Ffmpeg Sync when extracting parts of video

Question

This question can be seen as an extension of this question. I am using the ffmpeg 2.0.1 .

The following commands is useful when you want to extract parts of one big video file given time makers.

ffmpeg -i 00402.MTS -s 1280x720 -ss 0:00:05.38608 -vcodec libxvid -qscale:v 2 -acodec ac3 -ar 48000 -ab 256k -t 0:00:06.61391 m001.avi

My problem is that in the ouput file, the start value is 0 and the there is ~50ms of shift between audio and video. What argument should I use to take into account this 'start value' when extracting severals parts of the big video file ('.mts') ? What does this 'start value' mean ?

ffprobe output from big MTS video file

ffprobe m001.MTS 
ffprobe version 2.0.1 Copyright (c) 2007-2013 the FFmpeg developers built on Sep 24 2013 05:31:18 with gcc 4.8 (Debian 4.8.1-10)
configuration: --extra-cflags=-I../static/include --extra-ldflags='-L../static/lib -static' --enable-gpl --enable-version3 --enable-static --disable-shared --disable-debug --enable-runtime-cpudetect --enable-libmp3lame --enable-libx264 --enable-libspeex --enable-libvorbis --enable-libvpx --enable-libfreetype --enable-libxvid --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libtheora --enable-libvo-aacenc --enable-libvo-amrwbenc --enable-gray --enable-libopenjpeg --disable-ffserver
libavutil      52. 38.100 / 52. 38.100
libavcodec     55. 18.102 / 55. 18.102
libavformat    55. 12.100 / 55. 12.100
libavdevice    55.  3.100 / 55.  3.100
libavfilter     3. 79.101 /  3. 79.101
libswscale      2.  3.100 /  2.  3.100
libswresample   0. 17.102 /  0. 17.102
libpostproc    52.  3.100 / 52.  3.100
Input #0, mpegts, from 'm001.MTS':
Duration: 00:01:18.25, start: 0.455556, bitrate: 12217 kb/s
Program 1 
Stream #0:0[0x1011]: Video: h264 (High) (HDMV / 0x564D4448), yuv420p, 1440x1080 [SAR 4:3       DAR 16:9], 25 fps, 50 tbr, 90k tbn, 50 tbc
Stream #0:1[0x1100]: Audio: ac3 (AC-3 / 0x332D4341), 48000 Hz, stereo, fltp, 256 kb/s

ffprobe ouput from extract sample

Input #0, avi, from 'm001.avi':
Metadata:
encoder         : Lavf55.12.100
Duration: 00:00:06.64, start: 0.000000, bitrate: 4950 kb/s
Stream #0:0: Video: mpeg4 (Simple Profile) (xvid / 0x64697678), yuv420p, 1280x720 [SAR 1:1 DAR 16:9], 25 tbr, 25 tbn, 25 tbc
Stream #0:1: Audio: ac3 ([0] [0][0] / 0x2000), 48000 Hz, mono, fltp, 256 kb/s

Answer 1

I've come across this problem before, and the only solution is to use itsoffset and map.

ffmpeg -i 00402.MTS -itsoffset 0.455556 -i 00402.MTS -map 1:0 -map 0:1 -s 1280x720 -ss 0:00:05.38608 -vcodec libxvid -qscale:v 2 -acodec ac3 -ar 48000 -ab 256k -t 0:00:06.61391 m001.avi

This page has a nice explanation why and how it works.

http://www.kkoncepts.net/blog/fixing-out-sync-audio-and-video-ffmpeg
(No longer available at original link; now from web-beta.archive.org)

EDIT There are two main tools here to be aware of: 1) itsoffset 2) map

itsoffset exists to offset a stream that is out of sync, whereas map informs ffmpeg to use certain streams from specific inputs and apply them to the output.

In this example, we are using the same input twice, mapping the streams thusly: We use the video stream of first input (which seems to be starting at 00:00:00.00) We use the second input as the audio stream, but offset the input time 0.455556 seconds.

From the above mentioned site:

Here "-map 1:0" tells ffmpeg to use the second input file (1) and take its first stream (0) as the video source. Since that's the input that follows the "-itsoffset" switch, that's the stream that will be delayed. Then "-map 0.1" tells it to use the second stream from the first input as the audio. Since there's no offset applied to the first input file, it will be processed normally.