Steps to reproduce

1. Upgrade to avprobe 12.3 on Mac osX
2. Try to run AudioSegment.from_file() on an audio file

Expected behavior

AudioSegment should return segment data based on avprobe output

Actual behavior

AudioSegment hits a KeyError

  File "/dejavu/dejavu/decoder.py", line 51, in read
    audiofile = AudioSegment.from_file(filename)
  File "/dejavu/.venv/lib/python3.6/site-packages/pydub/audio_segment.py", line 664, in from_file
    if (audio_streams[0]['sample_fmt'] == 'fltp' and
KeyError: 'sample_fmt'

the new version of avprobe has added a line break to the line containing the sample_fmt.

Avprobe output used to look like this (partial output)

Input #0, mp3, from '/audio/track.mp3':
  Duration: 00:00:33.70, start: 0.000000, bitrate: 232 kb/s
    Stream #0:0: Audio: vorbis, 44100 Hz, stereo, fltp, 320 kb/s

but now looks like this (note the additional line break)

Input #0, mp3, from '/audio/track.mp3':
  Duration: 00:00:33.70, start: 0.000000, bitrate: 232 kb/s
    Stream #0:0: Audio: vorbis
      44100 Hz, stereo, fltp, 320 kb/s

so the parsing of this line https://github.com/markkohdev/pydub/blob/master/pydub/utils.py#L262 no longer works, consequently not setting sample_fmt on this line https://github.com/markkohdev/pydub/blob/master/pydub/utils.py#L291

Thus, this key does not exist and this line fails
https://github.com/markkohdev/pydub/blob/master/pydub/audio_segment.py#L664

The solution to this seems to be to update the utils file regexes to handle the new format, or to use a safe .get('sample_fmt', '<default>') in the audio segmenter

Your System configuration

• Python version: 3.6
• Pydub version: 0.22.0
• ffmpeg or avlib?: avlib
• ffmpeg/avlib version: 12.3

Is there an audio file you can include to help us reproduce?

schubert.ogg.zip