AudioKit

Audio normalizer

Make every episode, song or voice note play at the same perceived loudness, measured in LUFS.

  • Runs on your device
  • No upload
  • No sign-up
  • Free

Open a quiet recording, song or video

Stays on this device: nothing is uploaded. You can also drag a file here or paste it.

To normalize audio: open the file below, pick a loudness target (−16 LUFS for podcasts, −14 LUFS for Spotify and YouTube), press Play boosted to check, then Export. True peaks are kept below −1.5 dBTP.

Step by step

  1. Open the file. Peak and average levels appear straight away.
  2. Pick a target: Podcast −16, Spotify/YouTube −14 or Loud −11 LUFS.
  3. Press Play boosted to hear 15 seconds of the result.
  4. Export and Download.

Peak vs loudness normalization

Peak normalization raises a file until its single loudest sample touches a ceiling. Two files with the same peak can still sound very different in loudness, because one may be dense and the other mostly quiet with a few spikes.

Loudness normalization measures how loud the whole file sounds, using the K-weighted measurement in ITU-R BS.1770 that EBU R128 builds on, and moves it to a target in LUFS. That is what streaming services do on playback, which is why the targets here match theirs. AudioKit uses ffmpeg’s loudnorm filter with a −1.5 dBTP true-peak limit.

Which target to choose

Spotify and YouTube play music at about −14 LUFS. Apple Podcasts recommends about −16 LUFS for spoken word. If your file is much quieter than the target, the filter also limits peaks, so very dynamic recordings may sound a little more compressed.

Related: Volume booster · Remove silence · Audio editor · Merge MP3 files · back to the volume booster.

Questions people ask

What does normalizing audio do?

It adjusts the overall level so the file reaches a target. Loudness normalization targets perceived loudness (LUFS); peak normalization targets the loudest sample.

What LUFS should I normalize to?

About −14 LUFS for music on Spotify and YouTube, about −16 LUFS for podcasts. −11 LUFS is loud, close to many commercial masters.

Does normalizing reduce quality?

Changing the level doesn’t, but the export re-encodes if you choose a lossy format. Export WAV or FLAC to avoid that.

Can I normalize several files to the same loudness?

Run each one through with the same target. Each file is measured on its own, so all of them end up at the same LUFS.

More tools on the bench