Definition
The loudness war is the trend, which grew between the 1990s and the early 2010s, of mastering records louder and louder thanks to digital limiters and compressors, sacrificing dynamics. The idea was simple: at the same system volume, a louder track “sounds better” on first listen (equal-loudness contours make what is louder seem fuller) and stands out on the radio and in jukeboxes.
A brief history
- With vinyl and radio, volume was limited by the physics of the medium and by radio processors.
- With the CD (from 1982) the limit becomes 0 dBFS: from the 1990s, thanks to digital look-ahead limiters, the average loudness of CDs rises by many dB.
- Albums often cited as examples of heavily compressed masters: Oasis, “(What's the Story) Morning Glory?” (1995); Red Hot Chili Peppers, “Californication” (1999); Metallica, “Death Magnetic” (2008), whose master with audible clipping provoked fan protests (fans noticed the less compressed version prepared for the video game Guitar Hero).
- Counter-initiatives: the Pleasurize Music Foundation (with the DR meter), the “Dynamic Range Day” promoted by engineer Ian Shepherd, the EBU R128 standard in broadcasting (2010), and above all loudness normalization on streaming platforms.
Why normalization changed the rules
Spotify, YouTube, Apple Music, Tidal, Amazon, and Deezer turn down tracks that are louder than their target (around −14 LUFS for most, −16 for Apple Music). A master at −7 LUFS is turned down by about 7 dB on Spotify: it ends up at the same volume as a master at −14, but with less dynamics and smaller transients. In a direct comparison in a playlist, the dynamic master often sounds more powerful. The advantage of the “loud” master remains only where there is no normalization (some DJ sets, some players, the Spotify web player according to Spotify itself).
Perceived dynamics
Average loudness isn't all that matters, contrast does too:
- Macrodynamics: the difference in energy between sections (verse smaller than chorus). It is built mainly in the arrangement and mix.
- Microdynamics: the difference between the transient and the body of each hit (punch). Excessive limiting flattens it.
- Density: a dense, saturated mix can seem loud even at moderate loudness.
- Measurements: PLR (peak/loudness), LRA (loudness range), crest factor. A PLR of 8–10 dB in modern pop is common; below 6 dB punch suffers.
How to get “loud and good”
- A clear arrangement, without buildup in the low mids (loudness gets “eaten” there).
- A mix with controlled lows and defined transients.
- Light saturation and clipping spread around (on drums, bus, master) instead of a single limiter working 8 dB.
- A limiter working 2–4 dB; true peak −1 dBTP.
- Always compare at equal loudness (see “Review and A/B Comparison”).
Example
Two masters of the same pop track in A minor: A at −7 LUFS integrated (PLR 6 dB), B at −10 LUFS (PLR 9.5 dB). On Spotify both play back around −14 LUFS: A turned down by 7 dB, B by 4 dB. In the chorus B has a more incisive kick and a more “open” vocal; A sounds tighter but smaller. In a club without normalization A is louder at the same mixer volume, but a DJ simply turns up B's channel.
Common mistakes
- Chasing the loudness of references without considering the quality of their mix.
- Comparing masters at different loudness.
- Thinking −14 LUFS is a mandatory target: it's the point where the platform brings the track, not a binding mastering target.
Exercise
Master one of your tracks at −14, −10, and −7 LUFS integrated. Then normalize the three files to −14 LUFS with a gain plugin and compare them blind. Which do you prefer? Why?
Links
“Loudness, LUFS, and Normalization”, “Limiting, Clipping, and True Peak”, “Dynamics in Mastering”, “Mastering for Streaming”, “Psychoacoustics in the Mix”.