As a soundie a word of defense for my fellow craftspeople - there are several possible reasons contributing to this issue. Actors may want to whisper and mumble. On set there can be challenges with unwanted noise like traffic, or rustling wardrobe causing issues to the lavalier mics hidden under the clothes - if there is no budget or time to re-record (called ADR) problematic lines, the audio editors have to clean it up as best they can and this can lead to end results where also the desired audio signal is degraded.

When it gets to the mixing stage there might not be a lot of time, and a director or producer/studio may be pushing for the flashier or atmospheric sounds to be played louder. It can be hard to argue against this as the mixer will of course also want to land the next gig. And the sound mixer might also want to create dynamic mixes for the cinema etc - sometimes a "DVD mix" is also made with less dynamic range but this costs more money of course.

After mixing and delivery we still have more possible issues since the material is viewed through one of many streaming services with an endless combination of TVs, laptops, amplifiers and speakers, some of which may have their own varieties of automatic adjustment enabled, with the user not even realizing.

It is annoying, of course - I am used to subtitles myself and don't mind having them on if I need to view at a lower volume but I myself wish services would also include English subs without the closed captioning of sound effects etc.