Skip to main content

Topic: EAC... how perfect? How does it work? (Read 37171 times) previous topic - next topic

0 Members and 1 Guest are viewing this topic.
  • Patsoe
  • [*][*][*][*][*]
  • Members (Donating)
EAC... how perfect? How does it work?
Reply #25
Quote
Let me rephrase: in digital audio, a 1 that should be a 0 can result in
entirely different sound, or not? I can't see how errors could add/remove
bass, depth etc. from a song...

If you would simply send the digital values to a dac, even when you know that there are wrong values among them (and error correction circuits can tell you this), then yes. It will result in pops.

But, since most drives know (from error correction circuitry) when a value is wrong, they will not send the wrong value, but make a guess (by interpolation) instead.

Interpolation makes the wave as smooth as possible, so it is then possible to have something that is still clearly music, but less precise.

I think we're talking about the same things now? 

  • tigre
  • [*][*][*][*][*]
EAC... how perfect? How does it work?
Reply #26
Quote
Let me rephrase: in digital audio, a 1 that should be a 0 can result in entirely different sound, or not? I can't see how errors could add/remove bass, depth etc. from a song...

1. If errors are detected but the original information is not restorable (by e.g. using error recovery information) interpolation will be done.

Some (theoretical) example on the consequences:
1.1. Imagine, every 2nd sample is erroreous (1st correct, 2nd wrong, 3rd correct, 4th wrong ...) (and detected correctly). If best possible interpolation for this case is done, the result is the same as resampling a 22.05kHz signal to 44.1kHz. It's the same as applying a 11.025kHz lowpass. If the interpolation is not ideal for this case, e.g. linear: valueX=1/2(valueW+valueY), frequencies > 11kHz won't be cut away but "modified", causing aliassing and distortion. This is an extreme example, but basically it's the same if a few samples get interpolated. High frequency content is modified.

1.2. Imagine, you have a sharp transient, e.g. digital silence and suddenly a sample value of x000. If some samples before the 1st x000 sample are interpolated, the sharpness of the transient will be damaged, similar to pre-echo.

2.If errors are not detected e.g. because error detection information is not used, random errors can occur.

This will result in a single sound, described as "click", "pop", "digital kiss"  , depending on the error pattern.

You can try these things yourself using a hexeditor and/or a wave editor.

Corrections are welcome.
Let's suppose that rain washes out a picnic. Who is feeling negative? The rain? Or YOU? What's causing the negative feeling? The rain or your reaction? - Anthony De Mello

  • Radetzky
  • [*][*]
EAC... how perfect? How does it work?
Reply #27
I understand I've been somewhat redundant with my questions.  It is
because from the answers I got, I wasn't so sure people understood
correctly what I tried to say.  I wrote the same questions from
different angles.  I think I got the answers I was looking for now.

BTW, if you check "FILE/EAC OPTIONS/EXTRACTION/DELETE LEADING AND
TRAILING SILENTS BLOCKS", isn't that a good way to compare the quality
of extraction of different drives?  This way you eliminate the offset
discrepancy between the drives.  This should let you compare if the
actual musical information is correctly extracted.

Yesterday I tried to rip a track with an offset of -10000, then 0 and
then 10000.  I had also checked the option for deleting the leading
and trailing silents blocks.  EAC couldn't find a difference with the
wav files (using "TOOLS/COMPARE WAVS..."). 

Tonight, I decided to rip one track using my LGE low-profile combo
drive and a Plextor CDRW drive that I still have access to.  I also
configured EAC to remove the silent blocks.  The difference in the wav
files reported by EAC was "3 repeated samples 0:00:00.012" on the
track ripped from my drive.  I think this is pretty good!

I'll stop asking the same questions differently. Promess.

  • Patsoe
  • [*][*][*][*][*]
  • Members (Donating)
EAC... how perfect? How does it work?
Reply #28
Usually, deleting silence does the job, yes. You can do something similar by checking "EAC Options/Extraction/No use of null samples for CRC calculations".

On many of my cds, though, I still find different CRCs when not using offset correction, even though I use the mentioned option. Most probably, this is because some tracks are not silent on their boundaries (a good example are live recordings).

Oh, and don't let me hold you back from asking three, four or five times if things are unclear... I was just kidding there