Make every file play at the same volume. Measure each file's integrated loudness (LUFS) and peak level, normalize to a streaming, podcast or broadcast target or to a peak level, and download WAV or MP3. Everything runs in your browser, so your audio is never uploaded.
No file at hand? to see how it works.
Normalize by
Advertisement
Advertisement
How to use the Audio Normalizer
Add your audio files: drop MP3, WAV, M4A, OGG, FLAC or other files on the box. Each file is measured right away.
Choose a mode: Loudness (LUFS) makes files sound equally loud; Peak sets the loudest sample to a fixed level.
Pick a target: for example -14 LUFS for streaming music, -16 LUFS for podcasts or -1 dBFS peak.
Pick the output format (WAV or MP3) and press Normalize.
Listen and download each file, or press Download all (ZIP).
Loudness targets by platform
Streaming services and broadcasters play audio back at a fixed loudness. If your file is louder, they turn it down; if it is quieter, most simply leave it quiet. Normalizing to the right target before you upload means your audio plays the way you mixed it.
Use
Integrated loudness target
Typical peak ceiling
Spotify, YouTube, Tidal, Amazon Music
-14 LUFS
-1 dBTP
Apple Music (Sound Check)
-16 LUFS
-1 dBTP
Podcasts (stereo), audiobooks for many platforms
-16 LUFS
-1 dBTP
Podcasts (mono)
-19 LUFS
-1 dBTP
EBU R128 broadcast (Europe)
-23 LUFS ± 0.5 LU
-1 dBTP
ATSC A/85 broadcast (USA)
-24 LKFS ± 2 dB
-2 dBTP
LUFS (Loudness Units relative to Full Scale) and LKFS are the same unit. A difference of 1 LU equals 1 dB, so a file at -20 LUFS needs +6 dB of gain to reach -14 LUFS. Platform targets change from time to time, so check the current specification if you deliver to a strict broadcaster.
How the normalizer works
Decode the file to 32-bit floating-point samples at its original sample rate.
Measure the sample peak and the integrated loudness with the ITU-R BS.1770 algorithm: K-weighting filter, mean square in 400 ms blocks with 75% overlap, an absolute gate at -70 LUFS and a relative gate 10 LU below the ungated average.
Calculate the gain: gain (dB) = target − measured value. In Loudness mode the target is in LUFS; in Peak mode it is the peak level in dBFS.
Protect the peaks: if peak + gain would pass the ceiling, either a 5 ms look-ahead limiter with an 80 ms release reduces just those peaks, or the gain is lowered until the peak sits on the ceiling.
Re-measure and encode: the result is measured again (shown after "→" in the list) and saved as WAV or MP3.
Because the gain is the same for the whole file, the balance between quiet and loud passages stays exactly as it was. To reduce that range, use a compressor in an audio editor; to boost a single file by a set number of decibels, use the Audio Volume Booster.
Frequently asked questions
What is the difference between peak and loudness normalization?
Peak normalization scales a file so its loudest sample reaches a set level, such as -1 dBFS. Two files with the same peak can still sound very different in volume. Loudness normalization measures perceived loudness in LUFS over the whole file and scales it to a target, so files sound equally loud. Use loudness normalization to make a playlist or podcast episodes consistent.
What LUFS should I use?
Common targets are -14 LUFS for Spotify, YouTube and Tidal, -16 LUFS for Apple Music and most podcasts, -19 LUFS for mono podcasts, -23 LUFS for EBU R128 broadcast in Europe and -24 LKFS for ATSC A/85 broadcast in the US. If you are unsure, -14 LUFS for music and -16 LUFS for speech are safe choices.
How is loudness measured?
The tool follows ITU-R BS.1770: each channel is K-weighted (a high-shelf filter plus a high-pass filter), the mean square is computed in 400 ms blocks overlapping by 75%, blocks quieter than -70 LUFS are dropped, then blocks more than 10 LU below the average are dropped, and the rest are averaged. Loudness = -0.691 + 10 × log10(sum of channel mean squares).
Why is a limiter needed?
Raising a quiet but dynamic recording to -14 LUFS can push its peaks above 0 dBFS, which would clip and distort. With the limiter option, only those short peaks are turned down smoothly to the ceiling (default -1 dBFS) while the rest gets the full gain. If you prefer no limiting, choose Lower the gain: the file then stays below the target loudness instead.
Does normalizing reduce quality?
Normalization only multiplies the audio by one gain value, which does not change the sound apart from the volume. The limiter, if used, changes only short peaks. Saving as WAV keeps full quality; saving as MP3 adds normal MP3 compression, so use 192 kbps or more for music.
Is the peak a true peak?
The tool measures sample peak, the largest sample value. True peak (dBTP) also counts peaks between samples that can appear after conversion, and can be up to about 0.5 dB higher on loud material. The default ceiling of -1 dBFS leaves room for that.
Are my files uploaded?
No. Files are decoded, measured and processed in your browser with the Web Audio API. MP3 files are encoded with a JavaScript build of the LAME encoder on your device.