There is no single correct number. Most streaming services measure each track's integrated loudness and turn loud tracks down to a reference level, about −14 LUFS on Spotify, YouTube, Tidal and Amazon Music, −15 LUFS on Deezer and −16 LUFS on Apple Music. A master louder than that plays at the same level as a quieter one, with less dynamic range left. So choose the loudness at which your master sounds best when you compare versions at matched level, keep its true peak at −1 dBTP or lower, and check what each service will do to it.

What streaming does to your master's level

A streaming service reads the integrated LUFS of your file and changes its playback level by the difference from its reference. It changes the level, not the file. These are the reference levels:

ServiceReference levelNotes
Spotify−14 LUFSDefault setting. Loud is −11 LUFS, Quiet −19 LUFS. Can turn quiet tracks up as far as the peaks allow.
Apple Musicabout −16 LUFSSound Check.
YouTubeabout −14 LUFSTurns loud tracks down only.
Tidalabout −14 LUFS
Amazon Musicabout −14 LUFS
Deezerabout −15 LUFS

Only Spotify publishes its rules in full, so treat the other numbers as estimates. Instagram, TikTok and SoundCloud publish no loudness target for music. The LUFS targets for each streaming service guide covers each one in more detail.

A worked example

Take three masters of the same song at −9, −14 and −18 LUFS integrated, played on Spotify's default setting.

  • The −9 LUFS master is 5 dB over the reference, so Spotify turns it down 5 dB.
  • The −14 LUFS master plays unchanged.
  • The −18 LUFS master is 4 dB under. Spotify turns it up only if its true peak stays under −1 dBTP afterwards. With a true peak of −5 dBTP or lower it gets the full 4 dB. With a true peak of −3 dBTP it gets 2 dB and plays at −16 LUFS.

The first two now play at the same loudness. The −9 LUFS master went through 5 dB more limiting to get there, and the listener hears none of that extra level. On YouTube, which only turns tracks down, the −18 LUFS master plays 4 dB quieter than the other two.

The trade-off between louder and quieter

Pushing a master louder means more limiting and clipping, which shaves the loud moments down toward the quiet ones. That has two sides.

A louder master plays louder wherever nobody turns it down. That includes Spotify listeners who switch normalization off in the app's settings, and places that publish no target. Dense limiting can also be the sound you want, since some music is meant to feel flat-out and compressed.

A quieter master keeps more of the difference between its loud and quiet moments. Drums hit harder against the rest of the mix, and choruses lift more from the verses. On services that only turn tracks down, it isn't penalized. It does play quieter than other tracks anywhere it isn't turned up, such as YouTube when it sits under −14 LUFS.

So the real question is how much limiting the music can take before it sounds worse, measured at the level a listener will hear it. A number from a chart can't answer that for your song. A comparison can.

How to decide by comparing level-matched

Louder sounds better to almost everyone. Ears are less sensitive to lows and highs at low levels (the Fletcher-Munson curves, standardized in ISO 226), so the louder of two versions sounds fuller and brighter. A difference of about 1 dB can be enough to bias you. If you compare a −9 LUFS master against a −14 LUFS one at their own levels, the −9 one will usually win for that reason alone. Turn them to the same loudness first, the way a streaming service will.

  1. Export two or three versions of the master with different limiter settings. For example, one at about −14 LUFS, one 3 dB louder and one 3 dB louder again. Keep everything else the same.
  2. Measure each version's integrated LUFS and true peak.
  3. Turn every version down to the loudness of the quietest one. Don't turn the quiet one up, or you may push it into clipping.
  4. Play them in sync and switch between them at the same moment in the song. Listen to the loudest section and to the transitions into it, where heavy limiting shows first.
  5. If the louder version sounds the same or worse at matched level, use the quieter one. You lose nothing on the services that turn loud tracks down. If the louder version sounds better to you even at matched level, keep it. The density is part of the sound, and it survives normalization.
  6. If you can't tell them apart, or don't trust your choice, run a blind A/B test so you don't know which file is which.

In ABSeq, load the versions as tracks A to D and leave Auto Gain on. It measures each file's integrated loudness and turns every track down to the quietest one, so the switch compares sound, not level. The guide to comparing masters goes through the same process for comparing engineers or revisions.

Set the peak ceiling

Loudness and peak level are separate decisions. Spotify recommends masters peak no higher than −1 dBTP, and −2 dBTP for masters louder than −14 LUFS. True peak matters because the waveform between samples can rise above the highest sample. A master at 0 dBFS sample peak can go over 0 dBTP, which can clip in the playback converter or when the file is encoded to AAC, MP3 or Ogg. The true peak guide explains how to measure it and why a limiter's ceiling doesn't always match the meter.

Peaks also set how far Spotify can turn a quiet master up, as the −18 LUFS example showed. A quiet master with high peaks gets less help than one with peaks well under −1 dBTP.

Check what each service will do

Once you've picked a version, work out the change on each service. Subtract your master's integrated loudness from the service's reference level. A −11 LUFS master is turned down 3 dB on Spotify, YouTube, Tidal and Amazon Music, 4 dB on Deezer and 5 dB on Apple Music. ABSeq shows this on every track. Tap the loudness readout and it lists the change for Spotify, Apple Music, YouTube, Tidal, Amazon Music and Deezer. "Spotify −1.8 dB" means Spotify would turn the track down 1.8 dB. The track readouts section of the help guide explains each value.

You don't need a separate master for each service. One master gets turned down by a different amount on each, and the sound stays the same.

Questions

How loud should my master be for Spotify?

Spotify plays tracks at −14 LUFS integrated on its default setting, so a louder master is turned down by the difference and gains no level. Master at the loudness that sounds best when compared level-matched, and keep the true peak at −1 dBTP or lower, or −2 dBTP if the master is louder than −14 LUFS.

Should I master to −14 LUFS?

You don't have to. −14 LUFS is the level Spotify, YouTube, Tidal and Amazon Music turn loud tracks down to, so mastering there means those services leave the level alone. A louder master is fine if it still sounds better to you at matched level; it is turned down and keeps its sound. A master well under −14 LUFS plays quieter on services that only turn tracks down.

How loud should my master be in dB?

Peak level in dB doesn't tell you how loud a master is. Two masters that both peak at −1 dBTP can be 6 LU apart in loudness. Measure loudness in LUFS integrated, and use dB for the true peak ceiling, which Spotify recommends at −1 dBTP or −2 dBTP for masters louder than −14 LUFS.

Is a master louder than −14 LUFS bad?

It isn't penalized, it just isn't rewarded. A −10 LUFS master plays 4 dB quieter on Spotify than its file level, at the same loudness as a −14 LUFS master, with whatever dynamic range the extra limiting removed. Whether that is worse depends on the music, so compare the two at matched level and decide by ear.

What is mastering loudness for streaming?

It is the integrated loudness you choose for a master, knowing the services will change its playback level. Spotify uses −14 LUFS, Apple Music about −16, YouTube, Tidal and Amazon Music about −14 and Deezer about −15. Instagram, TikTok and SoundCloud publish no music loudness target.