How to Measure Audio Loudness

How loud is your audio, really? Peak meters lie — two tracks peaking at 0 dBFS can sound 8 dB apart. To measure audio loudness properly you need a LUFS meter: perceived loudness with K-weighting and smart gating. This guide shows how loudness measurement works, how to read the numbers, and how to check loudness of any audio file free online.

Every streaming platform — Spotify, YouTube, Apple Music, TikTok — measures your music in LUFS and adjusts playback gain accordingly. If you don't measure loudness before release, the platform measures it for you, and you don't get a vote on the result. A two-minute loudness check is the cheapest quality control in audio.

Why peak meters can't measure loudness

Your DAW's channel meters show peak level: the highest sample value. Useful for avoiding digital clipping, useless for loudness. Human hearing is not a peak detector — we're most sensitive around 2–5 kHz, barely notice sub-bass, and integrate sound over time. A bass-heavy track and a midrange-forward track can share identical peak readings while one feels twice as loud.

RMS meters tried to fix this by averaging power, but RMS has no perceptual weighting and counts silence fully — a track with a 30-second quiet intro measures "quieter" than it sounds. LUFS solved both problems: K-weighting shapes the measurement to match hearing, and gating excludes silence and background from the average. That's why every platform standardized on it. Full technical breakdown: What is LUFS?

The three loudness measurements: integrated, short-term, momentary

A proper audio loudness meter shows three time scales. They answer different questions:

Integrated: the official number

Integrated LUFS averages the entire program with EBU R128 gating: 400 ms blocks, discard anything below -70 LUFS absolute, then discard blocks more than 10 LU below the first average. This is the number Spotify uses, the number broadcast specs cite, and the number you put on a delivery sheet. Always measure the full file as delivered — including silence at the head and tail the way the listener gets it.

Short-term: the section number

Short-term uses a 3-second sliding window with no relative gating. If your verse reads -16 short-term and your chorus reads -9, that's a 7 LU jump the listener feels — useful information whether you're balancing a mix or checking that a podcast interview doesn't lurch between speakers.

Momentary: the live number

Momentary reacts every 400 ms with no gating. It's jumpy by design — watch it while adjusting a limiter, ignore it when judging a finished track.

True peak: the safety number next to loudness

Every loudness measurement session should also check true-peak (dBTP). Loudness tells you how it sounds; true-peak tells you whether it clips. Sample peaks lie: after digital-to-analog conversion or lossy encoding, inter-sample peaks can exceed the highest sample by 1–3 dB. A file peaking at -0.1 dBFS sample-peak can hit +2 dBTP — real clipping on playback.

Rule: keep true-peak at or under -1.0 dBTP for distribution (-1.5 for podcasts and lossy pipelines). If your loudness is right but true-peak is hot, you need a true-peak limiter, not more gain reduction. Deep dive: True Peak vs sample peak and dBTP explained.

How to use an online LUFS meter

You don't need to install anything. JSM Loudness is a free LUFS meter online that runs 100% in your browser:

1. Open the Analyzer.
2. Drop your audio file — MP3, WAV, M4A, FLAC, or OGG.
3. Read the results: integrated LUFS, short-term max, momentary max, LRA, and true-peak dBTP appear instantly.
4. Compare against your target (table below). If you're off, normalize to your target and re-measure.

Because measurement happens locally via the Web Audio API, your audio never leaves your device — no upload queue, no privacy concern, and it works offline once the page is loaded. The implementation uses exact ITU-R BS.1770-4 coefficients, verified at -14.05 LUFS against EBU reference material.

Reading your results: what the numbers mean

Integrated:  -11.8 LUFS   → a touch hot for Spotify (-14 target);
                               expect ~2.2 dB turn-down. Fine if intentional.
Short-term max: -8.4 LUFS  → loudest 3-sec section; big verse/chorus
                               contrast if 6+ LU above integrated.
Momentary max:  -6.9 LUFS  → transient peaks; watch while limiting.
LRA:            7.2 LU     → healthy dynamics for modern pop.
True-peak:     -0.8 dBTP   → TOO HOT. Limit to -1.0 before release.

Notice the workflow: integrated tells you where you stand vs the target, short-term tells you about section balance, LRA tells you about dynamics health (see LRA / PLR / PSR explained), and true-peak is the pass/fail safety check. All five numbers together — never one alone.

Loudness targets cheat sheet

These are delivery references, not creative mandates. A -10 LUFS master isn't "wrong" — Spotify will just turn it down 4 dB. What's wrong is not knowing your number and being surprised by it.

Common loudness measurement mistakes

• Measuring a snippet. Integrated LUFS over the chorus is not your track's loudness. Measure the full program, head to tail.
• Comparing ungated to gated meters. A meter without EBU R128 gating reads lower on material with quiet sections. Always compare integrated-to-integrated with the same gating.
• Forgetting true-peak. Loudness perfect, peaks clipping — the most common failed QC in self-releases.
• Measuring the MP3 instead of the master. Lossy encoding shifts both loudness (slightly) and true-peak (sometimes a lot). Measure the WAV you deliver; spot-check the encode.
• Chasing -14.0 exactly. ±1 LU is inaudible and within tolerance everywhere. Mix with your ears, verify with the meter.

Frequently asked questions

What is a good LUFS level for music?

For streaming, -14 LUFS integrated is the reference — Spotify and YouTube normalize to it. In practice, punchy modern masters land between -12 and -9 LUFS integrated; dynamic acoustic or classical material sits around -16 to -20. There is no single correct number: match the genre's dynamic range, keep true-peak under -1 dBTP, and let the platform handle playback gain.

Can I measure LUFS online without installing software?

Yes. JSM Loudness is a free online LUFS meter that runs 100% in your browser via the Web Audio API — drop an MP3, WAV, M4A, or FLAC and get integrated, short-term, momentary, LRA, and true-peak instantly. Your audio never leaves your device, so it's private and works offline once loaded.

Why does my DAW loudness reading differ from an online LUFS meter?

Three usual causes: (1) different time windows — your DAW may show momentary while the meter shows integrated; (2) gating — EBU R128 integrated uses -70 LUFS absolute and -10 LU relative gates, and meters that skip gating read lower on material with quiet sections; (3) true-peak vs sample-peak — oversampled true-peak reads up to 3 dB higher than sample peak. Compare integrated-to-integrated with the same gating and the numbers agree within 0.1 LU.

What is the difference between LUFS and RMS when measuring loudness?

RMS measures average electrical power with no perceptual weighting and no gating — silence and sub-bass count fully. LUFS applies K-weighting (emphasizing 2-5 kHz where hearing is most sensitive) and EBU R128 gating that discards silence and quiet passages. Two tracks at identical RMS can differ by 6 dB in LUFS; LUFS matches what you hear, which is why streaming platforms standardized on it.