[Paranoia] Problem while overreading with offset correction set

Артём Попов artfwo at gmail.com
Sat Apr 28 22:06:13 PDT 2007


Hello!

I'd like to report a problem with overreading to lead-out... I have
been recently comparing my Paranoia rips with several "reference"
rips, originally done with EAC on a Windows machine. The drive was
exactly the same in both cases:

Pioneer DVR-110D, Firmware: 1.41+TDB.
("accurate stream", 2000kb cache, read offset correction: +48, drive
overreads into lead-out)

cdparanoia version: svn (10pre1), built with STATIC=TRUE
kernel: 2.60.20-15 (the default Ubuntu 7.04 setup)

I think, I also should mention, that I'm using a native amd64 version
of Ubuntu. cdparanoia sources have been slightly altered as specified
here:
http://www.nabble.com/cdparanoia-and-cache-t2390032.html

cdparanoia command:
cdparanoia -v -O 48 [.0]-

Well, most of the rips completely matched those done with EAC: md5
sums were different, but audio frames are the same to a single byte :)
Scratched CDs have unmatched frames here and there (in places, that
EAC marks as "suspicious"), I suppose that's quite normal.

But I have a CD with the last track filled with non-zero samples to
the very end. In the EAC rip, all the samples are in place (last 48
samples are non-zero), but in Paranoia rip, the last 48 samples are
all zeroed (offset correction is +48, remember?).

Does it mean, that cdparanoia fails to force drive to overread into
lead-out? I suppose so, but what do you guys think? Maybe I'm just
doing something wrong? I also remember reading about some problems
with latest libata-enabled kernels and cdparanoia on lkml, could this
be the case? Thanks in advance for any information on this matter!!

Best,
--Artem.


More information about the Paranoia mailing list