What is the difference between integrated, momentary and short-term loudness?

Integrated is the gated average across the whole programme and is the figure platforms compare against. Momentary is the level over a four hundred millisecond window and short-term over three seconds, so the two maxima tell you how loud the file gets rather than what it averages.

Why does gating matter?

Without gating, a long quiet introduction or a stretch of silence would pull the average down and make the loud section look acceptable while it is not. The absolute gate removes very quiet material and the relative gate removes the quieter parts of the programme, so the measurement reflects what listeners actually hear.

Why does my file show a positive difference against a target?

Because it is louder than that target. A positive number of LU means the platform will apply attenuation when it normalises, so the track will play back at the level the service prefers rather than louder than the rest of the catalogue. That is a warning, not a failure.

Should my master be exactly at the target?

Matching a target is a reasonable aim for consistent playback, and podcast and broadcast work often requires it. For music the target is a reference point rather than a rule: what matters most is knowing where the file sits so the decision is deliberate instead of accidental.

Does this check peaks as well?

No. Loudness and peak level are separate measurements, and a file can be perfectly on target for loudness while still exceeding a delivery ceiling in inter-sample terms. Run the true peak meter alongside it before delivering anything that has a specified maximum.

Loudness Meter (LUFS)

You Might Also Need

PPM Meter

Peak program meter

True Peak Meter

Measure true peak levels

Loudness Normalizer

Normalize to LUFS loudness standards

Streaming Platform Loudness Guide

Platform-specific loudness targets

Loudness Meter (LUFS) measures a file the way streaming platforms do, reporting integrated, momentary and short-term loudness and placing each against the targets the major services normalise to. It answers why one upload sounds quieter than another with a number.

What is Loudness Meter (LUFS)?

Drop a file and three figures appear: integrated loudness for the whole programme, the highest momentary reading over a four hundred millisecond window, and the highest short-term reading over three seconds.

The measurement and the readings it produces:

Integrated LUFS, the number to compare against a platform's target, measured according to the ITU-R BS.1770-4 algorithm.
Momentary maximum, taken over the peak four hundred millisecond window found in the file.
Short-term maximum, taken over the peak three second window, which is close to how a listener experiences a loud passage.
A comparison strip against targets including minus fourteen for Spotify, YouTube, Amazon and Tidal, minus fifteen for Deezer, minus sixteen for Apple Music sound check and podcast standards, and minus twenty three for EBU R128 broadcast.

Two stages of K-weighting modelling the head's response are applied before the level is measured, and the measurement is gated twice. An absolute gate at minus seventy LUFS discards very quiet stretches so that a long silence cannot drag the average down, and a relative gate ten units below the ungated loudness discards the quieter parts of the programme so the integrated figure describes the material that is actually being listened to rather than including its pauses. Each target row is reported as a signed difference in LU: a positive value means the file is louder than the target and the platform will turn it down, and a negative value means it is quieter and will usually not be turned up. That asymmetry is the reason loudness normalisation is worth measuring before delivery rather than after. The platform targets shown are the ones the major services normalise toward, and they cluster in a narrow band, which is worth noting: minus fourteen, minus fifteen and minus sixteen are close enough that a master aiming at one is rarely badly wrong for the others. Broadcast's minus twenty three is a different world entirely, which is why a piece of work cannot sensibly be prepared for both without measuring against each of them in turn.

What the meter deliberately does not do:

It measures only; nothing is gained or normalised, so a file that is too loud needs a separate correction step.
Targets are shown for comparison and are not rules imposed on the material.
A file shorter than three seconds cannot produce a meaningful short-term reading.
True peak is not reported here, so a loudness-compliant master can still exceed a delivery ceiling in inter-sample terms.
Nothing is written to the file's metadata, so a delivered master carries no loudness tag as a result of measuring it.
Movement within the programme is summarised as maxima rather than as a curve.

Use it before uploading anything to a platform, and before mastering work where a target has to be hit to within a fraction of a unit. When the measurement shows the file is quiet or loud, the normaliser beside it applies the gain safely, and the range and histogram tools show whether the loudness inside the programme moves more than the material warrants.

How to use Loudness Meter (LUFS)

Drop an audio file onto the panel; the measurement runs without any settings to choose.
Read the integrated figure as the number to compare with a platform's target.
Check the momentary and short-term maxima to see how loud the file gets at its peaks.
Look down the target list for signed differences in LU, then apply gain elsewhere if the file needs correcting.

When to use Loudness Meter (LUFS) vs related tools

Loudness Meter (LUFS) answers the delivery question directly, so run it before uploading to any platform that normalises. Loudness Range (LRA) Meter shows how much the loudness moves through the programme, and True Peak Meter checks the ceiling that loudness targets usually come with. When the numbers call for a change, Loudness Normalizer applies the gain and Streaming Platform Loudness Guide explains what each service does with it.

Privacy & Security

This tool runs entirely in your browser — no data ever leaves your device. There is no server round-trip, no upload, no logging, and no account required. Your input is processed locally using client-side JavaScript and is never stored, transmitted, or accessible to anyone else. When you close the tab, everything disappears.