Get one good rip from multiple damaged copies of a CD

[dseomn]

I currently have two copies of the same CD, with the same disc ID, and they both fail to rip because of scratches. Would it make sense to add a feature to get one good rip from multiple damaged copies of the same CD, assuming there's no single sector that's damaged in all copies? I think #649 would help if the damage is in different tracks, but not different sectors of the same track.

[drjaska]

I have found CDs which with some drives consistently read a glitched track due to a poorly placed scratch, whipper does not find anything wrong with it as the CD does not have an AccurateRip database entry. I would not trust scratched discs for bit perfect CD digitalisation unless the CD has an AR db entry, in which case #667 would help more than this feature request. I would recommend using cdparanoia if you just want a copy with no audible glitches.

[dseomn]

That would definitely help too, but I think using a combination of multiple discs and AR would help even more. That way completely unreadable scratches on one disc could be read from another, and the final result checked against AR.

[drjaska]

You would in my opinion benefit more from #681 being fixed or whipper image verify *.cue.

The first approach would enable you to delete the log file and previous DAMAGED tracks from the path and then rip the CD in a way which reuses the old tracks instead of reripping them and writes a new log file.

The second would allow you to pick various tracks in flac/wav to assemble to be a disc and then verify it for integrity (at least partially, idk how many integrity checks there are) and AccurateRip checksums.

[dseomn]

Thanks, that does look useful. #649 in combination with whipper image verify would let me manually splice together rips of the same track with different scratches, then check it against AR.

[drjaska]

I do not understand how #649 would help you at all unless you wish to collect possibly garbage data, assuming that you already know of -k / --keep-going "continue ripping further tracks instead of giving up if a track can't be ripped".

[dseomn]

I haven't tried --keep-going yet, but I thought that #649 was needed to keep the good data around the errors in a failed track? Otherwise I don't see how I could splice together multiple rips of a track with errors in different places. (I also haven't tried that hard yet. I'm going to see if I can get a 3rd copy to get a single good rip from. If that's damaged too, I'll look more into splicing.)