title
Products            Buy            Support Forum            Professional            About            Codec Central
 

interesting format question ~ not really for support

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • willyhoops
    dBpoweramp Enthusiast
    • Mar 2007
    • 63

    interesting format question ~ not really for support

    sorry not a support question but if you have questions about audio formats i can't see where else one would put them....

    if one takes a say 256kbps mp3 and converts to a lossless format like flac, and then converts back to mp3 at 256kbps do you get a file with the same audio quality as the original?

    my initial thought was no and then I began thinking one probably does.
  • LtData
    dBpoweramp Guru
    • May 2004
    • 8288

    #2
    Re: interesting format question ~ not really for support

    Remember that once you convert to a lossy format (mp3, WMA, etc.) you can never regain what was lost. Converting a mp3 format to either a higher-bitrate mp3 or to a lossless file is just a waste of space.

    Some more information on audio formats can be found here: http://www.dbpoweramp.com/spoons-aud...de-formats.htm

    Comment

    • xoas
      dBpoweramp Guru
      • Apr 2002
      • 2662

      #3
      Re: interesting format question ~ not really for support

      If one takes a say 256kbps mp3 and converts to a lossless format like flac, and then converts back to mp3 at 256kbps do you get a file with the same audio quality as the original?
      Interesting theoretical question.
      My reasoning would be this:
      Some audio information (audio quality) was lost in the conversion from the original source to mp3 at 256 kbs, although the amount of loss might not necessarily be perceptable to any given listener (not for most listeners, but there would probably be a few listeners who would determine a difference).
      No audio quality should be lost in the conversion to lossless.
      In the reconversion to mp3 at 256 kbs my educated guess is that additional audio information would be lost but that the degree of loss would likely be less than the amount of loss in the original conversion (because the same codec and same settings are involved I would think that there would be less loss than if the last conversion were from lossless to a different lossy codec-like Ogg Vorbis or Musepack-because the algorithms for the conversion would be the same as those used for the original .mp3).
      So I think there would be some loss but that the loss would be even less apparent to a listener than the loss from the original source to the original 256 kbs .mp3.

      I am just about positive that the degree of loss would be exactly the same whether you first converted from 256 kbs mp3 to 256 kbs mp3 or if you converted to lossless in the middle. I think that either way more information would be lost than if you simply made a copy of the first mp3 file. This would make make the conversion to from mp3 to lossless a misguided exercise (as LtData points out you gain no quality but use more space). This leads to the principle that a lossy file is always its own best copy.

      I willingly admit that I have no hard data for my answer on this, only what I have read about here and elsewhere about conversions and a few periods of extensive listening tests on my part regarding different formats and settings. I would be interested in seeing any other facts or opinions on this issue that others might have to offer.

      Best wishes,
      Bill

      Comment

      • willyhoops
        dBpoweramp Enthusiast
        • Mar 2007
        • 63

        #4
        Re: interesting format question ~ not really for support

        i was thinking that in the conversion of lossless back to mp3 the infomation that is already missing would be exactly removed again - ie there would be no loss at all as long as the bitrate was the same. also if the bitrate was a little lower it wouldn't be so different from moving from the original to that lower bit rate in one step. and if the bit rate was a little higher it might make no difference becuase the extra information it would have preserved is already missing.

        So quite different from analouge copies, for example, where every conversion causes a drop in quality.

        I tried finding the answer on google but couldn't get it. If you know a better place to post it please let me know. It's a thing which an audio compression expert would know the answer to right away.

        I think that this is an important question becuase say we have some old mp3 tracks hanging around but keep everthing else in flac. Then its fine to convert the mp3 to flac. But what if we occasionally export our flac collection back to mp3 for playback on a small drive player. If there is a big generational drop on every conversion this would be a real bad thing to do. But if it's always removing the same kind of audio information then this is not a bad thing to do.
        Last edited by willyhoops; May 01, 2007, 08:21 AM.

        Comment

        • willyhoops
          dBpoweramp Enthusiast
          • Mar 2007
          • 63

          #5
          Re: interesting format question ~ not really for support

          I did find this on the internet:

          -----------------------------

          (1) On lossy compression:

          Repeated applications of lossy compression and decompression can cause generation loss. Some lossy compression algorithms are much worse than others in this regard. Generation loss caused by lossy compression can be made worse if the parameters used are not consistent across generations. For example, with JPEG, changing the quality setting will cause different quantisation constants to be used, causing a lot of extra loss.

          (2) On a thread discussing exactly mp3->wav->mp3 :

          MP3s don't come out the same when you recompress them because although the decompression->compression process is mathematically an identity transform, computationally it is not. The MDCT must be approximated on a computer, and in addition, all compressors use heuristics to speed up the encoding process since it is impractical to try every possible encoding. The inaccuracy also applies to decoding. Only if you had a perfect encoder and decoder that used infinite precision arithmetic would you get no degredation. Unfortunately, we don't have such an encoder/decoder.

          -----------------------------

          However it doesn't really answer it properly for me because it says "not quite" but not "how much". Since it's an important number there should be a figure which says for each audio format how much of a loss is introduced by compressed->uncompressed->compressed (Eg orig->256kbps mp3->uncompressed->256kbps mp3->uncompressed is typically approximately equivalent to orig->248kbps mp3->uncompressed??). Also general comments about going compressed mp3 265kbps->uncompressed->compressed mp3 128kbps for example.

          I just haven't found the right web site that correctly and authoritatively answers these questions.
          Last edited by willyhoops; May 01, 2007, 09:01 AM.

          Comment

          Working...

          ]]>