putting 2 identical mp3s with different metadata into spek produces 2 slightly different looking spectrograms, even when the underlying audio data is identical. (i stripped away the metadata of both files and they had matching md5 hashes. the spectrograms also matched perfectly when their metadata was removed)
File A
original mp3 (metadata mistakenly showed a duration of 74:33:55 when it was really 2:40)
File B
original mp3 -> metadata modified by mpTrim to show the correct duration which is 2:40 (underlying audio data of B is still identical to A)
File C
original mp3 -> metadata cleared with TagStripperororiginal mp3 -> metadata modified by mpTrim -> metadata cleared with TagStripper
the spectrograms i got from spek for A and B were different, yet the filesizes were very similar, so i cleared the metadata of both files using TagStripper. the 2 resulting files with cleared metadata had identical md5 hashes, so basically they were the same file. thus, the underlying audio data of A and B must also have been the same
also curious was that neither A nor B's spectrogram matched up with C's. this feels like a pretty widespread bug that should be easy to duplicate
bug was produced using mac version 0.8.2.3
putting 2 identical mp3s with different metadata into spek produces 2 slightly different looking spectrograms, even when the underlying audio data is identical. (i stripped away the metadata of both files and they had matching md5 hashes. the spectrograms also matched perfectly when their metadata was removed)
File A
original mp3 (metadata mistakenly showed a duration of 74:33:55 when it was really 2:40)
File B
original mp3 -> metadata modified by mpTrim to show the correct duration which is 2:40 (underlying audio data of B is still identical to A)
File C
original mp3 -> metadata cleared with TagStripper or original mp3 -> metadata modified by mpTrim -> metadata cleared with TagStripper
the spectrograms i got from spek for A and B were different, yet the filesizes were very similar, so i cleared the metadata of both files using TagStripper. the 2 resulting files with cleared metadata had identical md5 hashes, so basically they were the same file. thus, the underlying audio data of A and B must also have been the same
also curious was that neither A nor B's spectrogram matched up with C's. this feels like a pretty widespread bug that should be easy to duplicate