normalizing audio

i use ffmpeg to normalize the audio of my media files, particularly anything i want to sentence mine. i also like to normalize my anki media folder every month or so. be sure you’re on an up-to-date build of ffmpeg to use the following methods. (see updating ffmpeg for how i updated ffmpeg on pop os, which should work on other debian based distributions)

batch normalize audio files

to batch normalize audio files, install ffmpeg-normalize using pip install ffmpeg-normalize. inside the folder, run”

ffmpeg-normalize *.mp3 -c:a libmp3lame -ext mp3 -pr

mp3gain can be used change gain without the need to reencode the track. i like to use this in my anki media folder since it is much faster than ffmpeg-normalize. the link has downloads for windows, but on linux it can be installed from your distro’s repository. to run from the command line:

mp3gain -r -k -s r *.mp3

normalize video audio

for a single video:

ffmpeg -i infile.mp4 -filter:a speechnorm,loudnorm -c:a ac3 -c:v copy out.mp4 

to run the above command on a whole folder in linux or mac, use:

for i in *.mkv; do ffmpeg -i "$i" -filter:a speechnorm,loudnorm -c:a ac3 -c:v copy "normalized/$i"; done

or

ffmpeg-normalize *.mkv -c:a ac3 --dynamic --progress

the normalized videos will have the same name but will be in the subfolder titled “normalized”