Volume Normalisation Test: Free Online LUFS Meter
Overview
Volume Normalisation Test. Drop an audio file into the Volume Normalisation Test and it measures integrated LUFS, true peak and loudness range in your browser, shows how far each streaming platform will turn it up or down, and exports a normalised WAV — nothing is uploaded. Below that, Play Clip A (Loud) and Play Clip B (Quiet) check whether something in your own playback chain is flattening a known 20dB gap. The bass test takes under a minute and needs nothing but your browser.
Your show sounds fine on headphones, yet Spotify plays it quieter than the show before it. This Volume Normalisation Test works as a free online LUFS meter: it reads the perceived loudness of your audio and shows how far YouTube, Apple Music or Tidal will move it up or have it turned down, so you can fix the level before you upload.
How the Volume Normalisation Test Works
The test is a loudness analyzer that runs entirely in your browser. Your file is decoded on your own device and never uploaded, which matters for unreleased songs and private interviews. It works in any modern browser on a desktop or phone, with no app to install. Use the stereo & polarity test whenever you want a fast, repeatable way to confirm this.
- Choose an audio file in WAV, MP3, FLAC, AAC or M4A format.
- Wait a few seconds while the online LUFS analyzer measures it in the browser to the ITU-R BS.1770 standard.
- Read the loudness report and compare it with your target.
- Preview the level change before you export the corrected audio file.
Reading Integrated Loudness, True Peak and LRA
The report gives you three numbers. Integrated loudness (integrated LUFS) is the level of the whole file from start to finish, and it is the figure platforms act on. True peak is the highest level after digital-to-analog reconstruction, in dBTP, so it catches inter-sample peaks that a plain sample peak reading misses. The loudness range, or LRA, shows how far the quiet and loud passages sit apart.
LUFS, Perceived Loudness and Volume Levels
LUFS measures how loud a signal feels, not just how big it is. A peak meter in dBFS treats all frequencies the same, but your hearing does not: a 4 kHz tone seems far louder than a 60 Hz tone at equal level. K-weighting filters the audio to copy that, so the reading tracks what listeners actually notice. Two files with matching peak readings can still sound very different. That weighted reading is what the meter on this page reports, and it is why the test levels files by LUFS instead of by peak. The online channel balance check gives you a clear answer instead of guessing.
Why Loudness Units Full Scale Readings Are Negative
These readings are measured downward from 0, the digital full scale ceiling. Every reading is negative, and a lower number is quieter: -18 is softer than -12. One LU equals 1 dB of level change.
Short-Term LUFS and Momentary Readings
Short-term LUFS uses a 3-second window and follows the music as it moves, while a momentary reading uses just 400 ms. Integrated is the average loudness over the whole file. Watch the short-term trace to spot loudness fluctuations, such as a guest who speaks far softer than the host.
Choosing an Audio Normalization Method
Audio normalization measures a file first, then applies one fixed level change to reach a target. Any audio normalizer uses one of three yardsticks, and the one you pick decides how evenly the normalized audio plays next to other files.
Peak, RMS and LUFS-Based Methods Compared
- Peak normalization moves the loudest sample to your target. It is fast and keeps every relationship in the file, but two files with equal peaks can still differ in loudness.
- RMS normalization works from average energy, which is closer to what you hear and suits a DJ set or an album played track by track.
- EBU R128 uses integrated LUFS with a true-peak limit. A LUFS normalizer built on it gives the most consistent loudness between different kinds of content.
Normalizing Audio Versus Amplifying It
Amplifying adds a fixed gain you guess yourself, so a quiet file may stay quiet and a loud one may clip. Normalizing audio measures before it changes anything and makes exactly the volume adjustments needed. The level change itself does not reduce audio quality. Only the export format can: keep WAV or another lossless format if the file will be edited again.
Loudness Normalization Targets for Streaming Services
Most streaming services apply loudness normalization during playback, so each one has its own target loudness. Music and video streaming platforms mostly settle on -14 LUFS, podcasts on -16 LUFS, and television on much lower levels to keep room for dynamics.
| Platform or standard | Target loudness | Typical content |
|---|---|---|
| YouTube, Spotify, Tidal, Amazon Music | -14 LUFS | Music and video streaming |
| Apple Music, Apple Podcasts | -16 LUFS | Music and podcasts |
| Audiobooks (Audible) | -18 LUFS | Spoken word and voice recordings |
| EBU R128 broadcast (Europe) | -23 LUFS | Radio and TV |
| ATSC A/85 broadcast (US) | -24 LUFS | TV |
| Cinema | -27 to -24 LUFS | Film with wide dynamic range |
YouTube only lowers loud uploads and never raises quiet ones, so a spoken-word show mixed too soft simply stays soft there.
Loudness Standards and Tolerance
Loudness standards allow a small tolerance around the target level. The European broadcast rule accepts ±1 LU, while ATSC A/85 allows ±2 LU. Streaming platforms publish no formal window, but a file inside that window is effectively platform-compliant and will barely be touched.
Checking a Quiet Folk Single with the LUFS Meter
The question on the table: does a sparse acoustic single need a louder master before it goes to Spotify, or will the platform's loudness normalization lift it anyway? The answer turns out to be no, it won't be lifted all the way: as mastered, the song plays about 1.7 LU below its neighbours.
