Skip to content

Integrated vs short-term vs momentary loudness

Send feedback

Last updated 6 min readTags:MasteringLoudnessLUFS

Learn the difference between integrated, short-term and momentary LUFS, and which reading to use for whole tracks, choruses and sudden level changes.

A loudness meter can show several LUFS readings for the same audio. They differ because each covers a different period, and the whole-track measurement also excludes audio below certain thresholds.

For mix decisions, start with the reading that covers the passage you want to assess:

Reading Time measured Best used for
Integrated (I or LUFS-I) All audio measured since the meter was started or reset, with gating Checking overall track loudness
Short-term (S or LUFS-S) The previous 3 seconds Comparing verses, choruses and reference passages
Momentary (M or LUFS-M) The previous 0.4 seconds Checking brief changes in loudness

The momentary and short-term windows move forwards as the audio plays. Neither uses gating in EBU Mode. These time windows follow EBU Tech 3341.

All three readings use LUFS, so a less negative number means louder: −10 LUFS is louder than −14 LUFS. For more on the unit itself, see what LUFS measures.

Integrated loudness: measure the whole track

Integrated loudness gives you one figure for the audio you have measured. To obtain a whole-track result, reset the meter and play the song from beginning to end, or analyse the complete exported file.

During playback, the integrated figure is provisional. It can change as more of the song is measured. If you stop after the first chorus, the reading describes the audio so far, not the finished song.

What gating leaves out

Integrated loudness is a gated, energy-based measurement, rather than a simple average of the LUFS numbers on the display. The algorithm assesses overlapping 400 ms blocks. It first excludes blocks below −70 LUFS, then excludes those more than 10 LU below the loudness calculated after that first gate. The calculation is specified in ITU-R BS.1770.

This means silence and sufficiently quiet audio are left out. A quiet intro is not automatically excluded; it must fall below the gate. A softer verse that remains above it still contributes to the result.

Which reading matters for streaming?

Integrated loudness is the relevant reading for whole-track loudness normalisation. Spotify, for example, specifies −14 LUFS for its Normal setting. A track measured at −10 LUFS integrated would receive roughly 4 dB of attenuation when that setting applies. Album playback can use a shared gain adjustment, and normalisation is not active on every playback device. See Spotify's normalisation guidance.

Keep this separate from mix balance. If your vocal is too quiet or the bass overwhelms the chorus, changing the whole track's gain will not correct either problem.

Short-term loudness: compare passages

Short-term loudness is useful when a whole-track figure hides a difference between sections. You might want to check whether the second chorus is louder than the first, or compare your chorus with a commercial reference.

Suppose your chorus reads around −9 LUFS short-term and the reference's chorus reads around −11 LUFS. Your passage measures about 2 LU louder. For an initial listening comparison, turn your track down by about 2 dB with a gain control, then listen again. These figures are illustrative, not recommended targets.

Keep the monitor volume unchanged and compare similar passages. A sparse verse and a dense final chorus are poor choices for judging differences in tonal balance or punch.

For a more repeatable level match:

  1. Choose a representative passage from each track, such as eight bars of the main chorus.
  2. Measure the integrated loudness of each selected passage separately, with a reset between measurements.
  3. Reduce the louder passage by the measured difference using a gain or trim control.
  4. Compare by ear and make a small gain adjustment if one still sounds louder.

Here, integrated loudness describes each selected excerpt. Short-term loudness remains useful for seeing how the level varies within it. Avoid matching two passages from a single changing short-term value.

You can also compare verses and choruses within your own mix. A louder chorus may be exactly what the arrangement needs. Adjust it when the transition sounds wrong, rather than because the readings differ.

Momentary loudness: inspect a brief change

Momentary loudness reacts more quickly than short-term loudness. It can draw attention to a loud vocal phrase, an abrupt entrance or a transition that feels too forceful.

It still averages audio over its window. It does not measure the highest waveform peak; use true peak to assess peak level and headroom.

If a transition sounds too loud, replay the bar before and after it. Watch the momentary reading, then check which parts enter or change level. Audition those parts in the full mix and try a small fader or automation adjustment on the likely cause.

The meter cannot identify the instrument responsible. A rise could come from one part, several layers arriving together, or processing on the mix bus. A momentary value above the short-term value does not, by itself, prove that anything needs fixing.

How to use a loudness meter in your DAW

Place the meter on the master bus after any processing whose output you want to assess. For an export check, it should receive the same signal as the exported file.

  • For the whole mix: reset the meter and measure one complete pass. Do not include repeated choruses or extra playback in that result.
  • For section balance: compare similar points in each section with short-term loudness, then listen to the transitions.
  • For a brief level problem: use momentary loudness to locate the change, then inspect the individual tracks.
  • After an adjustment: compare at a similar listening loudness. An increase in volume can make a change seem better than it is.

If your meter offers automatic resets or transport-linked measurement, check how those settings behave before relying on the integrated result.

Common questions

Why is integrated LUFS lower than the chorus reading?

The whole-track result can include quieter verses, bridges and other sections that remain above the gate. A short-term reading taken during a loud chorus covers only that passage. The two figures need not match.

Can I measure a chorus on its own?

Yes. Reset the meter and measure just that section. Label the result as the chorus's integrated loudness so you do not confuse it with the whole-track figure.

Should I compress the mix until the readings are closer?

Only if you hear a dynamic problem that compression would address. The difference between integrated, short-term and momentary loudness is not a fault to correct. Judge whether the mix needs more consistent levels, and where, before choosing a processor.

For feedback on your own track, analyse your mix with Sonalix. Use the report's suggested fixes to address audible problems in your DAW, then compare the revised mix at a similar listening level.

    Type to search titles, headings and summaries.