Why −14 LUFS is not the right target for every song

Should you master every song to −14 LUFS? Spotify recommends that level for delivery, but it also accepts louder masters. The useful question is what happens to your song when playback turns it down.

Take a master that measures −9 LUFS. With normal track normalization, Spotify turns it down by about 5 dB. The compression and limiting are still there: quieter playback does not bring back the drum attacks or dynamics changed during mastering.

What Spotify actually does

Spotify’s artist guidance describes a normal playback level of −14 LUFS, measured under ITU 1770. Album playback preserves the level differences between tracks; shuffle and mixed playlists use individual track levels. Premium listeners can choose Loud (−11), Normal (−14) or Quiet (−19). Loud can apply a limiter to softer, dynamic recordings. The web player and some third-party devices do not use normalization.

For delivery, Spotify recommends −14 LUFS and a true peak below −1 dBTP. For louder masters, it recommends below −2 dBTP to leave more encoding headroom. It asks for one high-quality stereo master per track. These are Spotify’s recommendations; the listening trade-offs below explain why a producer might choose a different loudness.

Loudness still changes the sound

Compare two versions of the same mix at matched loudness. The more heavily limited version may have longer sustain, but the kick may hit less sharply. A lighter pass may keep more space between the attacks. Listen to the chorus and the quieter sections before deciding which version suits the song.

The eleven purchased masters in our August 2026 test measured −11.4 to −7.3 LUFS, with loudness ranges of 3.5 to 12.3 LU. They included sparse pop, hip-hop, acoustic music and band recordings. Those measurements describe this small set of releases. They do not establish an ideal target for another track.

How PolishMyTrack chooses a starting level

PolishMyTrack chooses a loudness target together with a true-peak ceiling. In our five-source test, adaptive Auto selected targets from roughly the low −12s to high −11s LUFS; the resulting PCM stayed at or below −2 dBTP. Already-loud commercial masters received constant downward gain, with no limiting or tonal processing. Their displayed LRA stayed unchanged.

Sometimes the peak ceiling prevents a master from reaching the requested loudness. The report shows the measured result so you can see that difference. Pushing harder just to hit the target can change the sound you wanted to keep.

Listen after matching the levels

Check how the master is likely to play after normalization, then listen at matched loudness. If the extra limiting makes the drums smaller or the vocal harder to listen to, back it off. If you prefer the denser version at equal volume, that is a useful reason to keep it.