[Paranoia] Ripping physically scratched CD's

[grarpamp]

On Fri, Dec 5, 2014 at 11:23 AM, Paolo Saggese <pmsa4-xiph at yahoo.it> wrote:
>  What a typical cdparanoia user expect is that the rip is either perfect
> (a perfect copy of the original) or the whole process fail.
> ...
> the software should fail returning a proper error rather than "hanging"

When called with no handling options other than say a device name
and track selection parameters (ie: 'ok cdparanoia, just do your
thing'), I'd expect it to 'fail' noticeably if the rip isn't both
perfect and self reproducible.

However, much more important to me is that the tool emit a documented
exit status for every possible exit, and emit a parseable log stream
of everything that goes on throughout the entire session regarding
data, metadata and the process from invocation to exit. That will
tell you programmatically if there is a fail or quality issue, the nature
of it, and where.

> If you don't care about rip quality, there are really no reasons to use
> cdparanoia (or libparanoia) in the first place. You can use cdda2wav or
> any other (unsafe) ripping software instead.

Try using two different make/model drives on different controller
busses (PATA/SATA/USB/SCSI), apply any offsets, run both cdparanoia
and cdda2wav, two rips each, eight total, then compare hashes and
logs. If they all match, you're good. Silent data differences do
still exist and happen, if you're paranoid you want to isolate and
catch them. (Hint, use ECC ram and ZFS checksums too).

Note that cdda2wav has a -paranoia option.