I drop the 24-bit WAV master into the analyzer and read three figures: -17.4 LUFS integrated, a true peak of -2.7 dBTP, and a loudness range of 9.6 LU. That wide range fits a fingerpicked guitar and one voice, and I don't want to flatten it.
Spotify's target is -14 LUFS, so the gap is -14 - (-17.4) = 3.4 LU. The catch is in Spotify's own published rule: on its default setting it raises quieter tracks only until the peak reaches -1 dB. With the peak at -2.7, the most it can add is 1.7 LU, so the single plays at -15.7 LUFS, not -14.
| Master option | Level change | Peak before limiting | Limiting needed | Spotify plays it at |
|---|---|---|---|---|
| As is (-17.4) | 0 | -2.7 | 0 | -15.7 |
| Raised to -15.0 | +2.4 | -0.3 | 0.7 | -15.0 |
| Raised to -14.0 | +3.4 | +0.7 | 1.7 | -14.0 |
I check the table by hand: each new peak is the old peak plus the level change, and limiting is whatever lands above -1.0. What stands out is the break-even: up to -15.7 the master gets louder for free, but every LU above that comes straight out of the peaks, since a limited peak at -1.0 leaves Spotify no room to add anything.
The number I watch here is the true peak, not the LUFS reading. A single transient, like a hard pick attack, can cap the platform's lift for the whole song. YouTube adds a second trap, since it never raises quiet uploads at all, so the same master would stay at -17.4 there. That tips the decision toward the middle row: 0.7 dB of limiting buys -15.0 on both services. My next step is that limit and a rerun of the test, expecting about -15.0 LUFS at -1.0 dBTP. If the guitar loses its snap, I'd go back to -15.7 rather than spend 1.7 dB of limiting to chase -14.
The Loudness Penalty on Loud Songs
The loudness penalty is how much a service turns your track down to match its distribution loudness, the level it plays its loudest music at. Take a finished track that measures -8.7 LUFS: at -14 LUFS on YouTube the penalty is -5.3, and on a service that targets -16 it grows to -7.3. Listeners hear the same level as every other song, but the heavy compression used in mastering to get it that loud stays in the sound. The penalty is not a score to chase. A less crushed version with more dynamics can come out louder in feel once the platform levels it.
Checking Reference Tracks Before You Master
Run a few commercial reference tracks from your genre through the same LUFS checker, then compare your mix with that reference material at matched level. If your version sounds flatter, you have traded dynamic range for loudness that the platform takes back anyway.
Normalize Audio Online Without Clipping
Raising a quiet file moves its peaks up by the same amount, so you should normalize audio with an eye on the highest peak, not just the LUFS figure.
Worked Example: Raising a Quiet Podcast Episode
Say your recording of a 38-minute podcast episode measures -19.6 LUFS integrated, with a true peak of -3.8 dBTP and a loudness range of 5.2 LU. Your host wants -16 LUFS. The level change is:
$$\text{Gain} = L_{\text{target}} - L_{\text{measured}} = -16 - (-19.6) = +3.6\ \text{LU}$$
The peak level rises by the same 3.6, to -0.2 dBTP. That breaks the usual -1.0 dBTP ceiling, so a limiter must pull the highest peaks down by 0.8 dB. The integrated level still lands at -16, but the file now has no headroom left above that ceiling.
Leaving Headroom for Lossy Formats
When a lossy encoder turns WAV into MP3 or AAC, the decoded waveform can overshoot the original samples. Before encoding an MP3, read the peak figure in the test's report: if it sits above -1 dBTP, use gentle limiting rather than pushing the whole file up.
Frequently Asked Questions
- What is LUFS?
- LUFS (Loudness Units relative to Full Scale) measures how loud audio sounds to people, not just its raw signal level. It is the unit streaming services and broadcasters use to set loudness targets, and one LU equals one dB of level change.
- What does the loudness penalty mean?
- It is how much a streaming service will turn your file down (a negative number) or up (a positive number) to match its playback level. A penalty of -4.0 on Spotify means your song plays 4 dB quieter than it was mastered.
- Does YouTube make quiet audio louder?
- No. YouTube turns down uploads louder than about -14 LUFS but leaves quiet ones as they are, which is why the YouTube penalty never goes above zero here.
- What loudness should I normalize to?
- Music and video streaming platforms play back at about -14 LUFS, podcasts at -16 LUFS, European TV at -23 LUFS under EBU R 128 and US TV at -24 LUFS under ATSC A/85. Pick the standard where the file will be heard.
- Does normalizing audio reduce quality?
- A level change on its own costs nothing. Quality only drops if the raised peaks need limiting or the file is exported to a lossy format such as MP3, so watch the true peak after gain.
- What is the difference between peak, RMS and EBU R128 normalization?
- Peak normalization moves the loudest sample to a target, RMS uses the average energy, and EBU R128 uses integrated LUFS with a true peak limit. EBU R128 gives the most even loudness from one file to the next.
- Why should true peak stay below 0 dBTP?
- Inter-sample peaks can go over 0 dB when audio is converted to analog or encoded to a lossy format, which causes distortion. A ceiling of -1 to -2 dBTP leaves room for that.