Audio loudness normaliser
Match the listening level of your recordings. Measure LUFS and true peak, choose a target, then hear the difference before you save.
Add recordings, match their level, keep the WAVs and report.
Different recordings. One listening level.
Recordings 0
Give every recording a fair level.
Drop audio here or use Add audio.
Measure one file or match a whole batch.
Measure first, then normalise to see the achieved level.
- Original true peak
- —
- Result true peak
- —
- Applied gain
- —
- Peak reduction
- —
The measurement report includes every recording in your queue.
Limits and keyboard controls
Up to 20 files, 100 MiB each and 300 MiB in the queue. Each decoded file must fit 256 MiB of samples. Device memory may impose a lower limit. Mono and stereo only, with at least 0.4 seconds of audio.
In the audio preview, Space plays or pauses, B switches versions, ← and → move five seconds, and Home returns to the start. The WAV is measured again after processing. This is not a certified compliance meter.
Signed in, typed text and settings save automatically to your private account. Text opened from files stays in this browser’s local storage. Files stay in your browser. Account saving
How to normalise audio loudness
Measure first. Listen before delivery.
Add your recordings
Drop one file or a batch. Each recording gets an integrated LUFS and true-peak reading before anything changes.
Choose a target and protect the peaks
Start with a preset or enter your own target. Preserve dynamics, or limit peaks when a gain change is not enough.
Listen, then keep the audio and report
Compare the original and result at the same playback position. Download individual WAVs or one ZIP with measurement reports.
Using the loudness workbench
Your queue holds the original recordings and their results for this tab. Settings apply to every recording.
Add and inspect recordings
Use Add audio or drop files onto the panel. Measurement starts automatically, without playing or changing your audio.
| Control | What it does |
|---|---|
| Add audio | Adds MP3, WAV, M4A, FLAC, OGG and other audio your browser can decode. You can select several files. |
| Recording row | Selects a recording for its detailed readings and preview. Original and result LUFS stay visible in the queue. |
| Remove recording, × | Removes that source and result from the tab. It does not delete your original file. |
| Clear queue | Releases all recordings and results. Your delivery settings stay in place. |
| Try a three-recording example | Creates synthetic clips at different levels. The third includes a sharp peak to demonstrate a constrained target. |
Choose a listening level
Starting point sets the target and ceiling together. These presets are references, so check your destination’s requirements.
- Target loudness accepts -40 to -9 LUFS. A less negative value is louder.
- True-peak ceiling accepts -9 to -0.1 dBTP. This limits reconstructed peaks, which can exceed individual samples.
- Preserve dynamics uses one gain change. When the ceiling gets in the way, the result stays below target.
- Limit peaks uses linked gain across channels to reduce peaks. A final true-peak check may lower the whole result.
Changing a setting clears old results. Use the Normalise button again to make outputs for the new settings. Cancel processing stops the current run and keeps finished results.
Read the result and compare the sound
On target means the result is within 0.2 LU of your chosen value. Below target means peak protection or the gain limit prevented it.
Original and Result
The large numbers show integrated LUFS. The scale marks both readings against your target.
True peak
Original true peak and Result true peak show the estimated reconstructed peak in dBTP.
Gain and reduction
Applied gain is the global gain change. Peak reduction is the largest additional reduction made by the limiter.
Use Play and the Playback position slider to listen. Original audio and Normalised audio keep your position when you switch. This compares actual listening levels, so the louder version will sound louder.
Export audio and a measurement report
Download selected WAV saves the recording you are inspecting. Download batch ZIP saves all finished WAVs and CSV and JSON reports together. With one result, it reads Download WAV + report.
Download report CSV also works after measurement alone. The report names each file and includes available readings, output settings and any processing error. It is created locally and never saved to your account.
Focus mode and keyboard controls
Focus mode gives the panel more room. Choose Narrow, Wide or Full, then use Done or Escape to return.
| Key, with the audio preview focused | Action |
|---|---|
| Space | Play or pause. |
| B | Switch between original and normalised audio. |
| ← / → | Seek five seconds. |
| Home | Return to the start. |
What this tool will not do
It will not repair clipping, remove noise or independently level different speakers within one recording. It changes whole-file loudness, with optional peak limiting.
- Output is 24-bit WAV. It does not write MP3, preserve embedded tags or add loudness metadata.
- Mono and stereo only. Files shorter than 0.4 seconds cannot produce the required integrated measurement.
- Silence and fully gated audio are measured without being amplified.
- Memory limits apply: 100 MiB per source, 300 MiB per queue and 256 MiB of decoded samples per recording.
- A WAV keeps its supported sample rate. Other formats decode at 48 kHz here.
- The meter is not certified. Check formal broadcast delivery with a reference meter and the full specification.
When to use another audio tool
Finish noise removal and editing before normalisation. Cutting pauses or changing the sound afterwards can change the loudness again.
Some competing tools offer MP3 export or broader mastering chains. This workbench focuses on batch WAV delivery and a clear report of what it achieved.
Frequently asked questions
Can I normalise several audio files to the same loudness?
Yes. Add up to 20 recordings and choose one LUFS target for the queue. Each file is measured and processed separately. The batch ZIP contains the successful WAVs and CSV and JSON reports. A recording that cannot reach the target is marked Below target, with its achieved level shown.
Does this measure LUFS or just turn the volume up?
It measures K-weighted integrated LUFS using 400 ms blocks with 75% overlap and absolute and relative gates. It also estimates reconstructed true peak with four-times oversampling. The result is measured again after gain changes and PCM quantisation. It is not a certified compliance meter.
What loudness should I use for a podcast?
The Podcast preset starts at -16 LUFS and -1 dBTP. Apple recommends around -16 LKFS with a one-decibel tolerance and a true-peak ceiling of -1 dBFS. Check your host’s specification, especially for mono delivery. The guide explains why a mono and duplicated stereo signal can show different LUFS readings.
Why is my result below the LUFS target?
Your loudest peak may prevent the required gain. Preserve dynamics lowers the gain to respect your ceiling. Limit peaks reduces peaks before a final true-peak check, but may still finish below target. Gain increases also stop at 40 dB. The tool reports the achieved value instead of treating a requested value as a measurement.
Can I export a normalised MP3?
This tool exports 24-bit PCM WAV. MP3, AAC and other supported input formats are decoded first. A new WAV avoids another lossy encoding step, but cannot recover detail already lost in the original. Convert it in your audio editor if you need a compressed delivery format, then check the encoded file’s peaks again.
What is the file size limit?
Each source file can be up to 100 MiB. The queue holds up to 20 files and 300 MiB of source audio. Each decoded file must fit 256 MiB of samples. At 48 kHz, that is about 11.6 minutes of stereo audio or 23.3 minutes of mono. Your device may run out of memory sooner. Split longer recordings before processing.
Something missing? Request a feature