Loading tool…
Recordings made on different days, mics and apps rarely play back at the same volume. Add one file or a whole folder, pick a loudness target such as Spotify’s −14 LUFS or a peak level, and download versions that match. Each file is measured before and after, so you can see exactly what changed.
Processing locally on your device
Overview
There are two ways to normalize. Peak normalization finds the single loudest sample and scales the file so that sample lands at a level you choose, from −0.1 to −6 dBFS. It guarantees no clipping but says little about how loud the file sounds: one stray clap keeps a whole recording quiet. Loudness normalization instead measures integrated loudness in LUFS (Loudness Units relative to Full Scale) and adjusts the file to a target, which is how streaming services and broadcasters level programs against each other.
The loudness measurement follows ITU-R BS.1770-4 and EBU R128. Your browser decodes the file, applies the standard K-weighting filters (a gentle high-frequency shelf and a 38 Hz high-pass that approximate human hearing), splits the audio into 400 ms blocks with 75% overlap and gates out anything below −70 LUFS and anything more than 10 LU under the average. The result is checked against FFmpeg’s own ebur128 meter. True peak is measured with 4× oversampling, because the waveform between samples can rise higher than any stored sample once it is converted to analog or encoded to MP3 or AAC.
Raising a quiet file to −14 LUFS often pushes its peaks past full scale. With the limiter on, the tool applies the gain and then a look-ahead true-peak limiter that turns the loudest moments down smoothly so nothing exceeds your ceiling, usually −1 dBTP. It then re-measures and corrects the gain, because limiting lowers loudness slightly. With the limiter off, only plain gain is used and the tool stops at the ceiling, telling you how far short of the target it landed. If the limiter has to work harder than about 6 dB, you get a warning, because heavy limiting flattens drums and consonants.
Everything runs on your device. Measurement and gain are done in JavaScript on the decoded samples, so WAV output is instant. MP3, M4A, OGG, Opus and FLAC are encoded with FFmpeg compiled to WebAssembly, which downloads once (about 31 MB) and is then cached. Batch mode processes files one after another and offers the results as a ZIP.
| Preset | Integrated target | Suggested ceiling | Typical use |
|---|---|---|---|
| Spotify | −14 LUFS | −1 dBTP | Music streaming |
| Apple Music | −16 LUFS | −1 dBTP | Music streaming (Sound Check) |
| YouTube | −14 LUFS | −1 dBTP | Video soundtracks |
| Podcast | −16 LUFS | −1 dBTP | Stereo podcast episodes |
| EBU R128 | −23 LUFS | −1 dBTP | European broadcast |
Drop one or more audio files (or videos) onto the normalizer.
Choose Loudness (LUFS) and a preset such as Spotify, Apple Music, YouTube, Podcast or EBU R128, or choose Peak and set a dBFS level.
Set the true-peak ceiling and decide whether the limiter may be used.
Select a file and use Prepare preview, then switch between Original and Normalized to compare.
Pick an output format and click Normalize; check the before/after table and download.
Use cases
Bring every episode, intro and ad read to around −16 LUFS with peaks under −1 dBTP, so listeners don’t reach for the volume between segments or between your show and the next one in their queue.
Check where a master lands and make a −14 LUFS reference version to hear what a streaming service will do to it. If the limiter has to work hard, that’s a sign the master is already loud.
Normalize a batch of lessons recorded on different days and microphones in one go, so students hear a consistent level throughout the course.
Hit EBU R128’s −23 LUFS with a −1 dBTP ceiling for broadcast or festival deliveries, then confirm the result in the after column before sending the file.
Pro tips
Answers to common questions about the normalizer.
Peak normalization scales a file so its loudest sample reaches a set level. Loudness normalization scales it so its average perceived loudness, measured in LUFS, reaches a target. Two files with identical peaks can differ in loudness by 10 dB or more, which is why streaming platforms use loudness.
Without the limiter, gain stops where the true peak would cross your ceiling, and the tool tells you how far short it landed. With the limiter, very dense or distorted material can still end slightly under the target after limiting; the after column shows the real result.
Plain gain only changes the level. The limiter only touches moments that would exceed the ceiling, and on most material it works on short peaks. If it has to reduce peaks by more than about 6 dB you will see a warning, because that much limiting is audible.
The meter implements ITU-R BS.1770-4 with EBU R128 gating and was checked against FFmpeg’s ebur128 filter on test signals, matching integrated loudness, short-term maximum and loudness range to within 0.1 LU.
Yes. Add as many files as your device’s memory allows. They are measured as they load, processed one after another with the same settings, and downloaded together as a ZIP.
No. ToolEnhancer processes files on your own device inside the browser. They are never sent to, stored on or seen by our servers.
Yes. The normalizer is free to use with no sign-up, no watermarks and no daily limits.