Need Help With Scratched CD - I Have Multiple Copies of the Same CD Options

[hurryup]

Hi,
I have multiple copies of the same audio cd ripped, each of them however are scratched to a certain degree, and ripping each cd individually causes some songs to have minor errors even in secure ripping, but since I have multiple copies of the CD, is there any program that error recovers by ripping/comparing multiple copies of the same CD? Thanks

This post has been edited by greynol: Dec 7 2009, 05:23

Reason for edit: Removed post icon.

[psycho]

This might not be a solution you're looking for, but...

Use some ripping app, that supports AccurateRip and rip all of the CDs with it. Hopefully you will get all of the tracks accurately ripped when put together. Offcourse, this will only work if the copies were made with correct read and write offsets!

Hopefully, someone will come up with a better solition for you.

[sauvage78]

If the pressing is the same (same CRC on remaining accurate tracks with fixed offset), the solution of psycho works, I used it myself in the past.

This post has been edited by sauvage78: Dec 6 2009, 14:20

[ophitoxaemia]

This solution does not work if the same track has errors in different parts, which is what the OP was asking about.

You need to read more wikis sauvage.

[sauvage78]

ophitoxaemia:

You need to re-read the original post before poping up & telling that psycho & me are making a misstake: the original post doesn't tell if the same track has error on both CD, that's why psycho said:

"This might not be a solution you're looking for, but..." in front of his post.

You introduced yourself the possibility that the same track would have a scratch on both CD & you didn't even mention that it was on both CD in your post N°#4.

You'd better learn reading before posting Mr Know-It-All.
Now plz leave me alone.

This post has been edited by sauvage78: Dec 6 2009, 15:22

[Akkurat]

Hurryup, you could try my "rev speed lowering trick" if you're using EAC (which you should with AccurateRip). Intro here, and instructions here.

[hurryup]

Hey guys, ophitoxaemia was right to a certain degree - the same few songs on each CD has errors, just on different parts of each song. Also, the CD that I am ripping is not in accuraterip. So what to do?

[sauvage78]

hurryup:
So ... I may be wrong but I think that you cannot do anything. Technically it would be possible but it would require combining frames from different rips & AFAIK no ripper implement such a complex feature for the simple reason that your case is very rare.

The other possibility would be to try to fix it after the rip by splitting the two bad songs in valid part at the same frame timings with an audio editor & then try combining these valid parts ... but I never heard of anyone trying to create such a monster file.

You would need to know at which frame exactly the damage occur to leave out bad frames & even then I am not sure if it would work (you would need the right lossless & frame accurate editor for the job) & without a valid accurate result in the database to check against you wouldn't even know if the end result would be valid ... a lot of trouble for a very uncertain gain IMHO ... sorry ...

[kritip]

buy another copy

[twostar]

You could try ripping multiple times with different rippers and their different rip modes: EAC with burst, secure C2, secure no C2, foobar2000 with standard and paranoid and Dbpoweramp. If any of the rips matches twice, that would be your best bet.

[skeptic]

Why not fix the scratches? I personally use Brasso Wadding. It seems to work for all but the deepest. Use at your own risk of course.

This post has been edited by skeptic: Dec 7 2009, 08:43

[odyssey]

If you know which parts of the tracks are bad and are willing to use a lot of time on it, rip both and slice them together using an audio editor - be sure to merge at exact bits. After that you can verify the tracks with Accuraterip using CUEtools or foobar2000.

[sauvage78]

Personnaly, I had some success using the teeth pasta trick on lightly scratched CD that I couldn't get accurate, but it only worked on light scratches, for deep scratches there is no miracle.

With some luck one of the bad CD have "fixable" scratches.

Personnaly, I still dunno for sure if the success was due to teeth pasta or just to re-trial. This is a very random method anyway. (Burst or low secure) + AR2 works best with this method. (If you don't want to stress your drive & waste your time for nothing ...)
But even with this method you would need an accurate confidence of 2 to check against ...

So whatever you try, in your situation an AR2 result to check against would greatly help, so you'd better wait for the CD to be in the AR database.

This post has been edited by sauvage78: Dec 7 2009, 10:33

[flacflac]

A year ago or so I had the same problem, and I was able to use EAC in sector mode and just rip 0-x with CD-1 and x+1-y with CD-2 - then I merged the two files (with Audacity, if I remember correctly) and burned a new disc with these then-complete files. The subsequent rips were accurate in the end.

ff

[2Bdecided]

Brasso has always worked* for me (on discs attacked by toddlers!) as long as the scratches aren't on the label side. Polish radially only. It'll make your CD look far worse, but play far better.

*worked = Accuraterip verified after polishing.


If you have three CDs with errors in difference places, then (as long as the rips line-up) a simple bit of code to "majority vote" each individual audio sample should do it.

Trivial (but slow!) in BASIC, MATLAB, etc. Actually, it might not even be that slow in MATLAB - put the three separate versions into the same vector and median filter it across the appropriate dimension. Make sure you handle the two stereo channels separately.

Cheers,
David.

[maggior]

Brasso has always worked* for me (on discs attacked by toddlers!) as long as the scratches aren't on the label side. Polish radially only. It'll make your CD look far worse, but play far better.

*worked = Accuraterip verified after polishing.

