https://trac.ffmpeg.org/wiki/AudioVolume#LoudnessNormalization
However, it is recommended to run the normalization with two passes, extracting the measured values from the first run, then using the values in a second run with linear normalization enabled.
ffmpeg -i UvL8OsG75f0.opus -filter:a loudnorm -pass 1 -f null /dev/null
ffmpeg -i UvL8OsG75f0.opus -filter:a loudnorm -pass 2 tmp.opus
The first command generated no "ffmpeg2pass-..." logs at all.
How to normalize the volume in a two-pass approach properly?