Compressed music might be harmful to the ears

Pop songs for example, are heavily compressed because "compressed" music sounds "better" on cheap external speakers with bad dynamic range. Kinda like increasing the saturation on a cheap TFT panel.

From this, what I would take away is not necessarily that compression on its own is harmful but that... there's almost certainly more to ear damage than just the loudness, and compression is one way this can exhibit through. So I'd say the title and angle of the article is a bit misleading.

Naturally, when I say compression I mean it in the audio effect sense, as they do in the article.

On top of that another problem with compression is that it is not neutral, bad compressors, especially in the digital domain can introduce aliasing.

Also search for the "loudness war". This is how we called the problem back in the mid 2000's.

Beyond any potential health effects, DRC also just sounds bad after listening for more than a few seconds. It kills the liveliness of the audio and the signal-to-noise ratio.

So why is it used? Loudness is like candy to your ears, we are drawn to it and people tend to perceive louder music as sounding better, even if it’s exactly the same except for a change to the volume knob. DRC+MG is a bit like turning up the volume knob on your behalf, except it’s worse in almost every way.

One legitimate use case for DRC+MG, and in fact one of the original justifications for its widespread use, was the rise of portable music, where you want to be able to hear all of the instruments even though you are in a loud environment like the subway, for example. But this devolved into the “Loudness Wars”.

Tasteful use of DRC, perhaps without much MG, can also help “glue” an instrument within an audio mix. I might use it on a bass guitar, for example, to keep its volume under control so it stays in the background, providing rhythm, and doesn’t accidentally become the lead.

Ideally, the base format for all audio would have little to no DRC, but there would be other versions or built-in metadata to help playback devices adjust the listening experience to the environment by enabling DRC when it makes sense to do so or if the user wants it. Just as movies can be watched with or without HDR and subtitles. Unfortunately, the audio world hasn’t progressed much in that direction.

Put very simply, it increases the amplitude of the input sound. Its parameters are controlled by internal timers, external triggers, timers and signals ("side-chaining"). For example; lowering the bass ("ducking") during the critical milliseconds of the kick-drum so that the volume of both can stay maximised while their sum stays within maximum amplitude.

On voice-over, in radio and podcasts it makes voices sound even and "boomy/strong".

Originally probably necessary way to make recordings listenable in cars and noisy environments with lower spec speakers.

Nowadays producers use it to "sound louder" and thus make stronger impressions.

It's much easier to deliver reasonably dynamic masters nowadays than, say, 10 years ago, but the fight is still not over. Loudness normalization is still inconsistent between streaming platforms and one lagging/idiosyncratic actor (usually some smaller, locally-important one) can spoil the impression, breaking uniform listening experience, and force an author to reconsider what loudness and dynamic range they aim for.