LUFS meter and loudness normalization with verified WAV export
Measure integrated LUFS, true peak and loudness range locally. Match mono or stereo WAV/FLAC files, compare before and after, and export verified 24-bit WAV audio.
Loudness Workbench
Use a real local file or the executable audio example. Compare saved before/after measurements and listen before accepting a normalization decision.
Supported scope: Mono/stereo only · 8–192 kHz · WAV / FLAC · 400 ms–10 minutes per file · 96 MiB input and 24 million decoded channel samples per batch · up to 3 files · 24-bit PCM WAV output. MP3/AAC, video, surround, extensible WAV and live microphone capture are outside this release.
Audio and delivery target
Required fields are marked with an asterisk.
Your result
Review your output, then save it to your device.
Make the next step clear.
Your findings, assumptions and next steps will appear here after you run the check.
Loudness Workbench · https://loudness.kitforma.com/ · KitForma
Integrated LUFS and oversampled true peak
Before/after audio and waveform comparison
Verified WAV plus batch CSV/JSON
Understand the evidence.
Focused guides explain the rules, the edge cases and the checks worth doing before you act.
Read LUFS and true peak without confusing them
Separate programme loudness, transient peaks and actual delivery requirements.
Read the guide ↗Field guide / 02Normalize a file while respecting a peak ceiling
Choose constant gain or deliberate limiting, then verify the encoded result.
Read the guide ↗Field guide / 03Compare several audio files with a reproducible report
Match intended loudness while keeping duration, dynamics and uncertainty visible.
Read the guide ↗Before you start
Does LUFS mean the same thing as peak level?
No. Integrated LUFS uses frequency-weighted signal energy and loudness gating across the programme. Sample peak checks stored samples; true peak estimates the reconstructed waveform between samples. A file can have quiet average loudness and very high transient peaks.
Why did the result miss my loudness target?
With preserve-dynamics processing, a peak ceiling can prevent the requested boost. The tool applies less gain and reports the achieved level. Limiting can bring average level closer, but it changes dynamics and can still require a final correction to respect the measured ceiling.
Does a verified ceiling guarantee every platform will sound the same?
No. The check applies to the actual exported PCM WAV and this meter. Lossy encoding, resampling, playback processing and different true-peak estimators can change peaks. Use the receiving platform’s current delivery requirements and leave appropriate headroom.
Why is loudness range provisional?
LRA uses a distribution of three-second windows. Below three seconds it cannot be measured here; during the first 60 seconds it is marked provisional because a short excerpt may not represent a programme’s full range.
Does the tool upload or resample audio?
No. WAV and FLAC are decoded locally at their stored sample rate. The meter supports mono and stereo and refuses surround layouts. Audio processing runs in a local worker; Cancel discards its pending result.
Can normalization repair clipping or improve a poor recording?
Lowering level creates headroom but cannot reconstruct clipped waveform detail. Limiting can introduce audible changes. Compare the original and exported audio, inspect the clipping counts and retain the source recording.
Sources & assumptions
Formats, algorithms and supported features are documented below. Check the stated scope and inspect your output before using it.
- EBU Tech 3341 ↗
Momentary and short-term windows, loudness gating, test definitions and LRA stability.
- EBU Tech 3342 ↗
Loudness-range gating and 10th/95th percentile calculation.
- ebur128 WebAssembly engine ↗
Apache-2.0 wrapper around the MIT Rust ebur128 implementation.
- libebur128 reference ↗
MIT reference K-weight filter derivation and true-peak interpolation.
Reviewed 2026-10-18. Read our methodology or report an issue.