e-Bay dirt cheap) was fine, but a year later started developing those tiny spots that seemed to be under the polycarbonate, on the foil itself. The CD would not rip without tons of clicks; and, cursing myself for not making a backup in time, I started looking for a replacement. Two more copies - one from e-Bay, the other from a GEMM seller - turned out to have the same spots, though in different places. Good thing none of them were too expensive.<g>

I started looking for the original Russian seller from who I bought the first copy. There were four more CDs in that series, all of which play and rip to this day, without any spots. But he was no longer registered. So I turned to the other e-Bayer, who also was Russian, described to him my predicament and offered to pay him double for a replacement, provided he personally examined it first. He was puzzled by my request, but sent a free (!) copy nonetheless (there are good ppl to be found anywhere); and it turned to be without spots and ripped just fine.

To give CD-Maximum due credit, out of around 20 of their CDs I got all look and sound simply excellent, at a fraction of the price of the (often very rare and pricy) original releases. The problem seemed to be limited to a particular batch of that one release.

So if your CD exhibits the same symptoms, don't bother trying to rip or repair it. Neither Brasso nor PlexTools will help. Either look for a replacement, or record your standalone digital out into your soundcard digital in, if that is an option. I found that my cheap Radioshack standalone played those spotty CDs without audible glitches. It does not have a digital out though, so I could not test recording from it.

And most importantly: as soon as you find a CD in your collection that does not produce matching CRCs in Test & Copy Burst mode - back it up to a quality CD-R (like these Tayio Yuden-made Maxell CD-R Pros). You'll save yourself money and hassle later on.

Notice