FS#44116 - [libmp3splt] 0.9.2-1 splits very badly
Attached to Project:
Arch Linux
Opened by thelonius (thelonius) - Monday, 09 March 2015, 17:12 GMT
Last edited by Tobias Powalowski (tpowa) - Wednesday, 18 March 2015, 15:27 GMT
Opened by thelonius (thelonius) - Monday, 09 March 2015, 17:12 GMT
Last edited by Tobias Powalowski (tpowa) - Wednesday, 18 March 2015, 15:27 GMT
|
Details
Description:
The current libmp3splt-0.9.2-1 splits very badly at least for VBR mode when splitting in equal parts as well as when splitting to certain length. BADLY means the following: I had one file like this (I had much more of course): 01.mp3 20:01 6.6 MB 45.8 kbit/s 32.0 kHz 20:01 is the duration. With the commandline mp3splt -q -f -o @f-split@n -S 4 ./01.mp3 which means "split the file in 4 equal parts" I got the following results: 01-split1.mp3 7:10 2.3 MB 45.2 kbit/s 32.0 kHz 01-split2.mp3 7:10 2.4 MB 46.3 kbit/s 32.0 kHz 01-split3.mp3 5:41 1.9 MB 45.9 kbit/s 32.0 kHz and this output: Processing file './01.mp3' ... info: file matches the plugin 'mp3 (libmad)' info: MPEG 1 Layer 3 - 32000 Hz - Joint Stereo - FRAME MODE - Total time: 28m.40s info: starting 'split in equal tracks' mode File "./01-split1.mp3" created File "./01-split2.mp3" created File "./01-split3.mp3" created Processed 33382 frames - Sync errors: 0 split in equal tracks ok After downgrading to libmp3splt-0.9.0-2 the same file got split like this: 01-split1.mp3 5:00 1.6 MB 45.0 kbit/s 32.0 kHz 01-split2.mp3 5:00 1.6 MB 46.0 kbit/s 32.0 kHz 01-split3.mp3 5:00 1.7 MB 46.4 kbit/s 32.0 kHz 01-split4.mp3 5:00 1.6 MB 46.0 kbit/s 32.0 kHz which is perfect and what I expect. Sorry, I didn't test the intermediate version libmp3splt-0.9.1a-1. The version of the mp3splt command line tool doesn't matter. This is probably easily reproducable with arbitrary mp3-files at least with vbr-mode files. |
This task depends upon
Closed by Tobias Powalowski (tpowa)
Wednesday, 18 March 2015, 15:27 GMT
Reason for closing: Upstream
Wednesday, 18 March 2015, 15:27 GMT
Reason for closing: Upstream
Comment by
Tobias Powalowski (tpowa) -
Wednesday, 18 March 2015, 15:27 GMT
You need to report this upstream. This is not a packaging bug.