Earlier I took a quick look at when the loudness wars started in recorded music, and took a stab at why they began once the crime rate started falling. However, now that I've found a very clear pattern of chiaroscuro tracking the homicide rate in visual art, I wonder if there are parallels in music.
Light must reflect off a surface for us to see it, so the equivalent in sound is volume, the most basic feature for us to hear it. Bright vs. dark is like loud vs. quiet.
A key part of the loudness wars that I didn't touch on earlier is that it's not simply that music recordings today are on average louder. It's that they achieve this mainly by taking the sounds that are supposed to be softer and quieter in the song, and jack up the volume on them. The result is a song with much more uniform volume -- uniformly loud -- instead of one that has greater "dynamic range," or the difference, variation, contrast between the louder and quieter sounds.
That seems like the more important part of the change -- loss of dramatic tension or contrast that comes from hearing a loud passage break in after a softer one, or the coming-down feeling of a soft passage following a loud one. As detailed in the post on chiaroscuro, there are reasons why people in rising-crime times prefer more theatrical presentation, and falling-crime people prefer more sameness to the presentation.
Here is a great review of these changes in dynamic range, plus an illustration comparing an original recording to two later remastered versions. (Only the first pair of links at the end of his article works for me.) You can hear what should be softer parts, whether vocal or instrumental, getting a lot louder in the remastered ones.
So, just as our visual culture has gotten more bland from using more uniform lighting, our musical culture has become drier from more uniform loudness. And as that article demonstrates, it's happening even to re-issues of music that was originally recorded with a wide range of quiet to loud sounds.
Unless you have multiple copies of a CD, you may not even notice that the newer remasters tend to sound louder, more distorted, and not as full of rich contrast. I had a suspicion about it, but it wasn't even based on comparing the same album, just noticing that any remaster was likely to sound that way, whether I could compare it to the original or not. Reading more about dynamic range compression, I'm convinced.
Now before I get a CD, I do a quick look-up on the neat DR Database to see if there are multiple releases, and if so, which ones will sound nice and which will sound loud and garbled. Not all newer releases sound bad; this quick check gives them a fair chance. You can also see whether you should swap out the copy you have for a better one. It may seem odd that earlier presses of an album sound better than newer ones, but remember also that CD players were high-end stuff back then too, not cheap junk.
Playing around with that DR Database, it seems like 1993 is close to the ground zero of the loudness wars. Find any band who's released CDs from the start in 1983 up through today -- including CDs of music they originally recorded in the '60s or '70s, but not released on CD until later. Click the "Year" column header to line them up in chronological order. Sometime around '93 or '94, it's pretty reliable that the sound quality will start going to hell, bottoming out in the 2000s and this decade.
That's another way in which the early '90s weren't so bad -- studio production still sounded pretty slick, even for grunge albums like Nevermind.