How loud should a master be for Spotify and Apple Music

For about twenty years, mastering engineers were asked for one thing above all others: make it louder than the last record. The logic was sound at the time. On a CD or a radio playlist, the louder track genuinely did sound bigger and more exciting in a direct comparison, and nobody was adjusting the volume between songs.

Streaming ended that. Spotify, Apple Music, YouTube, Tidal and Amazon all normalise playback: they measure how loud a track actually is and turn it up or down so everything in a playlist sits at roughly the same level. If your master is crushed to be very loud, the platform simply turns it down — and what you are left with is a quiet version of a squashed record, next to someone else's open-sounding one at the same volume.

What LUFS means, in plain terms

LUFS measures perceived loudness across a whole track, the way an ear hears it, rather than the peak value your DAW meter shows. A song at -14 LUFS integrated sounds about as loud as any other song at -14 LUFS, regardless of genre. Peak metering tells you nothing about this: a track can peak at 0 dBFS and still be quiet, or peak at -3 and feel enormous.

Where the platforms sit

• Spotify normalises to about -14 LUFS by default, and does not turn quieter tracks up beyond a limit unless the listener enables loud mode.
• Apple Music uses Sound Check at a similar target, around -16 LUFS.
• YouTube lands near -14 LUFS.
• Tidal and Amazon are in the same neighbourhood.
• The exact numbers change from time to time; the principle does not.

So what should you aim for?

Somewhere between -14 and -9 LUFS integrated suits almost every pop and rock record, and the range matters less than what the master does inside it. A song mastered at -10 with its transients intact will sound bigger after normalisation than the same song crushed to -6, because the platform turns the -6 version down and all you hear is what was taken away: the snare has no crack left, the vocal never lifts, the chorus does not arrive.

Genre matters too. A stomping, dense rock chorus can carry more limiting than a sparse acoustic ballad, where a listener hears the pumping immediately. The right number is whatever leaves the song sounding like itself when it is played at the same volume as everything around it.

True peak, and why it is not fussiness

Deliver masters with a true peak ceiling around -1 dBTP. Platforms encode to lossy formats (AAC, Ogg), and that process can push peaks slightly above where they sat in the WAV. If you left no room, the result is distortion that never existed in your file and that you will only ever hear after release, on someone else's phone.

What to actually do about it

• Stop mastering to a number. Master until the song sounds right, then check the number.
• Compare at matched loudness — turn your master down to your reference's level before deciding it sounds better.
• Keep the dynamic moments dynamic: the verse should be quieter than the chorus in the master, not just in the arrangement.
• Ask for a separate louder version if you need one for a club, a stall or a video. That is a legitimate second master, not a compromise on the first.

The short version: loudness is now the platform's job and dynamics are yours. A record that breathes wins the playlist.

← All notes