cdparanoia and drive offsets

2003-05-30 18:52:38

I know the read offset of my drive is -472, and I have no problem correcting that offset in EAC or AccurateRip.  Furthermore, when I rip in Linux using cdparanoia and no special options, the "WAV Compare" in EAC tells me that I have 472 repeated samples.  Now if I rip again with cdparanoia using the "--sample-offset -1" option,  comparing WAVs in EAC tells me I have 471 repeated samples.  So I'm on the right track, right?  But if I use a "--sample-offset -472" option for cdparanoia, the WAV Compare does not repeat any repeated samples, but does report "different samples" for what appears to be the whole track.  Calculating md5sums verifies that the two tracks are not identical.

Has anybody had problems using offsets in cdparanoia?  Any suggestions as to what my problem is, and how I might correct it?

cdparanoia and drive offsets

Reply #1 – 2003-05-30 19:13:15

There's something about posting a problem publicly that can suddenly give you the insight to solve your own problem...

My problem wasn't with cdparanoia, it was with EAC.  I had "Fill up missing offset samples with silence" checked in my EAC options, and doing so caused my EAC rips to not match either my AccurateRip or cdparanoia rips.  I'm not sure if that EAC option is checked by default or not, and I haven't heard it discussed much on these forums, so I'm not sure how important it is, but now all three rips match, so I'm leaving it unchecked.

cdparanoia and drive offsets

Reply #2 – 2003-05-30 19:24:07

When I said "all three rips match," I said so because EAC's WAV Compare identifies them all as the same track.  However, finding md5sums for each track (produced with EAC, cdparanoia, and AccurateRip) shows that the AccurateRip track is different.  In fact, the file produced by AccurateRip is 2 bytes larger than those made with EAC and cdparanoia.  Any idea why?

cdparanoia and drive offsets

Reply #3 – 2003-05-30 21:36:28

I seem to have made quite a mess of things, so let's see if I can straighten this out (for my own amusement...nobody seems really interested in this thread anyway...):

• My drive offset is 472, so my correction is -472, meaning it should start reading 472 samples earlier than it would normally
  
• I'm ripping tracks using EAC and AccurateRip for Windows and cdparanoia in Linux
  
• Track 1 fails with both cdparanoia and AccurateRip, most likely because the offset correction is making the drive try to read outside of where the drive thinks the audio data exists.  EAC produces different files depending on whether I check "Fill up missing offset samples with silence".
  
• Track 7 and Track 15 (the last track) rips the same with cdparanoia and EAC, but with AccurateRip each wave file is two bytes larger than the others.

Earlier I said I had to uncheck the "fill up missing offset samples with silence" option in EAC to produce a rip equivalent to what I get with cdparanoia, but further testing shows that isn't the case (and suggests I was a little off my rocker this morning).  All I can really conclude is that AccurateRip does something that is quite different than EAC and cdparanoia that is causing it to make files that are 2 bytes larger, and that its extraction routine has problems when offset correction is applied.  I've posted on the dBpowerAMP forums - I really like the AccurateRip concept, and want to do what I can to see it gain acceptance.

cdparanoia and drive offsets

Reply #4 – 2003-05-30 22:18:45

Nice talking to yourself, eh?

Anyway, glad we could help you.

cdparanoia and drive offsets

Reply #5 – 2003-05-30 22:44:35

Hi, i also have some problems with cdparanoia's --sample-offset parameter...
whenever trying to rip track 1 with a negative sample offset i get "Error parsing span argument"... any other track can be ripped this way
when i set the same negative offset in accuraterip all tracks can be ripped perfectly... any ideas for cdparanoia?

cdparanoia and drive offsets

Reply #6 – 2003-05-31 01:19:02

If EAC says they are the same and the files are two bytes bigger, there must be non audio data added to the wav file.

cdparanoia and drive offsets

Reply #7 – 2003-05-31 03:04:22

Spoon gave me a simple answer for the 2-byte difference - AccurateRip adds a different WAV header to the file than EAC or cdparanoia.  If I compress all my wave files to flac, I can compare sizes and md5sums with no problem.

So that only leaves one problem...AccurateRip and cdparanoia fail to rip the first track if I apply a negative offset correction.  AccurateRip occasionally fails on other tracks, probably for some other reason, but at least I'm down to one simple problem that can be easily solved by just using EAC.  Anyway, it's been a learning process, and I'm finally understanding offsets, overreads, and all that good stuff...

cdparanoia and drive offsets

Reply #8 – 2003-05-31 09:29:22

Ah right I know what it will be, I think your CD drive is unable to overread outside the disc start and end - not sure why EAC is not failling, it must just return 0s.

cdparanoia and drive offsets

Reply #9 – 2003-05-31 16:16:11

I think you're right, at least according to one of the online drive databases.  It's too bad, too...now that I feel like I really know what I'm doing, I find out that my hardware won't be able to do it anyway...