Why does the loudness measurement use gating?

Because a straightforward average would be dominated by the quietest parts of a file. The standard applies an absolute gate so near silence is ignored and a relative gate so quiet passages well below the main programme are excluded, which is why the integrated figure reflects the loud parts. That is the same method platforms use at ingest.

What happens if the ceiling blocks the gain?

The file lands short of the loudness target and the footer tells you so, naming the gain applied and the gain that was wanted. That happens when a file has peaks close to the rail and needs to be turned up. The fix is to bring the peaks down with a limiter first, then normalise again.

Does normalising change the dynamics?

No. The gain is one constant factor applied to every sample jointly across the channels, so the shape of the waveform is unchanged and only its level moves. Perceived dynamics are identical. If you want the file denser rather than merely louder, that is a compression job before the normalise.

Which target should I choose?

Match the destination. Podcast distribution generally expects about minus 16 LUFS, Spotify and YouTube work to about minus 14, and broadcast standards are much lower at minus 23 in Europe and minus 24 for the American standard. The presets name each of those so the choice is explicit.

Why is stereo balance preserved?

Because a single joint gain is computed for the whole file and applied identically to every channel. If each channel were normalised separately, a slightly louder left channel would be pulled down more than the right and the image would shift. Measuring jointly avoids that entirely.

Loudness Normalizer

You Might Also Need

Loudness Meter (LUFS)

Measure integrated loudness in LUFS

RMS Normalizer

Normalize audio using RMS levels

True Peak Meter

Measure true peak levels

Streaming Platform Loudness Guide

Platform-specific loudness targets

Loudness Normalizer measures a file's integrated loudness with the same standard platforms use, then applies one constant gain so it lands on the target you choose. It is the ingest normalisation a streaming service would apply, done locally first.

What is Loudness Normalizer?

Drop a file, pick a target and set a ceiling, then read the report. The footer shows the loudness before and after, the gain that was applied, and the resulting peak, so you can see whether the ceiling interfered with the calculation.

The controls:

Four target presets at minus 16 for podcasts, minus 14 for Spotify and YouTube, minus 23 for EBU R128 and minus 24 for ATSC A/85.
A target slider from minus 30 to minus 9 LUFS in single unit steps, defaulting to minus 14.
A true peak ceiling from minus 6 to zero dBFS in half decibel steps, defaulting to minus 1.
A footer reporting the before and after loudness, the gain applied and the final peak level.
An output WAV named after the source with a lufs suffix, limited to two channels.

The measurement is ITU-R BS.1770-4 integrated loudness, the same standard a platform applies at ingest, including its gating so that silence and very quiet passages cannot pull the reading down. The gain that follows is a single constant factor applied jointly across the channels, which is deliberate: because every channel is scaled by the same number, the stereo balance is untouched and only the level changes. A hot master gets turned down and a quiet mix gets turned up, exactly as a service would do, but here you see the numbers and decide the target. The ceiling exists because loudness and true peak are different things: a file can be far below the loudness target and still have peaks near the rail, in which case the requested gain would push it over. When that happens the report says so explicitly, naming the gain that was applied and the gain that would have been wanted. A silent file is reported as such rather than being scaled by zero.

What it cannot do:

It changes level only, so a squashed master arrives at the target still squashed; dynamics are untouched by design.
The gain is constant for the whole file, so loudness that varies inside a recording is not corrected.
When the ceiling limits the gain, the file lands short of the target, and the report is the only place that tells you.
It measures rather than models, so a file with unusual content may read differently from another tool's estimate of the same material.
Only two channels are accepted, so multichannel material is out of scope.
There is no limiter in the chain, so a very dynamic file cannot be squeezed to hit a loud target without a separate limiting step.

Use it when a file needs to arrive at a streaming platform already at the level that platform will use, because then its own normalisation has nothing left to do. Check the report for the ceiling case, since that is the one outcome where the file does not reach the target, and follow with a limiter if the peaks are what is standing in the way. Leaving the ceiling at minus 1 also keeps room for the overshoot a lossy encoder can add afterwards.

How to use Loudness Normalizer

Drop an audio file onto the panel; the tool accepts up to two channels.
Choose a target from the presets, or set your own value anywhere from minus 30 to minus 9 LUFS.
Set the true peak ceiling, leaving the default of minus 1 dBFS unless a platform asks for something else.
Read the before and after loudness in the footer and note any ceiling warning, then download the normalised WAV.

When to use Loudness Normalizer vs related tools

Loudness Normalizer is the tool when a file has to arrive at a specific loudness, which makes it the right finish for anything headed to a streaming service. Loudness Meter (LUFS) measures without changing anything, which is the right first step when you only want to know where a file stands, and Rms Normalizer targets an average level instead of a perceptual one. When the ceiling is what prevents the target being reached, True Peak Meter explains why, and Streaming Platform Loudness Guide documents what each service does with the result.

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.