Before and after
The editor shows the original peak, calculated gain, and predicted peak before export.
Normalize audio online
to a −1 dBFS sample peak
Measure the decoded file's highest sample, calculate one linear gain adjustment, preview it, and export from the same browser-based audio graph.
Sample-peak measurementAutomatic −1 dBFS targetLocal processing
You can also drag and drop an audio or video file anywhere on this page.•Compatibility depends on the file codecs and device: MP3, WAV, M4A, AAC, FLAC, OGG, WebM, MP4, MOV, MKV, and more
Peak normalization is a repeatable level adjustment
This tool finds the largest absolute decoded sample and applies the same gain ratio to every channel. It is useful for setting a predictable sample-peak ceiling, but it does not measure perceived loudness.
Channel balance
One multiplier is used for every channel, preserving the relationship between left and right.
Preview equals export
Playback and export use the same gain value in the shared processing graph.
How to normalize an audio peak
Use the measured values as a check, not just an automatic button. The result is most useful when you know what the target represents.
- 1
Open and measure the file
Choose the recording and wait for the browser to decode every channel. The original sample peak is measured before any gain is applied.
- 2
Review the calculated gain
The tool targets −1 dBFS with one multiplier for the whole file. Check the predicted peak and note when the 200% gain cap prevents the target from being reached.
- 3
Preview and export
Listen to quiet and loud passages, then export. If you need matched perceived loudness across several files, continue with a dedicated LUFS workflow.
Important limits
- This is sample-peak normalization, not LUFS loudness normalization or true-peak limiting.
- A file that already reaches 0 dBFS may already be clipped; lowering it cannot reconstruct flattened samples.
- The automatic gain is limited to 200%, so very quiet files may remain below −1 dBFS.
- Compressed export re-encodes the result and can introduce codec-dependent changes.
Useful for music, voice, and podcasts when the goal is a peak ceiling
- Bring a quiet recording closer to a consistent maximum without manual gain calculation.
- Reduce a hot file to leave a small sample-peak margin.
- Prepare clips for a later loudness workflow while keeping this step clearly separate from LUFS matching.
Sample peak, true peak, and LUFS are different measurements
Choosing the right measurement avoids treating peak normalization as a solution for every loudness problem.
MeasurementWhat it describesHow to use it here
Sample peak (dBFS)
The highest decoded digital sample in the file.
This is what the current normalizer measures and targets at −1 dBFS.
True peak (dBTP)
An estimate of peaks that can occur between samples during playback or conversion.
This tool does not calculate it; use a true-peak meter or limiter when delivery rules require dBTP.
Integrated loudness (LUFS)
A weighted measurement of perceived loudness over time.
Use it to match programs or episodes; a shared peak value does not guarantee equal perceived volume.
Normalization runs locally
The selected audio is decoded and processed in browser memory. It is not uploaded to the application server to calculate or export the normalized result. Site resources and consented analytics may still use network requests.
Peak-normalization questions
Clear answers about what changes, what stays intact, and when another audio process is needed.
Will every normalized file sound equally loud?
No. Two files can have the same sample peak and very different average energy, dynamics, and perceived loudness. LUFS matching is a separate process.
Can normalization repair clipped audio?
No. Lowering a clipped recording creates headroom but does not reconstruct waveform details already flattened at the recording or encoding stage.
Does normalization change the stereo balance?
The same linear gain is applied to all channels, so the existing level relationship between left and right is preserved.
Why can the predicted peak remain below −1 dBFS?
Automatic amplification is capped at 200% for safety. A very quiet source may need more gain than that cap allows.