Mastering Volume: Why Louder Isn't Always Better

It seems that every year, major label releases keep getting louder and louder.

Try this: go to your CD collection and select three albums. Find one that was released in the late 80's. Find one that came out in the mid-90's. Now, pick one that came out this year. Make sure these are real, store bought CD's and not copies from friends or mp3 clones (those won't fully illustrate this point).

The albums I picked are Toto IV, Jeff Buckley's Grace, and Train's For Me, It's You. Now, set the volume on your sound system to its middle setting. Not too loud, not too soft. Make sure the sound comfortably fills the room. When I put on Toto, things sound pretty good. When I switch to Grace, the CD seems louder. Now, when I switch to the Train CD, it sounds like my speakers are going to fly out of the room. It sounds almost twice as loud as the Toto CD.

What's up with that? What you are hearing is a difference in average volume levels. The technical term for this is RMS level (root-mean-square). Over the past 20 years, as CD technology has evolved, records have steadily become louder and louder. The RMS level of the Toto album is about -20 dB. The RMS level of the Jeff Buckley is about -10 dB. The RMS level of the Train is about -6 dB.

What does that mean? Because albums are getting louder and louder, dynamics and depth are becoming sacrificed. The main reason this has happened is because a louder album will stand out more when compared to a quieter album. The day major labels learned about this is the day that the volume wars began. Ricky Martin's last single topped the RMS charts with a ear-shattering RMS level of -3 dB.

This phenomenon is what mastering engineer Bob Katz calls "hypercompression." A little master bus compression can help glue a mix together, but jamming all of the life out of it so that it can compete with today's winners doesn't make a mix better. It degrades the quality of a master. It's a big no-no.

As a devotee of the study of classic albums, a quieter, well-produced album always sounds better to me than a crapped out hyper-compressed modern record. In spite of all of the hypercompression, I don't think it's possible for an album to rock harder than AC/DC's Back in Black. And, that's not a loud album.

Recently, I was finishing up John Gilllespie's album Endless Summer. I recorded and mixed the album at my studio. I also offered to master the project. I tried three different approaches to the final master. First, I tried a completely uncompressed version. I used an over-easy limiter to help match the levels. Second, I tried a hypercompressed version, taxed to the max. Third, I tried a version with bus compression on the final mixes with a brickwall limter set with the threshold to no more than -3 dB into the meters. The third version outperformed the other two on just about every system I tried. The third master is much more present than the first version and has more depth and dynamics than the second version.

The lesson learned is a simple one. Always, always trust your ears.

When you are finishing your band's next CD, I'm sure loudness will be a concern of everyone's. Just know that it's OK if your album isn't as loud as someone elses. Dynamics are a good thing.

Also, because most people use mp3 interfaces to listen to audio, these systems have built in "Sound Check" settings that play everything at the same levels. That is, assuming, that people will actually purchase your CD and not just rip it from a friend. If friend A's mp3 encoding is set to 128 kbps, it really doesn't matter if the CD has been mastered. The audio is already degraded to the point where RMS volume is the least of your problems.

The only time I would recommend jamming the average volume as high as it can go is when you are giving a demo to an A&R at a label. This will help it stand among the other hypercompressed stuff that's on her desk. Most A&R pass on 99% of the material they hear, so unless your feeling lucky, you might want to back it off a bit.