Free LUFS meter

What a LUFS meter tells you

A LUFS meter, also called a loudness meter or loudness analyzer, tells you how loud a song sounds to a person, not how high its waveform reaches. LUFS means loudness units relative to full scale. Two songs can peak at the same level and still sound very different in volume, and LUFS is the measurement that tells them apart. The method is ITU-R BS.1770, the one behind the EBU R128 broadcast rules and the loudness normalization of the big streaming services.

The meter first filters the audio the way an ear hears it: it lifts the highs by about 4 dB and cuts the deepest bass, because we are less sensitive to very low notes. Then it averages the energy over time and reports it in LUFS. One loudness unit, or LU, equals one decibel, so a song that is 3 LU louder is 3 dB louder in the way that counts. Quiet moments are gated out of the long measurement, so a silent intro or a pause does not drag the number down.

The numbers on this page and what they mean

Integrated loudness, or integrated LUFS, is the one number for the whole song. It is what Spotify, YouTube and the others compare with their reference level, so read it on the final master.

Short-term loudness looks at a sliding window of 3 seconds, and we show the highest one, which tells you how loud your chorus really is. Momentary loudness uses a window of 400 milliseconds, so its highest value is the loudest single moment.

Loudness range, or LRA, says how much the loudness moves between the quiet and the loud parts, in LU. It ignores the extremes by taking the spread between the 10th and the 95th percentile of the short-term values, as EBU Tech 3342 describes. A small number is a steady, compressed song and a big number is a dynamic one.

True peak is the highest level the signal reaches once the samples are turned back into a smooth wave, in dBTP. It can be higher than the highest sample, and that gap is what clips an MP3 or an AAC file after the encode. We oversample the audio four times to find it, which is what BS.1770 asks for.

PLR, the peak to loudness ratio, is the true peak minus the integrated loudness. A song at −14 LUFS with a true peak of −1 dBTP has a PLR of 13 LU. It is a quick read on how much punch is left after the limiting.

Streaming loudness targets in October 2026

Streaming services turn songs up or down so that a playlist plays at one volume. This is loudness normalization. Spotify uses a reference of about −14 LUFS for its default setting, YouTube about −14, Tidal about −14, Amazon Music about −14, Deezer about −15 and Apple Music, with Sound Check on, about −16. Clubs and DJ sets do not normalize at all, so a master for the dance floor can sit much louder.

If your song is louder than the reference, the service turns it down by the difference. A master at −9 LUFS plays about 5 dB quieter on Spotify, and it keeps the squashed dynamics that its loudness cost, so it can sound weaker than a −14 master that was made with care. This is what people call the loudness penalty. If your song is quieter than the reference, Spotify and Apple Music may turn it up as far as its peaks allow, while YouTube, Tidal, Deezer and Amazon Music are reported to only turn songs down.

The table above is a small LUFS calculator for streaming. It subtracts the reference of each service from your integrated loudness and shows the gain that would apply, and it checks your true peak against the ceiling each service asks for, usually −1 dBTP and −2 dBTP for Amazon Music. References change, so treat them as a guide.

Broadcast has its own targets: EBU R128 asks for −23 LUFS in Europe and ATSC A/85 asks for −24 LKFS in the United States. Podcast services often ask for around −16 LUFS.

True peak, sample peak and why −1 dBTP

The peak meter of your editor shows the highest sample. Between two samples the real wave can go higher, and when a player or an encoder rebuilds it the peak can pass 0 dBFS and clip, even though no sample did. A true peak meter finds those hidden peaks.

Streaming services re-encode what you upload into lossy formats, and that adds a little more. A ceiling of −1 dBTP is the usual advice, and −2 dBTP is safer for a loud master. This page shows the true peak and the sample peak side by side and marks on the graph where the true peak goes over −1 dBTP.

How accurate is this meter?

The loudness code is the one OBRA Audio Mastering uses to measure every master: K-weighting, 400 ms blocks with 75 percent overlap, an absolute gate at −70 LUFS and a relative gate 10 LU below the average, as BS.1770 and EBU Tech 3341 describe. We checked it against the EBU R128 filter of FFmpeg, and on the files we tried the integrated loudness, the loudness range and the true peak agree to within about a tenth of a unit.

Three things to know. A mono file is measured as if it played from both speakers, which is what a player does, and that reads 3 LU louder than counting the single channel once. Files with more than two channels are not supported yet. And a meter reads the decoded audio, so an MP3 and its WAV source can differ a little.

Who uses a LUFS checker

Mastering engineers use it to check a delivery before it goes out, with the integrated loudness, the true peak and the range in one place. Producers use it to see how loud a reference track really is and to match it. Podcasters and video makers check a mix against the level a platform asks for. Teachers use the graph to show what limiting does to a song. One button copies the results as plain text.

Tips for hitting a target without crushing the song

Do not chase a number. A master at −9 LUFS that the service simply turns down often sounds better than one pushed to −14 with heavy limiting. Judge by ear first and use the meter to confirm.

Leave headroom. Keep the true peak at −1 dBTP or lower, and lower still when the master is loud, so the lossy encodes do not distort.

Look at the graph, not only at the integrated number. A chorus 4 LU louder than the verse is dynamics. A flat slab at the top is a limiter at work.

Keep an album consistent, and compare masters at the same loudness: a louder version always sounds better for a few seconds.

Free, private and in beta

This tool is in Beta. It is free, needs no account and runs in your browser, so your song is not uploaded and nothing is stored. Files up to 15 minutes work. When you want the song itself to reach −14 LUFS with a safe true peak, OBRA Audio Mastering makes three masters that you can listen to for free.