The Loudness War (Why Old Records Were So Loud)

It's Tuesday, September first. On Sunday night I went to an EDM show, and I had this thought while standing in the crowd: I have no idea if this guy is actually doing anything up there. He could be playing a Spotify playlist. The lights were incredible, the bass was shaking my chest, but watching him twist knobs and tap buttons, I couldn't tell you what was live performance and what was theater.

That got me thinking about loudness. In a club, louder is better. You want the drop to hit like a wall. And that instinct — make it louder than the last guy — turns out to be the exact same instinct that broke recorded music for thirty years.

This is the story of the loudness war: how a perfectly rational incentives problem caused albums to get objectively worse decade after decade, how Metallica accidentally proved it with a video game, and how the war was finally ended not by artists, not by critics, not by angry fans, but by an algorithm nobody even thought of as a solution. The loudness war didn't start with digital audio. It started in the 1940s, in bars and diners, with jukeboxes.

A jukebox owner would set the volume knob once and leave it. Every record played at the same amplifier level. So if you were a producer and you wanted your song to jump out of the speaker and grab people by the collar, you didn't need a better song. You needed a louder record. Cut the single hotter than the competition, and when your track came on, it would physically hit harder. Nobody was touching the volume knob between songs. The louder record won.

Motown was notorious for this. Their engineers pushed 7-inch singles to the absolute limit of what vinyl could handle. One Motown engineer later admitted they were cutting "some of the hottest 45s in the industry." The strategy worked. Motown dominated jukeboxes, and from jukeboxes, radio.

Radio created the same incentive. Program directors at radio stations would audition stacks of singles. If your record was even slightly quieter than the one before it, it sounded weak. Producers learned fast: you mastered as loud as the vinyl grooves would allow.

But vinyl had a physical limit. Push the grooves too hot and the needle would literally jump out of the record. In 1969, mastering engineer Bob Ludwig cut a version of Led Zeppelin II that was so loud it caused some turntable cartridges to mistrack. The album had to be recalled and reissued at a lower level. The medium itself was the safety brake.

Then came the compact disc, and the brake came off. A CD can hold audio up to exactly zero decibels — digital full scale. You cannot go above zero. Hit zero and the waveform flatlines. It clips. Your music turns into a square wave, which is what distortion sounds like.

The obvious solution would be: don't hit zero. Leave some space. But that's not what happened. Instead, mastering engineers figured out that you could use a tool called a brickwall limiter to push the average loudness of a song higher and higher while keeping the peaks just under zero. The tradeoff was dynamic range — the difference between the quiet parts and the loud parts.

Think of it like this: a song has quiet verses and loud choruses. That's dynamic range. That's what makes the chorus feel like a release. A brickwall limiter squashes that difference. The verses get louder, the choruses stay the same, and eventually the whole song is just one flat, blaring block of sound. The waveform stops looking like a mountain range and starts looking like a brick.

You can see it. That's what makes this story unusual for an audio topic — there's a visual proof point. Pull up a waveform of any album from the 1980s and compare it to the remastered version from the 2000s. The 1980s version has peaks and valleys, quiet moments and explosions. The remaster is a solid rectangle from left to right, every second pushed to the ceiling.

In 1987, Guns N' Roses' Appetite for Destruction averaged about -15 dB RMS — that's the measure of average loudness. By 1994, the average rock record was at -12 dB. Then in 1995, Oasis released (What's the Story) Morning Glory? and hit -8 dB on multiple tracks. That's an enormous jump. Ian Shepherd, a mastering engineer who's been tracking this for decades, has called it the moment the CD loudness war got serious. The album sold over twenty-two million copies worldwide and nobody seemed to care that it was compressed into a brick. The precedent was set: loud albums sell. By 2008, the loudness war had a body count. Red Hot Chili Peppers' Californication, released in 1999, was so heavily compressed that you can hear audible distortion across the entire album — not artistic distortion, just ugly digital clipping. Fans had been complaining for years. But complaints don't change incentives. If a loud album is what gets programmed director attention, then loud is what gets made.

Then Metallica released Death Magnetic.

