cdmackay wrote: > This doesn't explain *why* it might sound better, though? > > We know it shouldn't sound (much) worse, and these tests seem to support > that, but why might it sound better? > The way most of these tests are being done is that you write a set of rules ("we can compress in this way"), use them to do the compression and then you let people listen to the result (blind ABX test). You then test for two things: whether people can discriminate between the compressed/uncompressed tracks (codec with less discrimination is better) and you test which ones are perceived to sound "better" (better is better).
As a result, you get a codec that is as hard to discriminate as possible but also you get a codec that amplifies characteristics people regard as sounding "good". One thing you can find with a lot of mp3 tracks is an actual perception of "increased transparency" which in turn is NOT a surprise since some frequencies simply get cut out. This is usually perceived as "sounding better", actually especially by people who are trained listeners (they are used to try to identify different characteristics in the music and this becomes easier with the cut-out frequencies). > > An interesting test would be to split the results into groups, depending > on how familiar the subjects were with listening to MP3. > > e.g. one might imagine that subjects that listen to a lot of MP3 might > be more likely to prefer MP3 over lossless, compared to those who > regularly listen go FLAC, or don't regularly listen to MP3. > Has been done at great length. Except for the identification of a few well-known artifacts there is generally no effect of better listening experience or personal preference. Audiophiles (if they ever agree to do ABX tests) don't test better or different than others. This is a bit different to tests for dynamics compression where having listening experience actually _does_ help. There is only one thing that strongly correlates with being able to identify mp3s and that's hearing defects. People who have strongly reduced hearing capabilities in higher frequency ranges can clearly identify mp3s because for them the frequency masking effect doesn't work anymore. ------------------------------------------------------------------------ pippin's Profile: http://forums.slimdevices.com/member.php?userid=13777 View this thread: http://forums.slimdevices.com/showthread.php?t=98010 _______________________________________________ audiophiles mailing list audiophiles@lists.slimdevices.com http://lists.slimdevices.com/mailman/listinfo/audiophiles