Skip to content

Loudness normalisation: streaming levels explained

Send feedback

Last updated 5 min readTags:MasteringLoudnessStreaming

How streaming loudness normalisation affects your track, why −14 LUFS is not a universal mastering target, and how to check loudness and peaks.

Loudness normalisation adjusts playback volume to reduce loudness differences between tracks. A loud master is turned down; a quieter one may be turned up, depending on the service and its peak limits.

This is why a streaming reference level such as −14 LUFS is not a universal mastering target. Choose the amount of compression and limiting that suits your music, then check the result at matched loudness and leave suitable peak headroom.

How loudness normalisation works

The service measures loudness and calculates a gain adjustment against its playback reference. For track normalisation, the measurement is usually integrated loudness: a whole-track reading in LUFS.

The basic calculation is:

Playback gain = reference loudness − measured loudness

At a reference of −14 LUFS, a −9 LUFS track needs a −5 dB adjustment. The same gain change applies throughout the song, so the differences between its quiet and loud passages remain. This fixed adjustment is distinct from compression, which changes gain over time. The AES explanation of loudness normalisation describes both track and album approaches.

Normalisation does not undo processing already in the master. If a limiter has softened the drum attacks or introduced distortion, turning the track down leaves those changes audible.

Nor does equal integrated LUFS guarantee identical perceived loudness. Bass balance, vocal prominence and the contrast between sections still affect what you hear. A whole-track reading cannot tell you whether two choruses will sound equally loud.

What happens to loud and quiet tracks?

These examples illustrate Spotify's Normal setting, with track normalisation active:

Master Playback adjustment Result
−9 LUFS Down 5 dB −14 LUFS
−14 LUFS None −14 LUFS
−20 LUFS, maximum true peak −5 dBTP Up 4 dB −16 LUFS; the peak reaches −1 dBTP

The quiet example stops short because further gain would use up the reserved peak headroom.

Loud masters retain their processing

A heavily limited master can sound less punchy than a less limited version at matched loudness. Listen for duller snare attacks, a kick that loses definition, or distortion on bass notes and cymbals.

However, high integrated loudness does not prove that a master is damaged. Some arrangements and production styles suit a dense, compressed sound. A low LUFS reading does not guarantee good dynamics either.

The useful question is whether the processing improves the track once the volume advantage has been removed. If the more limited version still sounds better, there is a musical reason to keep it.

Quiet masters can run out of peak headroom

A track with sharp peaks and a low average level has less room for a clean gain increase. Its peak-to-loudness ratio (PLR) is the difference between maximum true peak and integrated loudness. For example, −1 dBTP minus −20 LUFS gives a PLR of 19 dB.

That figure describes the gap; it does not tell you whether the music needs fixing. A sparse acoustic recording may need those peaks. An isolated kick hit that is much louder than the rest of a mix may deserve attention.

Simply turning down the export lowers the peaks and average loudness together. It does not change PLR or solve a peak-limited playback problem. To reduce that gap, you would need to change the mix or control the peaks, with an audible trade-off to assess.

Platform settings and recommendations

The following details come from the services' own documentation, checked in November 2026.

Spotify

Spotify publishes three playback levels: Normal −14 LUFS, Loud −11 LUFS and Quiet −19 LUFS. Loud can apply limiting to quiet, dynamic tracks.

It uses one gain adjustment for an album in sequence, and individual adjustments for shuffled tracks or playlists. Its documentation excludes the web player and third-party devices such as TVs and speakers from normalisation.

Spotify recommends −14 LUFS integrated for mastering, true peaks below −1 dBTP, and below −2 dBTP for masters louder than −14 LUFS. These are its published recommendations, rather than a universal release specification. See Spotify's loudness guidance.

Apple Music

Apple calls its loudness feature Sound Check. Its Apple Digital Masters brief describes loudness measurement, metadata and volume adjustment, including support for album-level playback. That document does not state a LUFS reference level.

YouTube

YouTube's Stable volume continuously adjusts level within a video. It is a separate feature from a fixed whole-track loudness adjustment. YouTube says Stable volume is off for YouTube Music and official music videos.

Stable volume being off does not mean that music plays without loudness normalisation. The cited page does not document a music normalisation target.

Should you master to −14 LUFS?

You can, if the track sounds right there. There is no need to force every release to that figure.

A streaming playback reference answers a different question from a mastering decision. The former sets a listening level; the latter concerns the sound of your release. Apple's mastering guidance likewise advises producers to preserve their intended sound regardless of playback volume.

Avoid adding limiting solely to reach a number. Equally, do not remove compression that is central to the sound merely to produce a quieter file. Compare the alternatives fairly before deciding.

How to check your track before release

  1. Measure the complete export. Read integrated LUFS and maximum true peak from start to finish. A chorus loop gives you a reading for that chorus, not the whole track.
  2. Compare two amounts of limiting. Export your current version and one with less limiter gain reduction. Turn the louder version down to match the quieter one, then compare the same passages. Listen to drum attacks, bass definition, vocal clarity and distortion.
  3. Use a suitable reference. Choose a track with a similar arrangement and intended sound. Match loudness before judging balance or punch. For a chorus comparison, check the level of those passages as well as the whole-track readings.
  4. Check true peak after export. A limiter's displayed ceiling is not a substitute for measuring the finished file. The AES streaming recommendation specifies a maximum of −1 dBTP at the input to lossy encoding; individual service or distributor guidance may call for more headroom.
  5. Investigate isolated peaks before limiting everything harder. If one hit uses most of the available headroom, adjust that hit or its source channel and compare again. Keep the change only if it improves the track.

You can analyse your track with Sonalix to check integrated loudness and true peak. Use those readings alongside the listening comparison to decide whether the export needs an adjustment.

Was this page helpful?

    Type to search titles, headings and summaries.