search

nstanger / process_podcast

Python script for processing and assembling a podcast from various input sources
0 stars 0 forks source link

[CLOSED] process_lectorial_capture: concatenation problem #5

Closed nstanger closed 8 years ago

nstanger commented 8 years ago

Issue by nstanger Friday Jul 15, 2016 at 03:17 GMT Originally opened as https://github.com/Otago-InfoSci-Database/Tools/issues/5

The temporary files are generated OK, but don’t concatenate correctly. ffmpeg spews “non-monotonous DTS” warning, and the output file only correctly captures the first scene. The rest are garbled and somehow spread over another 20 minutes or so, instead of perhaps 30 seconds.

nstanger commented 8 years ago

Comment by nstanger Friday Jul 15, 2016 at 03:21 GMT

The concat filter seems to work (see https://trac.ffmpeg.org/wiki/Concatenate):

ffmpeg -y -i temp_INFO321_2016-07-13_00_a.mp4 -i temp_INFO321_2016-07-13_00_b.mp4 -i temp_INFO321_2016-07-13_01_a.mp4 -filter_complex '[0:v] [0:a] [1:v] [1:a] [2:v] [2:a] concat=n=3:v=1:a=1 [v] [a]' -map "[a]" -map "[v]" test_filter.mp4

It supports different codecs on the input, but does re-encode everything, which may be an issue, especially if Otago Capture re-encodes it again. Audio seems fine but small/thin text may be a problem.

nstanger commented 8 years ago

Comment by nstanger Sunday Jul 17, 2016 at 22:18 GMT

Enhanced version that includes adjusting the audio volume:

ffmpeg -y -i temp_INFO321_2016-07-13_00_a.mp4 -i temp_INFO321_2016-07-13_00_b.mp4 -i temp_INFO321_2016-07-13_01_a.mp4 -filter_complex '[0:a] volume=volume=+1dB [a0]; [1:a] volume=volume=+1dB [a0j]; [2:a] volume=volume=+1dB [a1]; [0:v] [a0] [1:v] [a0j] [2:v] [a1] concat=n=3:v=1:a=1 [v] [a]' -c:a pcm_s16le -ac 1 -c:v h264 -map "[a]" -map "[v]" test_filter.mov

nstanger commented 8 years ago

Comment by nstanger Sunday Jul 17, 2016 at 22:19 GMT

Otago Capture does re-encode the video and audio, but scales down the size, which mitigates some of the visual issues anyway.

nstanger commented 8 years ago

Comment by nstanger Monday Jul 18, 2016 at 02:17 GMT

Otago Capture supports .mov as an input format. That means we can upload a file with uncompressed audio, which means that the only re-encoding event for audio will be when Otago Capture processes it. Not much we can do about the video, but the issues there are less obvious anyway.