Death Magnetic was supposed to be Metallica's return to form — their first album with producer Rick Rubin, their first with bassist Robert Trujillo, a deliberate step back toward the thrash metal of their early years. The songs were well-received. The sound was not.

The CD master of Death Magnetic was compressed so aggressively that it was physically painful to listen to at any volume. The dynamic range had been almost entirely eliminated. Every instrument was fighting for space at the zero-decibel ceiling, creating a wall of distortion that made the songs sound like a blown speaker.

But here's where it gets incredible. Death Magnetic was also released as downloadable content for Guitar Hero III. The Guitar Hero version needed to be delivered as separate instrument tracks — what are called stems — so the game could raise or lower each instrument depending on what the player was doing. Those stems were sent to the game before the final mastering process.

In other words, the Guitar Hero version had not been brickwalled.

And fans noticed immediately. The video game version of the album sounded objectively better than the CD they paid for. It was ten decibels quieter and had actual dynamic range. You could hear the drums breathe. The guitars had space. When the band dropped into a heavy riff, it actually felt heavy instead of just feeling loud.

A petition appeared online demanding Metallica remix the album. Ted Jensen, the mastering engineer credited on Death Magnetic, posted on a message board: "Believe me I'm not proud to be associated with this one, and we can only hope that some good will come from this in some form of backlash against volume above all else."

Jensen also clarified that the problem wasn't at the mastering stage — the mixes were already crushed when they arrived at his studio. By 2008, the loudness war had become so embedded in the production pipeline that a song could be destroyed before it ever reached the mastering engineer. For years, audiophiles, critics, and even some artists had been begging for an end to the loudness war. Nothing worked. The incentives were too clean: louder records got noticed on radio, in playlists, in clubs, in shuffle. If you mastered a dynamic album, it sounded quiet and weak next to a brickwalled competitor. The market punished restraint.

Then the market changed. Not because of a consumer revolt or an artistic manifesto, but because of something boring: streaming services started normalizing loudness.

Spotify, Apple Music, YouTube, Tidal — they all apply loudness normalization to everything they play. The standard is measured in LUFS, Loudness Units relative to Full Scale. Spotify targets -14 LUFS. Apple Music uses -16 LUFS. YouTube sits around -14.

Here's what normalization does: if you upload a track mastered to be extremely loud, the streaming service turns it down. If you upload a track with natural dynamics, the service leaves it alone or turns it up slightly. The result is that every song plays back at roughly the same perceived volume regardless of how it was mastered.

This completely destroyed the incentive to brickwall. You can't "win" on a streaming platform by being louder. The algorithm makes you the same volume as everybody else. So if you compress your track to death, all you're doing is removing the dynamics that make music feel alive — for no competitive benefit whatsoever. You're making your song sound worse, and the streaming service is turning it down anyway.

Mastering engineers have publicly acknowledged the shift. They now master differently for streaming than they did for CD. Ian Shepherd, the same engineer who documented Metallica's Guitar Hero controversy, launched a campaign called Dynamic Range Day to celebrate the return of dynamics to recorded music. In interviews, producer after producer has said they now aim for a natural, dynamic master because there's simply no reason not to.

The loudness war lasted roughly thirty years. It damaged thousands of recordings. It made people's ears tired without them knowing why. No legislation stopped it, no industry standards body stopped it, no artist boycott stopped it. It was ended by a volume knob in the cloud — a few lines of code in a server farm somewhere that quietly made sure every song played at the same level, and in doing so, removed the only reason anyone had to ruin their own music. I still don't know if that DJ on Sunday was actually mixing or just pressing play. But I do know that the music coming out of those speakers was probably mastered to sound good at full volume in a club, which is a completely different target than sounding good on headphones. And the fact that the same song might be mastered differently for those two environments — that's progress. That's a music industry that finally learned you don't have to destroy your song to get it heard.

The loudness war was a classic tragedy of the commons: everyone acting rationally in their own interest produced a collectively worse result. It took thirty years and a complete restructuring of how we listen to music to fix it. If you ever want to hear what we lost, pull up the CD version of Death Magnetic and the Guitar Hero version back to back. One sounds like music. The other sounds like a mistake.

Stay informed, stay curious, and we'll see you tomorrow.