joatca;187132 Wrote: > I wondered if there might be a problem like that. To check it, I made up > another example. This time I've included 320Kbs, 128Kbs and VBR as well > as a 56Kbs MP3 of the original so you can hear approximately what it > sounds like (it's Wolfstone's "Tinnie Run" - 15 second guitar riff then > lots of drums and acoustic guitars). > > I also included a screenshot from Audacity zoomed right in to the > sample level on one peak on the original and each of the three > encoded/decoded files. I'm no audio expert, but they sure don't look > like they are out of sync in the time domain, but I can see subtle > differences in the waveforms. I didn't do anything to them other than > load them up in the same project then zoom in. > > The files are all here, with filenames that should be > self-explanatory: > > http://qqqq.ca/files/audio-comparison/
I agree that there could be time differences, but I would think the method would be extraordinarily sensitive to this and it would be very dramatically audible for even a very small time difference. Plus, this latest round of tests does confirm that 128 kbps drops a lot more material, and VBR still sounds different than CBR in terms of what it drops and on which parts of the musical passage. If the subtraction effects were overridden with time differences, the effects would be random but 128 kbps clearly drops more material than 320 kbps in this case, as you would expect. In fact the 320 difference test shows almost no differences in the guitar riff until the big cymbal crash. Interestingly, the cymbal crash drops a lot of material for all tests. -- Mark Lanctot ------------------------------------------------------------------------ Mark Lanctot's Profile: http://forums.slimdevices.com/member.php?userid=2071 View this thread: http://forums.slimdevices.com/showthread.php?t=32576 _______________________________________________ audiophiles mailing list [email protected] http://lists.slimdevices.com/lists/listinfo/audiophiles
