I think the success rate is like 98%?! 
All good down here now 
F.
3 Likes
That one, singular, checksum match, is you!
Recent versions of EAC will send checksum data to the CueTools Database [CTDB]… if EAC did not detect any hardware errors from you CD-ROM drive while ripping… EXCEPT!.. There was a bug in EAC where it would just always send everything to CTDB even if the disc was completely full of read errors… So now the CTDB is full of checksums of randomly scratched CDs.
Also, each time someone with a scratched CD re-rips their scratched CD with EAC, EAC sends yet another copy of the the idiopathic checksums to the CTDB. It’s not removing duplicate submissions by the same user.
(So, yeah, the CTDB can be flooded with bad data, completely by accident if someone keeps trying to rip their scratched CD over and over again.)
Anyway…
If you’re using EAC to rip, and then CueTools itself afterwards; Mentally subtract your own number of attempts from the (first number) matching count.
3 Likes
I’ve encountered this exact situation with CueTools, and after recovering the zero out of a thousand matched tracks by various means. I’ve been able to compare the raw PCM data between the matching and non-matching rips… and… The files are actually different! But not by very much. Usually it’s a only a few samples here and there, which, curiously, are not too far off from what the correct value should be.
That is to say, they are not wildly off by 70db, making a 30ms wide spike in the audio waveform… the errors looked like the CD-ROM drive itself had silently extrapolated (smoothed) over the read error. So when you listen to the audio with your human ears, you won’t hear anything wrong.
Don’t trust everything that a CD-ROM drive tells you about a CD.
3 Likes