4

I'm doing some auditing of my music library (using Mutagen/Python), including looking for duplicates. In particular, I am writing a program to look for duplicates and I am using track-length as part of the way I compare two tracks. In the process of doing that, I'm finding some small discrepancies (say 2 seconds) between the track length of the mp3 version versus the m4a version.

Is this "real" (i.e., the different encoding distorts the track length - yikes!) or is this something to do with how the tags are created or how they are read?

If the difference is real and unavoidable, what comparison tolerance should I be setting for equality?

D.W.
  • 3,382
  • 7
  • 44
  • 110
RFlack
  • 436
  • 1
  • 5
  • 19
  • 3
    You haven't told us how you converted from mp3 to m4a. Is it possible that the conversion introduced slight changes in the length of the track? Separately: Have you tried computing the track length using any other tools or libraries, to cross-compare and try to validate the numbers you are getting from Mutagen? – D.W. Dec 20 '13 at 23:22
  • Good points. Id have to check the specific tracks as Ive used various conversions at different times, as well as simply two different uploads. Ill do that and report back; also check Mutagen results against what my players report. My question really is - if all such timings are +/- 2 or 3 seconds (which does sound a lot to me) then its a non issue. Im just not sure what to expect. – RFlack Dec 21 '13 at 00:11
  • 1
    @Andrew... I will have to extract the relevant code snippets, which I can do (may take a little while, Im going in to weekend mode!). I will probably regret saying this, but I don't think the code is the issue. – RFlack Dec 21 '13 at 00:13
  • Apologies for lack of progress, seasonal activities plus ice storm have been interfering. However I have established that iTunes, at least, is reporting the same discrepancy at least in some cases. So that suggests it's not mutagen but the conversion / ripping processes. – RFlack Jan 04 '14 at 16:52

0 Answers0