Thirty-Two Bars of Slander

There’s a strip of little gray bars under the player on the front page — a spectrum display, thirty-two of them, bouncing along with the single. I put it there because a song about static should get to watch itself move. Mostly it minds its own business.

Then I noticed it was all bottom. Play Echoes and Static and the left edge of the display stood up like a stadium wave that never travels — three bars doing all the work, twenty-nine barely breathing. Which reads, to anyone who has ever sat in front of studio monitors at an irresponsible hour, as: your mix is bass-heavy. Mud. Too much low end. I did what you do with an accusation like that. I took it personally.

So before touching a single fader I put the record on trial. Asked my collaborator (still an AI, still being transparent about it) whether the master needed work; it subpoenaed the meter instead. Ran the actual file through the actual math, band by band, octave by octave. Verdict: innocent. The energy peaks right where warmth lives, 120 to 250 hertz. The sub territory underneath is quieter than that — four decibels quieter, no flab down there at all. From the peak, the whole record slopes off at about two and a half decibels per octave, which is more or less the shape of every record you have ever loved. Loudness sits square in the streaming pocket. There is no mud.

The meter was lying. And it was lying honestly, which is the worst kind.

A spectrum analyzer counts frequency the way a machine counts: evenly. One hertz is worth exactly as much as any other hertz. Split the audible range into thirty-two equal slices that way and the first slice runs from zero to about seven hundred hertz — the kick, the bass, and the fundamental of nearly every instrument we own, all crammed into bar number one. Everything you could whistle lands in the first three. The remaining twenty-nine divide up cymbal shimmer and air. On a display like that, every record ever mastered looks like a subwoofer demo. It wasn’t measuring our mix. It was measuring its own design, precisely.

An ear doesn’t count hertz. An ear hears octaves — doublings — the same distance from 100 to 200 as from 5,000 to 10,000. The fix was to make the bars listen the way ears do: each octave gets the same amount of screen. Now the display shows the record as it actually sits — warm at the shoulder, sloping off into the highs like tape.

Field note, filed under things the manual already knew: the tools you build to watch your own work will lie to you in the shape of their own design. The meter wasn’t broken. It measured exactly what it was built to measure, which happened to be the wrong thing, perfectly. We very nearly remastered a record to please a bar chart.

The bars are honest now. Or at least they lie in octaves, like the rest of us.

— Eric