ffmpeg: split video and extract audio

This is the command to extract a piece from a video file in the most efficient manner:

ffmpeg -ss $START -i "My 21st Birthday" -t $DURATION -c:v copy -c:a copy "Funny moment.webm"

Where

• $START is the start moment, in hh:mm:ss[.mmm] format

• $DURATION is the length in the same format as above

Note that in this case the order of parameters matters! If the input file is given before -ss, ffmpeg will still decode the streams from the beginning of the file. Instead, if -ss is given first, the streams are first seek’d, and only after the start point decoding begins.

The -c switch is parameterized with a stream index, which is a number like 0, 1, 2. In this case -c:v is a special syntax for all video streams. Following the swich, one should name an explicit decoder; in this case, the special word copy is used to keep the source format instead of decoding/encoding again.

To extract the audio stream from a video file, one can use

ffmpeg -i "my video.webm" -map 0:a -c:a copy "soundtrack.ogg"

The -map switch explained:

•  the first number is a 0 (zero) which identifies the source file id. In this case only one file is given, so its index will be zero

• the second character is a stream identifier. Again, here we use the special character a to identify all audio streams

• -c:a copy has the same meaning as above: keep the audio encoding

This little Ruby function can help to compute the $DURATION parameter

# first argument is start time, second is end time
def time_diff(start, end)
  t = Time.at( Time.parse(end) - Time.parse(start) )
  (t - t.gmt_offset).strftime("%H:%M:%S.%L")
end