Extracting PCM WAV files (from CD) and going back to CD

Discussion in 'Computers' started by MarkHastings, Nov 16, 2005.

  1. MarkHastings

    MarkHastings Executive Producer

    Joined:
    Jan 27, 2003
    Messages:
    12,013
    Likes Received:
    1
    Question, if I extract an audio CD as a WAV file (16 bit / 44 kHz / stereo) with PCM audio (i.e. no compression) and then burn that back to a CD...Is there ANY loss at all?

    I'm almsot positive the answer is no, but is there anything that can make it yes? Again, I'm not talking about glitches in the ripping or mistakes, but am I wrong in assuming that it is a 100% flawless transfer? I do realize there is the 'error correction' to worry about, but what else?
     
  2. Marko Berg

    Marko Berg Supporting Actor

    Joined:
    Mar 22, 2002
    Messages:
    857
    Likes Received:
    0
    In theory, the waveform file should be a bit-perfect copy of the original. Whether it's possible something can be altered due to errors in reading the original file and writing the copied file to the hard drive, I don't know.

    I've never been able to hear a difference, and if you can't either, I wouldn't worry about it. For all practical purposes, an audio CD you burned yourself from ripped tracks is identical to the original.
     
  3. MarkHastings

    MarkHastings Executive Producer

    Joined:
    Jan 27, 2003
    Messages:
    12,013
    Likes Received:
    1
    Again, agreed. My question comes from a work related question.

    Our sound engineer was asked this question by the owner of the company. We have CD's and the owner wants the engineer to rip them, do a little editing, then put them back onto CD. He then asked if these new CD's are exact copies.

    Now, I'm sure when he says "exact", I don't think he means 100% exact (unlike burning MP3's to CD which is an obvious no-no). I'm sure those few bits, that may get lost in translation, shouldn't matter into the equation. I figured this was the case, but just wanted to make sure I've got my bases covered here.

    I told the engineer: "I'm sure there's SOMEONE out there who will tell you that it isn't 100% exact, but for all intents and purposes, it IS an exact copy." - but if it isn't 100% exact, it would be helpful to explain to the owner (in technical terms) what sort of loss (no matter how insignificant it may be) could occur.
     
  4. PhilBoy

    PhilBoy Second Unit

    Joined:
    Sep 30, 2003
    Messages:
    427
    Likes Received:
    0
    I've been re-ripping my entire catalog with Exact Audio Copy for the past week. I want them all in .wav form for a HTPC project (CD's on demand etc.) and the software indicates that most rips are 100% (???), a few are 99.9%.

    EAC does it's own error correction upon ripping and there are a few tracks that it 'lingers' on (Bowie's Ryko discs for some reason).

    Just a thought.
     
  5. Ken Chan

    Ken Chan Producer

    Joined:
    Apr 11, 1999
    Messages:
    3,301
    Likes Received:
    0
    Depending on the CD authoring tool, there might be differences in the track gaps, for example. I'm also not sure how the error detection/correction codes are derived (these are part of the control data, not the "real" data). I suppose if the copy is not sample-accurate, the checksums would be different. The control data also contains a unique track ID, but I'm not sure how unique it is.

    However, if you extract the data from the copy and discard the gaps between songs, the songs themselves should be identical to the originals, as long as the ripping and burning are error-free. You can even try that yourself to verify. Data CDs (including CDs with .wav or .mp3 files) are more robust than Audio CDs, because a single wrong bit in a program can crash it, while a single bit off in a song doesn't really matter.
     
  6. MarkHastings

    MarkHastings Executive Producer

    Joined:
    Jan 27, 2003
    Messages:
    12,013
    Likes Received:
    1
    I think I know what you're saying, but let me see if I got it....

    Since audio CD's allow for error's, there is a chance that there could be a slight difference in a copy, where a data CD has to be a 1 for 1 match. Any slight error will cause a bad burn.

    Correct?
     
  7. ChristopherDAC

    ChristopherDAC Producer

    Joined:
    Feb 18, 2004
    Messages:
    3,729
    Likes Received:
    0
    Put it this way: the audio samples themselves are the same, except for the occasional uncorrected error [~1 per 60 minutes, per spec']. The non-audio data on the disc -- subcodes, timecode, grpahics track, timestamp, you name it -- will be different.
    And, yes, the error correction on CD-ROM is more robust than that on Audio CD because "error concealment" does not work: you can't take the average between two bytes in, say, a text file and have the result come out to approximate the value of the missing byte in the middle.
     
  8. Kimmo Jaskari

    Kimmo Jaskari Screenwriter

    Joined:
    Feb 27, 2000
    Messages:
    1,528
    Likes Received:
    0
    Anyone who wants really accurate EAC rips should also look into offsets.

    Coasterfactory.org has a nice tutorial that contains info about that, worth checking out.

    Offset info at coasterfactory.org
     
  9. Ken Chan

    Ken Chan Producer

    Joined:
    Apr 11, 1999
    Messages:
    3,301
    Likes Received:
    0
    Right. If you're burning a data CD, a single error is not acceptable, just like with an disk-to-disk copy. But for an audio CD, you probably wouldn't even notice, so they're allowed. It's designed into the spec, e.g. the error concealment that Christopher mentions. Audio CDs store more "user" bytes than data CDs, but all the bytes are (slightly) more suspect. Data CDs allocate some storage space to error correction, among other things, making them more suitable for storing data.

    In your case, you're copying from an audio CD, writing back to an audio CD, and then someone eventually reads from the copy. If you think about it that way, there are three places where things might not be 100% correct -- but again, not that you'd notice. With a data CD, you'd have the same three places, but the odds that each are not 100% are much much lower -- it "shouldn't" happen, although obviously it does from time to time.
     

Share This Page