Tutorial: Fix a broken rip with another one |
![]() ![]() |
Tutorial: Fix a broken rip with another one |
Sep 6 2010, 20:54
Post
#1
|
|
|
Group: Members Posts: 698 Joined: 6-March 10 Member No.: 78779 |
By request I share some recent experience. A severely scratched CD could not be saved with the usual polishing methods. All tracks ripped fine except for one with several nonrecoverable errors. I was lucky to find the same CD on Rapidshare, but with messed up gaps and also several ripping errors. With the following steps able to repair the broken track, so that the whole image passed an AccurateRip verification again.
Requirements:
Preparation: You don't need to care about small offset differences, these will be aligned automatically. Only if the download differs significantly from your rip, and has maybe a lot of leading silence added (a few seconds) or a lot of leading silence removed, make it roughly the same. Sample precision is not necessary. Several thousand samples difference are no problem. Do only modify the downloaded sample, not your own ripped track. To reduce the number of samples, that need to be fixed, try a brute force extraction first. If you have a C2 capable drive, you can set dBPoweramp to try a very high number of re-reads, e. g. 2000, per broken sector. EAC won't do so many retries, but you can at least set the error recovery quality to "high". Processing:
EAC will now list all different intervals, that differ in both files. Offset differences are automatically taken care of. Now for each differing interval that is also listed in your rip log click "Replace". If you blindly click replace for all entries, you might import errors from the downloaded file, which were ripped correctly in you version. When you now save the result and check it along with the other ripped tracks in CUETools against the AccurateRip DB, with some luck, you will have an exact match. It has worked for me, despite the very bad quality of the downloaded track. This post has been edited by greynol: Sep 7 2010, 19:28
Reason for edit: Removed the word "unknown".
|
|
|
|
Sep 7 2010, 13:34
Post
#2
|
|
|
Group: Members Posts: 256 Joined: 29-April 10 Member No.: 80274 |
Interesting. I tried with CEP something similar, like replacing (mix paste: replace) damaged part from scratched CD track with one provided by "massive distributed backup system" but no matter what I did (and I correctly worked in sample units with regard to start offset etc), there was always power burst at left side on pasted range. OTOH my knowledge is limited and maybe this can be done comfortably in CEP/Au, but the way EAC editor finds diffs and corrects them is great tip. Thanks again
EAC is mystery in itself |
|
|
|
Sep 7 2010, 18:52
Post
#3
|
|
![]() Group: Super Moderator Posts: 9365 Joined: 1-April 04 Member No.: 13167 |
http://www.hydrogenaudio.org/forums/index....showtopic=83418
Make sure options to smooth are disabled. I don't think you need to disable options to dither, but I could easily be wrong. -------------------- Everything sounds the same until it is proven otherwise.
|
|
|
|
Sep 7 2010, 19:11
Post
#4
|
|
|
Group: Members Posts: 698 Joined: 6-March 10 Member No.: 78779 |
Thanks. I just gave up on Audition in the end, but didn't know why it had failed. The smoothing option is very likely. But I'm also somewhat helpful that it failed, else I wouldn't have found EAC's one-click solution.
|
|
|
|
Sep 8 2010, 09:43
Post
#5
|
|
|
Group: Members Posts: 256 Joined: 29-April 10 Member No.: 80274 |
Somehow I never noticed that option. "Smooth Delete and Cut boundaries" was offender
I tried this on a deteriorated CD with a good small scratch, then ripped problematic track (2:22 s length) with EAC high quality (it took around 6 h on TEAC W58E, C2 capable) and then quick burst rip without EC. Bursted rip had less corrections to be made: http://db.tt/hF5uA19 and greater time saver - at least I didn't went to 2000 repeats with dBP Also I could continue that bursted rip needed overall 8.7 s replacements against 9.5 s with EC, while max interval is 66 ms (~3 K samples) and min 0 ms (< 44 samples) (sorry but I was bored) Then, I didn't feel like taking picture from the CD (main reason) then upload it on my PC then on net, so I googled for picture and guess what - I found it on HA wiki article: http://wiki.hydrogenaudio.org/index.php?ti...CDex_SecureMode So in such cases, pasting larger part from confident correction file with CEP also works fine (although in this extreme case where scratch is over whole track, some other method is more reasonable) |
|
|
|
![]() ![]() |
|
Lo-Fi Version | Time is now: 19th June 2013 - 07:54 |