How to normalize audio
- Drop your audio files into the box or choose them. Each one is normalized on its own.
- Choose Peak to set the loudest moment to a level, or Loudness (LUFS) to make files sound equally loud. For peaks, −1 dB is the safe choice for MP3 and M4A. For loudness, −14 LUFS suits music streaming, −16 podcasts and −23 broadcast.
- Check the note next to each file, which says how many dB were added or taken away, then listen and download.
Peak and loudness normalization
There are two kinds of normalization, and they answer different questions.
| Peak normalization | Loudness normalization | |
|---|---|---|
| Measures | The single loudest sample | Average loudness over time (LUFS) |
| Goal | Use the full range without clipping | Make files sound equally loud |
| Changes the sound | No, only the volume | Only the volume, unless peaks need limiting |
| This tool | Yes | Yes |
Peak normalization is predictable and never clips, which makes it right for getting the most out of a quiet recording. Its limit is that one short spike, like a clap or a plosive “p”, decides the result for the whole file. If a recording barely gets louder after normalizing, a spike like that is usually the reason.
Good times to normalize
- Quiet recordings: voice memos, lectures and field recordings often peak far below full scale. Normalizing brings them up as far as they can go cleanly.
- Sound effects and samples: giving every clip in a set the same peak level makes them easier to work with in a game or sampler.
- After mixing to mono: a recording with sound on only one side comes out about 6 dB quieter when converted to mono, and normalizing wins that back.
- Before joining clips: it gets files into the same range before you join them, though speech and music may still sound different in loudness.
Normalizing can’t repair audio that was already clipped when it was recorded: the flattened peaks stay flat, just at a different level. Your files are processed in your browser and never uploaded.