Normalize audio to a loudness target (LUFS) or a peak level
This audio normalizer measures how loud a recording really is and adjusts it to the level you need: −14 LUFS for music on streaming services, −16 LUFS for podcasts, −23 LUFS for European broadcast (EBU R128) or −24 LKFS for US television (ATSC A/85). Add one file or a whole folder of episodes or tracks and every one comes out at the same perceived loudness, ready to download separately or as a ZIP.
Loudness is measured the way broadcasters and streaming platforms measure it, with the ITU-R BS.1770 method: a “K-weighting” filter that hears like the ear does, 400 ms measurement blocks and gating that ignores silence and quiet pauses. When reaching the target would push peaks past your true-peak ceiling, a look-ahead limiter catches them — or, if you prefer, the gain stops at the ceiling instead.
Classic peak normalisation is here too, for when you just want the loudest sample at −1 dBFS. Nothing is uploaded: measuring, processing and encoding all run in your browser.
How to use the Audio Normalizer
- Drop one or more audio files onto the box. Each is measured: loudness in LUFS and true peak in dBTP appear in the list, with the planned gain.
- Under Normalize by choose Loudness (LUFS) and a Target loudness, or Peak level and a target peak.
- Keep the True-peak ceiling at −1 dBTP (−2 dBTP for ATSC) and decide what happens if peaks would go over it.
- With several files, choose Each file to the target or Album: one gain for all, which keeps the differences between tracks.
- Preview the selected file, pick the output format and press Normalize. Download each file from the list, or all of them as a ZIP.
Audio Normalizer features
Standards-based measurement
ITU-R BS.1770-4 integrated loudness with K-weighting for any sample rate, 400 ms blocks with 75 % overlap, −70 LUFS absolute and −10 LU relative gates — verified against the EBU Tech 3341 test signals.
Presets that match the platforms
Streaming −14 LUFS, podcasts −16 LUFS, ReplayGain 2.0 reference −18 LUFS, EBU R128 −23 LUFS, ATSC A/85 −24 LKFS, or any custom target from −40 to −5.
True-peak safe
4× oversampled peak measurement and a look-ahead limiter; after limiting, the loudness is re-measured and corrected so the result lands on the target.
Batch and album mode
Up to 50 files in one go, each normalised to the target or all given the same gain so an album keeps its dynamics.
Peak normalisation
Bring the highest sample peak or true peak to any level from −30 to 0 dB.
Clear before and after
Measured loudness and peaks for every input and output, a preview, and MP3, M4A, WAV, FLAC or OGG export.
Which loudness target should I use?
LUFS (loudness units relative to full scale) and LKFS are the same unit under two names; one LU equals one decibel. Platforms only turn loud tracks down (and some turn quiet ones up), so delivering at their target means your audio plays as you mixed it.
| Where it goes | Target | True-peak limit | Source |
|---|---|---|---|
| Spotify (default setting) | −14 LUFS | Below −1 dBTP (−2 dBTP if the master is louder than −14) | Spotify for Artists |
| YouTube and other streaming | About −14 LUFS | Below −1 dBTP | Widely measured; YouTube publishes no figure |
| Apple Podcasts | −16 LKFS ± 1 dB | Not above −1 dBFS true peak | Apple Podcasts audio requirements |
| Music library players | −18 LUFS | — | ReplayGain 2.0 reference |
| European TV and radio | −23 LUFS ± 0.5 LU | −1 dBTP | EBU R128 |
| US television | −24 LKFS | −2 dBTP | ATSC A/85 |
Loudness normalisation vs peak normalisation
Peak normalisation looks only at the single loudest sample. A track with one sharp drum hit and a track of steady speech can both peak at −1 dBFS yet sound very different in loudness. Loudness normalisation measures the energy across the whole programme, weighted the way we hear, so files normalised to the same LUFS sound equally loud — which is why streaming services and broadcasters use it.
One detail of BS.1770: channels are summed, so a mono file measures about 3 LU lower than the same sound copied onto two stereo channels. If a mono podcast will be played as dual-mono, convert it to stereo first or aim about 3 LU lower.
How the result reaches the target
The tool computes the gain (target minus measured loudness) and checks where the peaks would end up. If they stay under the ceiling, the gain is simply applied. If not, the limiter shaves the peaks; because that removes a little energy, the result is measured again and the gain nudged up to land within a fraction of a dB of the target. Heavily limited audio can sound squashed, so for very dynamic recordings choose “Lower the gain” or a quieter target.
To make a single file louder by a fixed amount, use the Audio Volume Booster; to clean hiss before normalising, try the Audio Noise Remover; and to combine normalised episodes, the Audio Joiner.
Audio Normalizer FAQ
How do I normalize audio to −14 LUFS?
Open your files, choose Loudness (LUFS) with the −14 LUFS target, keep the true-peak ceiling at −1 dBTP and press Normalize. Each file is measured with ITU-R BS.1770, adjusted, limited if needed, and re-measured so it lands on −14 LUFS.
What LUFS should a podcast be?
Apple Podcasts asks for −16 LKFS with a ±1 dB tolerance and true peaks no higher than −1 dBFS, and −16 LUFS is a common choice for podcasts in general.
Is this audio normalizer free and private?
Yes. It is free with no limit on files or sign-up, and the audio never leaves your browser.
Does normalizing reduce quality?
Changing the gain is lossless in itself. Quality is only affected if peaks have to be limited heavily, or by the lossy format you save to — choose WAV or FLAC for an exact copy.
What is album mode?
Album mode measures all files together and applies one gain to every track, so quiet interludes stay quieter than loud songs while the album as a whole hits the target.
Why is my mono file 3 LU quieter than expected?
BS.1770 adds up the channels, so a single mono channel measures about 3 LU lower than the same sound on both stereo channels. Aim lower for mono, or convert it to stereo first.