"Dynamic range compression" - reduce the difference between loudest and quietest sound

Some people are very in-character when narrating so they end up screaming, and sometimes they whisper. This can be pretty annoying, especially when I'm trying to sleep. Either I don't hear things, or I get a heart attack by someone screaming.

I’d love a feature to make the sound more consistent so the quieter sounds are louder, and the louder sounds are quieter. This seem to be called dynamic range compression. One might need to do something else too, to e.g. reduce/remove the sound of breathing.