NORMALIZE

Manual

1. What NORMALIZE does

   NORMALIZE adjusts the level of an audio file to a target value. It has two modes: peak normalize, which matches the loudest point of the waveform to a target, and loudness normalize, which matches LUFS — a measure closer to perceived loudness — to a target.

2. How peak normalize works

   It detects the loudest point (peak) in the file and applies a single gain change so it lands exactly on the target peak level (-1.0dBFS or -0.1dBFS). The relative dynamics of the waveform don't change.

3. How loudness normalize works

   It measures LUFS (Loudness Units Full Scale) based on the EBU R128 / ITU-R BS.1770 standard and processes in two passes to hit a target loudness. The first pass measures the file; the second pass uses that measurement to normalize accurately.

   Preset	Target	Typical use
   Streaming	-14 LUFS	Most streaming platforms, including Spotify and YouTube
   Apple Music	-16 LUFS	Delivering to Apple Music
   EBU R128 broadcast	-23 LUFS	TV and radio broadcast

4. Peak vs. loudness

   Peak only looks at the single loudest instant, while loudness measures perceived loudness across the whole track. Two tracks with the same peak level can have very different loudness if one is quiet and sparse and the other is dense and loud. Loudness normalize suits delivery to streaming platforms; peak normalize suits simple clipping prevention.

5. Use case: matching loudness for a streaming platform

   Before uploading a track to Spotify or YouTube, run loudness normalize with the -14 LUFS preset to avoid the platform's own loudness normalization changing the volume in an unpredictable way. Measuring first with the LOUDNESS tool helps you see how far off the target you currently are.

6. Use case: leveling volume across multiple tracks

   When you play a batch of files recorded at different times or with different gear, the volume differences between them can be distracting. Normalizing each one to the same target — peak or loudness — reduces that gap during playback.

7. How it works: in-browser processing and privacy

   Measurement and processing happen entirely in your browser (a WebAssembly build of ffmpeg). Before/after values are shown side by side in the Readout, and the file is never uploaded anywhere.

8. Limits and troubleshooting

   Loudness normalize is a precise two-pass method, but true-peak limiting (to avoid clipping) can leave a margin of about ±0.5dB from the target. Processing runs in your device's memory, so a very large file (roughly over 1GB) can fail due to low memory.

9. FAQ

   Which mode should I use?

   For delivering to Spotify, YouTube, Apple Music and similar platforms, use loudness normalize. If you just want to even out volume differences, peak normalize is simpler.

   It didn't land exactly on the target.

   Loudness normalize is a precise two-pass method, but a margin of about ±0.5dB can remain. True-peak limiting (to avoid clipping) can also nudge the result slightly away from the target.

   Which file formats are supported?

   WAV, MP3, FLAC, AAC, M4A, OGG and most other common audio formats that ffmpeg can read.

   Can I just measure the level first?

   Yes — the LOUDNESS tool measures integrated loudness (LUFS) and true peak. Measuring first, then normalizing to a target here, is a good workflow.

   Does it overwrite the original file?

   No. The result is created as a new file inside your browser; the original file is never modified until you download it.