[paranoia] cdparanoia inconsistent results

[J. Scott Edwards]

I have been using cdparanoia for years and I guess I have operating under
a delusion (fantasy?) that the ripping results would be identical.  But
now that I've started checking I'm finding it's not.

Sorry this story is a little long...

This all started several months ago when I started using Grip.  Prior to
that I had always run cdparanoia from the command line by hand with good
results.  After ripping and encoding several dozen CD's with Grip I
discovered that many of the cd's I had ripped had a weird thing which
sounded like a random sector inserted into the middle of the music.

Since grip had deleted all of the wave files I couldn't go back and see if
they were messed up.  So I tried re-ripping them but had Grip save the
wave files.  The wave files had the weird random sector thing thing as
well.  I played around and tried ripping from the command line and it
seemed better (but now I can't remember if there were still problems or
not).  I was running RedHat 7.2 at that time and I thought maybe it was a
problem that cdparanoia wasn't getting good responce time so I installed
the low-latency and pre-emptive kernel patches to see if that helped.  It
didn't seem to make any difference.

Sometime after that I switched from RedHat to Mandrake and the problem
went away.  But after that I always had Grip keep the wave files and made
a backup CD of them just in case.  Later I switched that machine to Gentoo
Linux and although I have only ripped a few CD's on it, it seems to have
worked as well.

So last month I got an old machine that I wanted to use as a juke box, so
I installed BeOS on it (this is another long story) and started ripping.
Everything went well until the disk filled up.  Since it doesn't have a cd
recorder, I was going to copy the files over my lan to another machine and
back them up there.  After messing around for some time (yet another long
story) I gave up on that idea and decided that I could re-rip them on the
other machine, but them to a cd and then compare them to the files on the
BeOS machine, verify they were the same and delete them.

Well of course they didn't match.  The files were the same length, but
they were not the same.  I looked at the hex dumps and it appeared that
perhaps the lead in was just a different length.  So I wrote a program
that stripped off the lead in and lead out and just left the actual data.
I used it on the files and the first two tracks from the CD were then
identical, so I thought all was well.

But then the next three tracks were different.  They were longer on one
machine than the other.  They appear to be identical up to where the one
file ends, but the other one continues on with additional non-zero
samples.

I have ripped it now on a third (OpenBSD-Pioneer DVD-ROM drive) machine
and files are again different than the files ripped on the Linux-Plextor
Plexwriter drive machine.  I am going to compare them now to the BeOS
ripped files.

My guess is that cdparanoia isn't always starting in the same place, but
always rips the same length (based on the toc?) and on the tracks where
one is longer it started too early read too much lead-in and then cut off
the last of the track?

Thanks
  -Scott

<p>--- >8 ----
List archives:  http://www.xiph.org/archives/
Paranoia homepage: http://www.xiph.org/paranoia/
To unsubscribe from this list, send a message to 'paranoia-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.