What does the shape of the histogram tell me?

How the programme's loudness is distributed. A single tall spike means it stays at one level throughout, which is typical of levelled speech or a heavily compressed master. A broad, flat spread means the loudness genuinely moves, as it does in classical and film material.

What does the excluded percentage mean?

It is the share of analysis windows removed by the gates before counting, mostly very quiet material and windows well below the programme's own average. A high figure is worth noticing, because it means much of the recording sits far below the loudness the chart is describing.

Why do two separate clusters appear sometimes?

Usually because the programme was assembled from takes recorded or processed at different levels, leaving the material grouped around two loudness points. It is a specific and useful finding: the aggregate range figure shows the symptom, and this chart shows the cause.

Why are very quiet windows left out?

A relative gate is applied twenty units below the programme's own average so that a fade-out or a long quiet section cannot stretch the chart into a tail that accounts for almost none of the listening. The percentage of excluded windows is reported so the effect is visible rather than hidden.

Does the histogram show where in the file a level occurs?

No, it is a distribution and discards time by design. To see when the loudness changes rather than where it concentrates, the range meter's time series shows the short-term loudness plotted against the file's duration.

Loudness Histogram

You Might Also Need

Loudness Range (LRA) Meter

Measure loudness range

Loudness Meter (LUFS)

Measure integrated loudness in LUFS

Loudness Normalizer

Normalize to LUFS loudness standards

Streaming Platform Loudness Guide

Platform-specific loudness targets

Loudness Histogram charts the distribution of a programme's short-term loudness, so you can see whether a recording lives at one level or sweeps across many. It also reports the range and how much of the material the gates removed.

What is Loudness Histogram?

Each bar is one loudness unit wide, so the height of a bar shows how many short-term windows of the programme landed at that loudness, and the shape of the whole chart is the character of the material at a glance.

What the view reports and how the distribution is built:

A histogram of short-term loudness grouped into single unit bins, drawn with the low end of the scale on the left.
The loudness range in units, taken from the same gated distribution the range meter uses.
The percentage of analysis windows excluded by the gating, which is a useful honesty check on the measurement.
One bar per occupied bin, with the axis labelled at the lowest, middle and highest bin centres.

The distribution is built from short-term loudness windows measured on a gated scale, then filtered by a relative gate set twenty units below the power mean of the programme itself before being counted into bins. That gate is what keeps a single quiet passage or an extended fade from stretching the chart into a long, uninformative tail, and the excluded percentage makes the effect visible instead of silent. Reading the shape is the point of the tool. A tall narrow spike means the material holds one loudness throughout, which is what a levelled podcast or a heavily compressed master looks like; a broad spread means the loudness genuinely moves, as it does in classical, film or acoustic recordings. A chart with two separate clusters is often the signature of a programme assembled from takes that were not matched, which is a problem the aggregate range figure cannot reveal on its own. Counting in LU bins of one unit each is a deliberately coarse choice, fine enough to show the shape and coarse enough to stay readable on a canvas of fixed width. The chart is a distribution rather than a timeline, so it answers how the loudness is arranged and leaves when it changes to the range meter's series. Reading it alongside those figures gives a fuller picture of a programme than any one of them provides alone, and a chart that is barely visible at all usually means the file is too short for the analysis to have a distribution worth drawing.

What the chart does not cover:

It describes loudness over time only in aggregate, so it cannot say where in the file a bin occurs.
Bin width is fixed at one unit, so finer movement is merged together.
Material too short for a useful number of windows produces a sparse or empty chart.
The gate is applied before counting, so excluded windows are represented only by the percentage figure.
There is no interactive readout per bin, and no export of the counts.
It reports no target and applies no correction.

Use it when a programme needs to be consistent rather than merely on target: the integrated loudness says where it sits, and this chart says whether it stays there. It is also a quick way to see the effect of a processing chain, since the same material measured before and after levelling produces visibly different shapes.

How to use Loudness Histogram

Drop an audio file onto the panel; the short-term distribution is measured and drawn automatically.
Read the overall shape: a single tall spike means one loudness, a broad spread means genuine movement.
Check the range card and the percentage of windows excluded by the gates next to the chart.
Look for more than one cluster in the shape, which usually indicates material that was not levelled consistently.

When to use Loudness Histogram vs related tools

Loudness Histogram is the view to choose when the shape of a programme's loudness matters more than its summary figures. Loudness Range (LRA) Meter reports the same distribution as a range and a time series, and Loudness Meter (LUFS) gives the integrated figures and platform targets. When the chart shows a spread that should be tighter, Loudness Normalizer or Dynamic Range Compressor is the tool that acts on 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.