ffprobe

The normal text output of ffprobe goes to stderr, so you usually need to add 2>&1 to parse the output. The alternative is to use the complex ffprobe options to format the output which then goes to stdout.

Start with a look at this page: https://trac.ffmpeg.org/wiki/FFprobeTips

ffprobe -v quiet "$in" -print_format json -show_entries stream_tags:format_tags

(for -print_format, there is also flat, xml, csv, etc.)

ffprobe -v warning -select_streams v -show_entries stream=codec_name,height,width,pix_fmt -of csv=p=0 "$in"

https://stackoverflow.com/questions/69029439/a-good-way-to-detect-alpha-channel-in-a-video-using-ffmpeg-ffprobe

for f in *.mov; do
  pixfmt=$(ffprobe -v 0 -select_streams v:0 -show_entries stream=pix_fmt -of compact=p=0:nk=1 "$f")
  alpha=$(ffprobe -v 0 -show_entries pixel_format=name:flags=alpha -of compact=p=0  | grep "$pixfmt|" | grep -oP "(?<=alpha=)\d")
  if (( alpha )); then echo "$f : yes, has alpha"; else echo "$f : NO"; fi
done

ffprobe "$in" 2>&1 | perl -nle '/^\s+Stream #(\d+:\d+).*?: Audio/ && print $1'

with codec and number of channels:

ffprobe "$in" 2>&1 | perl -nle '/^\s+Stream #(\d+:\d+).*?: Audio: (\S+).*(?:(\d+)\s+channels)?/ && print join("\t", $1, $2, $3)'

or to have the tracks in a Bash array:

audio_tracks=( $(ffprobe "$in" 2>&1 | perl -nle '/^\s+Stream #(\d+:\d+): Audio/ && print $1') )

or the "official" way:

ffprobe -v quiet -select_streams a -show_entries stream=index -of csv=p=0 "$in"

ffprobe -v quiet -select_streams a -show_entries stream=index,codec_name,channels -of csv=p=0 "$in"

audio_tracks=( $(ffprobe -v quiet -select_streams a -show_entries stream=index -of csv=p=0 "$in") )

ffprobe -v quiet -select_streams s -show_entries stream=index,codec_name:stream_tags=language -of csv=p=0 "$in"

(but for mkv, better see mkv)

For example to detect silent tracks/channels

ffmpeg -i "$in" -map 0:a -af "astats=measure_overall=none:measure_perchannel=Max_level" -f null -

ffmpeg -i "$in" -map 0:a -af "astats=measure_overall=none:measure_perchannel=Max_level" -f null - 2>&1 | grep -E '(Channel|Max level):'

ffmpeg -i "$in" -vn -sn -dn -map 0:a -af "astats=measure_overall=none" -f null - 2>&1 \
| perl -nle 'if (($id, $var, $val)=/^\[Parsed_astats_0 \@ 0x([\da-f]+)] (Channel|Max level): ([\d\.]+)/) { if ($var=~/Channel/) { if (!$t || $id ne $h{$t}{id}) {$t++; $h{$t}{id}=$id;} $c=$val;} elsif ($var=~/level/) {print "$t.$c $var = $val";} }'

Or with volumedetect, but it mixes all channels into a single value

ffmpeg -i "$in" -map 0:a -af "volumedetect" -f null - 2>&1 | grep max_volume

for track in $(ffprobe -v error -select_streams a -show_entries stream=index -of csv=p=0 "$in"); do \
  ffmpeg -i "$in" -vn -map 0:$track -c:a copy -f md5 - 2>/dev/null \
  | while read md5; do \
      printf "%32s  Audio Track %2d %s\n" $md5 $track "$in"; \
    done; \
done

ffprobe -v error -show_entries stream=index,codec_name:stream_tags=timecode -of csv=p=0 "$in"

-sections
    Print sections structure and section information, and exit. The output is not meant to be parsed by a machine.

-show_data_hash algorithm
    Show a hash of payload data, for packets with -show_packets and for codec extradata with -show_streams.