Account

What new tool do you want?

Video & Audio

Audio Normalizer (LUFS Loudness)

Normalize loudness to broadcast and streaming targets — −14 LUFS for Spotify/YouTube, −16 for podcasts, −23 for EBU R128 — measured to the ITU-R BS.1770 standard.

How to use

  1. 1Add your audio and pick a target (or enter your own LUFS).
  2. 2See the measured loudness before and after, preview, then download.

How it's calculated

Integrated loudness uses K-weighting and 400 ms gated blocks (ITU-R BS.1770-4). Gain = target − measured; a limiter keeps true peaks below the ceiling.

Frequently asked questions

What is LUFS?

Loudness Units relative to Full Scale — a measure of perceived loudness used by streaming services and broadcasters.

Which target should I use?

Spotify and YouTube normalize around −14 LUFS, Apple Music −16, podcasts −16 (−19 mono), and EU broadcast −23 LUFS.

What's the difference from peak normalization?

Peak normalization matches the loudest sample; loudness normalization matches how loud it sounds, which is what listeners notice.

Why 'make it louder' and 'normalize loudness' aren't the same job

A simple gain boost or peak normalizer only looks at the single loudest sample in the file and scales from there — which means one stray clip or transient can leave the rest of the track sounding quiet by comparison. This tool instead measures integrated loudness using K-weighting and 400 ms gated blocks per the ITU-R BS.1770-4 standard (the same measurement Spotify, YouTube and broadcasters use), so the result reflects how loud a human actually perceives the whole track to be, not just its single highest peak.

Picking the right target isn't just a number — it's where the file is going

The six presets exist because different platforms re-normalize differently: Spotify, YouTube and TikTok target −14 LUFS and will turn your track down if you deliver it hotter, wasting any extra loudness you mastered in. Podcasts conventionally ship at −16 LUFS stereo (or −19 mono, since mono content measures louder for the same perceived volume), and EU broadcast delivery requires −23 LUFS under EBU R128 — submitting audio at −14 to a broadcaster that expects −23 can fail their technical QC outright.

Worked example: fixing a quiet interview recording

An interview recorded quietly measures at −20 LUFS integrated. Targeting −14 LUFS (the Spotify/YouTube preset) requires a gain of target minus measured: −14 − (−20) = +6 dB applied across the whole file. If that +6 dB push would send any true peak above the −1 dBFS ceiling, the built-in limiter clamps those peaks rather than letting them clip — so the loudness shows the full +6 dB gain, but the loudest transients are gently limited instead of distorting.

The peak ceiling isn't decorative

Hitting a precise integrated-loudness number while ignoring true peak is how 'loudness normalized but clips on playback' files happen — a common real-world mastering mistake. The ceiling field (default −1 dBFS) guards specifically against this: codecs like MP3 and AAC can overshoot sample peaks slightly during encoding, so mastering with a small amount of headroom below 0 dBFS (rather than normalizing right up to it) avoids audible clipping once the file is re-encoded for streaming.

Why a 400 ms gate matters for the measurement itself

BS.1770 doesn't average loudness over the whole file in one pass — it breaks the audio into overlapping 400 ms blocks, applies K-weighting (a filter that approximates how the human ear perceives frequency) to each, and then gates out blocks that are quiet relative to the rest before averaging what's left. That gating step is what keeps a few seconds of silence at the start of an interview, or a quiet pause between songs, from dragging the overall loudness reading down — without it, two files with identical spoken-word levels but different amounts of silence would report different LUFS values for no audible reason.

Related tools

More: Video & Audio