Logo

Audio Normalizer - Hit a Loudness Target

Normalize a track to the LUFS level streaming platforms and podcast apps actually expect, with a true-peak ceiling to keep it clean. Processed on your device and exported as MP3, WAV or M4A.

The loudness analysis and the normalizing pass both run in this tab. Your file is never uploaded.

音声ファイルをここにドロップ

またはクリックして端末から曲を選ぶ

MP3・WAV・M4A・AAC・OGG・FLAC · 200 MB まで

LUFS, true peak, and what loudnorm actually does

This tool runs FFmpeg's loudnorm filter, which measures your track's integrated loudness in LUFS, a unit built around how loud human hearing perceives a signal averaged over time, then applies gain so the result matches your chosen target. That target is separate from the true-peak ceiling, which caps the single highest point the waveform reaches, including inter-sample peaks a normal peak meter can miss but a real digital-to-analog converter can still clip on. A file can be correctly normalized in LUFS and still clip on some devices if the peak ceiling is set too high.

This is a single-pass loudnorm run, so the filter estimates the needed gain as it goes instead of scanning the whole file first for an exact correction. That keeps it fast enough to run entirely in a browser tab, and it lands close to the target on most tracks; very short clips or material with extreme swings between quiet and loud sections are most likely to end up slightly off the number, worth a quick listen to confirm.

Picking a target for where the file is going

−14 LUFS matches what Spotify and YouTube normalize uploaded audio to, so a track already sitting near that target will not be turned down when it reaches the platform. −16 LUFS is the common target for Apple Podcasts and general podcast delivery, slightly quieter to leave more headroom for spoken word. −23 LUFS follows EBU R128, the broadcast standard used across much of Europe, and is noticeably quieter than either of the streaming targets, built for consistency across an entire broadcast schedule rather than for a single track played on its own.

The true-peak ceiling works alongside whichever target you pick. −1 dBTP, the default, leaves a small safety margin for most consumer playback chains. Pushing the ceiling closer to −0.5 dBTP squeezes out slightly more perceived loudness at a higher clipping risk on some hardware; pulling it back toward −3 dBTP is the safer choice for material that will be played through a wide range of unknown devices, like a podcast episode.

Why loud masters get turned down anyway, and when to use something simpler

Streaming platforms apply their own loudness normalization on playback, turning tracks mastered louder than their target down to match it. Mastering or exporting a file far above −14 LUFS does not make it play louder on Spotify or YouTube; it only gives up dynamic range for no audible gain once the platform's own normalization kicks in. Normalizing to the platform's target yourself gets a predictable result instead of guessing how much a service will pull the level down.

Normalizing is the wrong tool when you just want a quiet recording turned up by a fixed amount rather than matched to a standard; a straightforward dB increase is simpler for that, which is what the volume booster is for. Reach for the normalizer when the destination cares about a consistent loudness level, whether that is a streaming upload, a podcast feed or a broadcast spec.

Audio normalizer blueprint

What each loudness target and the peak ceiling mean in practice.

Loudness targets−14 LUFS (Spotify, YouTube), −16 LUFS (Apple Podcasts), −23 LUFS (EBU R128 broadcast).
True-peak ceiling−3 to −0.5 dBTP, default −1 dBTP.
Best forUploads to streaming platforms, podcast episodes, broadcast delivery.
What it changesOverall gain to hit a target loudness, plus a hard limit on the highest peak.
Quality noteSingle-pass loudnorm; very short or highly dynamic clips may land slightly off target.
OutputMP3, WAV or M4A.
LimitsUp to 200 MB per file, processed one file at a time.

How to Normalize Audio Loudness Online

Three steps, all inside your browser

1

Add your audio

Drop in an MP3, WAV, M4A, AAC, OGG or FLAC up to 200 MB. It is read from your disk and never uploaded.

Tips: Normalizing works best on a finished mix; do it after any editing, not before.

2

Pick a loudness target and a peak ceiling

Choose −14 LUFS for Spotify and YouTube, −16 LUFS for Apple Podcasts, or −23 LUFS for broadcast, then set the true-peak ceiling between −3 and −0.5 dBTP.

Tips: The default −1 dBTP ceiling leaves enough headroom for most playback systems without sounding noticeably quieter.

3

Preview and download

Listen to the normalized result, then save it as MP3, WAV or M4A.

Tips: If the source was already very loud, normalizing will turn it down; that is expected and matches what streaming platforms do anyway.

Audio Normalizer Questions

How loudness targets work, and when to reach for something simpler.

No. The loudness analysis and the normalizing pass both run through a WebAssembly build of FFmpeg inside this tab, so the file stays on your device from start to finish. You can confirm this in your browser's network panel while normalizing.
MP3, WAV, M4A, AAC, OGG and FLAC up to 200 MB go in, and you can export the normalized file as MP3, WAV or M4A regardless of the source format.
LUFS measures perceived loudness averaged across the whole track, which is what streaming platforms actually match your file against. The true-peak ceiling is separate: it limits the single highest point the waveform reaches, including peaks that only appear once a DAC reconstructs the signal between samples, so the file does not clip on playback even after it is normalized to your loudness target.
This tool uses FFmpeg's loudnorm filter in a single pass, which adjusts gain and limits peaks in one run rather than analyzing the whole file twice first. It is accurate for most tracks, but very short clips or highly dynamic material can land slightly off the exact target; a full re-listen after normalizing is worth doing before you rely on the number.
A fixed volume boost adds the same number of decibels no matter how loud the file already is, so two tracks boosted the same amount can still end up at very different loudness. Normalizing sets every file to the same target level, which is what you want before uploading to Spotify, YouTube or a podcast feed, where the platform would otherwise turn an overly loud master down anyway.
Use the volume booster instead. It applies a straightforward dB increase without analyzing loudness, which is faster and more predictable when you are not delivering to a platform that enforces a loudness target.

Other one-click audio effects

Same idea, different part of the sound: bass, volume, fade and reverse tools that run entirely in your browser.

Just need a quiet file turned up by a set amount? AI 音楽生成

料金

あなたに合ったプランを選んでください。隠れた料金や予期せぬ費用はありません。

スターター

音楽の旅を始めよう

14.99
1ヶ月
USD
600credits1ヶ月
200 曲1ヶ月
100 回1ヶ月
2 個(並行タスク数)
AI音楽ジェネレーター
AI歌詞ジェネレーター
AIボーカル除去
AIステム分離
AIマスタリング
音楽アナライザー
AIカバージェネレーター
1000 回(音楽ダウンロード)
商用ライセンス
リミックス&編集
AI歌声モデルトレーニング
メールサポート
人気

ホビー

創造力を解き放とう

29.99
1ヶ月
USD
1500credits1ヶ月
500 曲1ヶ月
250 回1ヶ月
5 個(並行タスク数)
AI音楽ジェネレーター
AI歌詞ジェネレーター
AIボーカル除去
AIステム分離
AIマスタリング
音楽アナライザー
AIカバージェネレーター
無制限(音楽ダウンロード)
商用ライセンス
リミックス&編集
AI歌声モデルトレーニング
メールサポート

プロフェッショナル

プロ品質の音楽制作

49.99
1ヶ月
USD
3600credits1ヶ月
1200 曲1ヶ月
600 回1ヶ月
10 個(並行タスク数)
AI音楽ジェネレーター
AI歌詞ジェネレーター
AIボーカル除去
AIステム分離
AIマスタリング
音楽アナライザー
AIカバージェネレーター
無制限(音楽ダウンロード)
商用ライセンス
リミックス&編集
AI歌声モデルトレーニング
メールサポート