I've used Brasso to clean up badly scratched discs from the library. I've also used toothpaste. Toothpaste is a more coarse polishing compound than Brasso. For badly scratched discs, I've used toothpaste first to work out the deeper scratches (at this point, disc smells minty fresh!). Then, I used Brasso to polish the disc (disc no longer so minty fresh.). In the end, the disc ends up looking worse off (it will be very cloudy looking), but it is readable by a CD drive - which is what is important.

The worst types of scratches are circular ones. If your discs have any of these, buffing these out will help tremendously.

If you search these forums for both "brasso" and "toothepaste", you'll find some good info on these methods.

[pdq]

I've used Brasso to clean up badly scratched discs from the library.

Were these purchased from the library, or just on loan. Personally I don't think I would have the guts to polish a disc that I didn't own.

[maggior]

I've used Brasso to clean up badly scratched discs from the library.

Were these purchased from the library, or just on loan. Personally I don't think I would have the guts to polish a disc that I didn't own.

It was on loan from the library. The disc was in such bad shape, I figured I certainly couldn't make it worse. As it turns out, I should have sent them a repair bill :-). Though I have to admit I was concerned when the disc got all cloudy after using Brasso on it.

This post has been edited by maggior: Dec 7 2009, 20:05

[Eli]

I have been asking spoon to add a feature to do just this to dBpoweramp:

http://forum.dbpoweramp.com/showthread.php?t=19495

- re-rip the same disc in multiple drives
- rip different copies of the same disc to build one good copy (my understanding is that this should be possible and verifiable, even with different pressings)

Add your support.

[greynol]

Yes, definitely possible. I've done this manually (and successfully!) on a few occasions. The trick is a good audio editor with a decent zoom function that lets you view multiple files at the same time that are time-synched. The ability to conduct a mix-paste is essential to making the process easy when doing this manually.

[2Bdecided]

I don't think the ripper needs to be part of the solution. This can easily be a post-ripping process. With three copies of the disc, you don't need the ripper to tell you where the errors are. With two copies it would help though!

Cheers,
David.

[greynol]

Sometimes using a second drive with the same disc can solve the problem.

Either way, doing this manually requires a considerable amount of work that is surely possible to automate with software.

Regarding the post about ddrescue (containing a plug for an OS boot disc), I believe it was spam and removed it and the related posts. If the guy who started it wants to read TOS #14 and then explain to me how it will work for CDDA via personal message, I will gladly put it back.

This post has been edited by greynol: Dec 9 2009, 16:53

[knutinh]

Having 3 or more versions of the same ripping with uncorrelated errors would be quite simple and fast to implement in matlab, just like 2Bdecided suggested.

The following lines seems to do the trick:

CODE

%% load 3 separate distorted files
a = wavread('file1.wav');
b = wavread('file2.wav');
c = wavread('file3.wav');

%% find most likely
d = median(cat(3,a,b,c),3);

%% write output
wavwrite(d);

As indicated by this test-code:

CODE

%% test-code
load handel
y = [y flipud(y)];
a = y;
a(100,1) = 1;
b = y;
b(1000,2) = -1;
c = y;
c(1,:) = 1;

tic
d = median(cat(3,a,b,c),3);
t=toc;
samples_on_cd = 44100*60*73;
time_spent_processing_one_CD = samples_on_cd*t/length(y)

assert(all(d(:)==y(:)))
assert(~all(d(:)==a(:)))
assert(~all(d(:)==b(:)))
assert(~all(d(:)==c(:)))

linear scaling of results suggest that processing one CD would take 1 minute of processing time - excluding disk i/o. In practice you might run out of memory for such large files. Frame-based i/o and processing is just a few lines more.

If the error model include sample insertion/dropping, then you need a much more complicated algo, doing correlation or something.

-k

[Eli]

I don't think the ripper needs to be part of the solution. This can easily be a post-ripping process. With three copies of the disc, you don't need the ripper to tell you where the errors are. With two copies it would help though!

Cheers,
David.

I didn't say the ripper needs to be part of the solution. But the process would lend it self very well to automation and would be a very slick feature that would make sense in a ripper. It is the ripper after all that will be used with multiple copies of the disc or with different drives. And this would seem like a very good "pro" feature in an app like dBpoweramp that would make it stand out. Spoon is already doing something like this internally for re-ripping different parts of a disc, it would seem easily ported to carry this same code across ripping sessions.

[knutinh]

Given that 3 separate rippings of the same waveform with uncorrelated error is easy to fix (given that 2 "good" samples can be found for every sample, and that no insertion/dropping occurs), what about the more challenging cases?

1. Selecting the most probable sample from N channels when all N channels are different (choose the sample that minimize high-frequency content within a window)

2. Given one ripping with errors (known or unknown sample indexes) and accurate-rip, is it possible to find the original waveform brute-force?

3. Given that sample#M is known to be wrong, what is the optimal concealement technique? I have seen a Philips spec-sheet where they used 1st-order interpolation for single samples, and sample-hold for error runs. Surely this can be improved?

Intuitively, I would guess that when a single sample is missing, some fc=fs/4 windowed ideal brickwall filter would be a good starting point. But should you just fill inn the missing sample based on those coefficients, or should you also alter neighbor (known good) samples to avoid discontinuities?

-k

This post has been edited by knutinh: Dec 10 2009, 11:32