Hi, people!

I've been doing some more testing on the EAC vs. cdparanoia/cdrdao subject (the original post is here). This time I've checked the EAC rips against cdparanoia ones and got some strange results once again!

The CD drive is once again Pioneer DVR-110D (1.41+TDB):
Read offset correction: +48
Caches audio (2000 Kb)
Overread: Lead-out
Ripping 1 CD as a single wav.
Cdparanoia version: 10pre1 (svn) with an altered readahead value, OS: Ubuntu Feisty

Here's what I have found out so far:

1. md5sums of EAC and paranoia rips never match, but see #2;
2. audio data is BYTE IDENTICAL, found this out by comparing individual frames;
3. audio data is sometimes different on scratched places (which EAC calls suspicious)
4. With offset correction set, EAC successfully overreads into lead-out, but FAILS to "overread into data track": when dealing with cdextra/cdplus media, the rip length is not a multiply of 588 frames - there are 48 missing samples in the end, which means that the rip is truncated! Shouldn't it ALWAYS divide by 588 (2352 bytes)?
5. When ripping the same (cdextra) CDs with cdparanoia, I get the rips with filesize = 44 + 2352 * frames. At least, cdparanoia always outputs redbook-compatible audio files.
6. cdparanoia in fact did not overread to lead-out, but appended 48 zero frames to the end of rips. I've stumbled upon this with a CD with a last track filled with noise to the very end. Byte-comparing the rips revealed, that last 48 frames in paranoia rip are zeroes. EAC rip contained some audio in this place...

So, EAC failed to overread into data track, cdparanoia failed to overread into lead-out. This makes me think that neither one is the perfect tool. What do you guys think?

P. S. Forgot to mention, that most rips were actually 100% sample-identical, so I think, it's still okay to rip with cdparanoia if you're not after perfect backups

First, 2000kb of cached audio? How did you determine this?

Second, I have a few drives that can overread into the lead-out (Plextor, Yamaha), and none of them can overread when there is a data track.

Third, you need to configure EAC to fill up missing offset samples with silence.

As far as what I think, two words come to mind: obsessive and anal. Before taking this the wrong way, consider that I was once this way. There are many other things to worry about in life than 48 samples of data that are inaudible; and if they were audible, then the guy who mastered the disc did a really terrible job.

Regarding cdparanoia, I'll defer to someone who stands behind it. In the event that I find EAC's results questionable, I turn to dBpowerAMP and/or Plextools.

With the cachex utility.



Wow, thanks for the info! So, the the only way to rip cd-extra is to fill missing offset samples with silence indeed. It seems that cdparanoia does just that without asking.

Yes, I'm quite anal about CD backups. I just want to be sure that the hard disk copy EXACTLY matches a CD, not for listening but for archiving purposes...

If I'm not mistaken, this problem might get fixed in the next release of EAC...

http://www.digital-inn.de/124409-post10.html