My name is Matt Glenn. I am a student of music technlogy and sound engineering at the University of Michigan in Ann Arbor. Outside of class (and sometimes during) I do a ton of thinking about music and audio engineering. This blog is a my attempt at organizing my mental maelstrom.

Matt Glenn

Friday, December 26, 2008

A modern take on the noise wars of popular music

A lot of people who are not familiar with music production do not know about the noise wars, so really briefly: over the past couple decades, a lot of engineers have been mixing and mastering popular music to be louder and louder, and thus are squashing the natural dynamics of the music in order to maintain a more constantly-loud feeling. Metallica was recently humiliated when their new album "Death Magnetic". The album was mixed not only in this fashion but was terribly distorted as a result; of course they are not the first album to have this problem, but even uneducated listeners got a comparison with the guitar hero mixes of certain songs on the album, which were produced with much more clarity. Yeah, they got a lot of shit for that.

Anyway, the noise war trend seems to affect rap and hip-hop more blatantly than it does more acoustic or jazzy pop. I've been listening to a bunch of Sara Bareilles lately, and her stuff is mixed to be pretty loud, but you can still feel the dynamics. When compared to a lot of the pop that comes out nowadays, this is not a bad thing. But when I look at the level meters, the song is still pretty crushed.

My analysis: I think we as listeners have become used to sensing dynamics in popular music from other stimuli besides actual dynamic level range. I mean, when you listen to Queen's "We are the champions", the perceived jump up in dynamic right at the pre-refrain (when the full band comes in on top of the piano and bass) is an *actual* jump up in volume—and a pretty significan jump. Of course, that was the 80s—in the first verse of Sara Bareilles' "Love Song", the full band that enters after "even I know that" doesn't really pump the meters much higher. From my listening, I see a trend of engineers resorting to alternative techniques to create the illusion of dynamics. Some of the things I've noticed...
  • number of instruments - more will inherently sound louder because of association and common sense, but also because of....
  • frequency range - if a recording contains a group of instruments that balance across the full range of frequencies, then it will sound louder than a more frequency restricted group (or a smaller number of instruments). Thank Fletcher and Munson and the perceived loudness phenomenon for that.
  • reverb - there is clearly more reverb when we yell in a church than when we whisper—as such, it's possible to increase the perceived loudness of a recorded part by increasing the amount of added reverb on it
  • bus compression - movies do this all the time: when there is a loud sound effect, the other tracks often "duck" out to make room for it (whether this is done manually or with a ducker varies). In music, if a pop singer belts, for intance, sometimes the bus compression on the whole song will make it seem like the other tracks duck down a bit, making the vocal appear louder than the rest of the tracks temporarily
I'm not knocking this approach—I mean if the song needs to be as loud as possible, at least the artist's dynamic interpretation will come through *somehow*...

No comments:

Post a Comment