Why Old Records Sound So Quiet

Drop an original 1987 pressing of an album next to a 2015 release on the same system and you'll reach for the volume knob. The old record isn't broken, and your ears aren't. It was simply mastered in a different economy — one where average loudness was low, peaks were left intact, and the listener was expected to supply the gain. What happened in between is a thirty-year arms race that engineers call the loudness war, and its casualties are sitting in your library right now, disguised as quiet records.

The mechanics are simple. Digital audio has a hard ceiling at 0 dBFS — push a sample above it and it clips. A recording's peak level tells you almost nothing about how loud it feels, though. Perceived loudness tracks average energy, so the trick of the war was to squeeze the space between the quiet parts and the peaks: first with compression, then with brickwall limiters that shave transients flat and let you push the whole waveform up against the ceiling. An early CD master might sit around -18 LUFS integrated with real crest factor. By the mid-2000s, major singles were shipping at -7 or even -6 LUFS, the waveform drawn as a solid brick. On paper both peak at zero. In practice one is dramatically louder, and dramatically more tiring.

The incentives were perverse and honest at once. Human hearing isn't flat — the equal-loudness contours mean that, all else equal, the louder of two tracks sounds brighter and fuller, and in any A/B comparison louder almost always wins. Labels knew this from the jukebox era, when Motown cut its 45s hot so they'd leap out of the speaker. CDs and radio raised the stakes: nobody wanted their record to sound weak when a changer shuffled it against a competitor. The casualties are famous. Oasis's Morning Glory is an early brick. The Chili Peppers' Californication distorts straight off the CD. Metallica's Death Magnetic clipped so hard that fans preferred the version bundled with Guitar Hero, which had headroom left.

What actually ended the war wasn't taste; it was normalization. Spotify, Apple Music, and YouTube now turn every track to a common playback target — roughly -14 LUFS for Spotify, -16 for Apple. Once the platform controls the gain, a crushed master buys you nothing: it just gets turned down, and at matched loudness it tends to sound smaller and more fatigued than a dynamic one. The arms race lost its reason to exist. Though the dialect survived. Plenty of modern pop and hip-hop still slams limiters past their knees, because clipped 808s and crushed drums are now a texture people want, not a defect. That's a choice. The war was the absence of one.

If you write for games or film, none of this is history — it's your delivery spec. Console platform holders publish loudness targets around -23 to -24 LUFS long-term with true-peak ceilings; streaming services for picture want dialogue-gated levels near -27 LUFS. Your score also isn't the only element in the chain: it ducks under dialogue and effects, and it may get chopped into loops and stems inside Wwise or FMOD. A composer who ships stems crushed to -8 LUFS hasn't made the game louder — they've made the audio director's mix bus a fight. Deliver clean, dynamic stems at sane levels and let the mixer decide how hard the final bus gets hit.

For curation, the lesson is sequencing and remaster skepticism. A playlist that jumps from a 1975 master to a 2015 one without gain-matching will make the older track sound like it's apologizing. Streaming normalization smooths this for listeners; if you program radio, DJ, or build catalog playlists, match levels yourself. And when a catalog gets remastered, don't assume newer is better. A louder remaster can be the same mix with the life squeezed out — the community Dynamic Range database will tell you in seconds whether a reissue kept its transients or traded them for level. Vinyl reissues sometimes get gentler masters because the format physically punishes out-of-phase bass and hot sibilance, so the 'warmth' people hear is often just headroom.

So the old records aren't quiet because their engineers lacked nerve. They were cut in a world that trusted the listener with the volume knob — still the cleanest, cheapest dynamics processor ever shipped. Give those masters the gain they were designed for and they'll surprise you. The drums hit harder when the chorus earns it. That was always the point.