Audio Normalizer (LUFS & Peak)
Hit −14, −16 or −23 LUFS on the dot — measured, limited and checked.
Try before you buy.
- Free preview: the first half of the normalised sound (up to 30 seconds) with a short MySmartCoPilot chime every 10 seconds; the measured loudness, peaks and gain stay hidden.
- Locked until you unlock it: download and copy.
- Unlock: Pro pass, ₹179 for 30 days, a one-time payment that never renews.
Ways to unlock shows how to get the full result.
Printing this result is locked in the free preview.
| Measurement | Before | After |
|---|
Target
From −40 to −5. Audiobooks often use −18 to −20.
Your normalised file
Locked in the free preview. Opens the ways to unlock this result.
Locked in the free preview. Batch runs unlock with a pass.
Locked in the free preview. Query results unlock with a pass.
About the Audio Normalizer (LUFS & Peak)
Bring a recording to a standard loudness: −14 LUFS for music on streaming services, −16 LUFS for podcasts, −23 LUFS for broadcast (EBU R 128) — or set the loudest peak to a level, or choose your own loudness. The page measures the file first, exactly as ITU-R BS.1770-5 describes (K-weighting and gating), and shows its integrated loudness, loudness range, true peak, sample peak and loudest moments.
It then applies one fixed gain to the whole file, so the balance between quiet and loud passages stays as it was. Where that gain would push peaks over the true-peak ceiling (−1 dBTP by default), a look-ahead true-peak limiter holds them under it; or turn the limiter off and let the gain stop at the ceiling instead. Listen to 15 seconds of the loudest part before and after to compare.
The result keeps the file’s format where the browser can write it (MP3 stays MP3, WAV stays WAV, FLAC stays FLAC), and a video keeps its picture, copied unchanged, with only the sound re-encoded. Afterwards the page decodes the finished file and measures it again, so the “after” values are those of the file you download. Nothing is uploaded.
Without a pass, the result is a free preview: the first half of the normalised sound (up to 30 seconds) with a short MySmartCoPilot chime every 10 seconds — for a video, its sound alone — while the measured loudness, peaks and gain stay hidden. The full file and every measurement, before and after, come with a pass, or after unlocking this result.
How to use it
- Drop an audio file or a video on the box, or press Choose an audio file or a video. The page measures its loudness and peaks.
- Pick a target: Streaming, Podcast, Broadcast, Peak level or Custom, and a true-peak ceiling. The line below shows whether the limiter will work, and the gain with a pass or once a result of this file is unlocked.
- Press Play 15 seconds and switch between Original and Normalised to hear the difference (without a pass, the normalised side carries a short MySmartCoPilot chime).
- Choose how to save it and press Normalise: with a pass, or after unlocking this result, the page measures the finished file, shows the before and after values and offers Download (and Copy report for the numbers). Without a pass the page plays a free preview instead: the first half of the normalised sound (up to 30 seconds) with a short MySmartCoPilot chime every 10 seconds; for a video, its sound alone.
Examples
episode.mp3 · −22.4 LUFS integrated · true peak −6.1 dBTP
episode (normalised).mp3 · +6.4 dB · −16.0 LUFS · true peak −1.0 dBTP (the limiter held the few loud peaks)
master.wav · −8.3 LUFS · true peak +0.4 dBTP
master (normalised).wav · −5.7 dB · −14.0 LUFS · true peak −5.3 dBTP (turned down: no limiter needed)
interview.mp4 · −17.2 LUFS
interview (normalised).mp4 · picture copied unchanged · sound −5.8 dB → −23.0 LUFS (EBU R 128)
Which target to choose
- Streaming, −14 LUFS: Spotify plays music at −14 LUFS in its normal setting and recommends masters at −14 LUFS integrated with true peaks below −1 dBTP — or below −2 dBTP for masters louder than −14 LUFS (Spotify for Artists).
- Podcast, −16 LUFS: Apple recommends that podcast audio stays around −16 LKFS, give or take 1 dB, with true peaks not above −1 dBFS (Apple Podcasts audio requirements).
- Broadcast, −23 LUFS: EBU R 128 normalises programmes to −23.0 LUFS with true peaks of at most −1 dBTP (EBU R 128).
- Peak level: the loudest peak at −1 dBFS (or another level) — the classic “normalise” of audio editors. It does not make files equally loud, only equally peaky.
LUFS and LKFS are the same unit. Services that normalise playback turn loud files down anyway, so making a file louder than its target gains nothing and costs dynamics.
How it is measured
Loudness follows ITU-R BS.1770-5: each channel is K-weighted (a high shelf that models the head, then a high-pass), mean squares are taken over 400 ms blocks every 100 ms, blocks below −70 LUFS are ignored, then blocks more than 10 LU below the rest. Loudness range is measured as EBU Tech 3342 defines it, from short-term (3 s) loudness. True peak uses the 4× over-sampling filter of BS.1770-5 Annex 2, which finds peaks between the samples that a sample meter misses. The meter is checked against the test signals of EBU Tech 3341 and Tech 3342.
What the limiter does — and does not do
The limiter looks 5 ms ahead and turns the gain down smoothly just before a peak that would pass the ceiling, then lets it recover over about 150 ms. It counts inter-sample peaks too, so the result stays under the ceiling after digital-to-analogue conversion. Below the ceiling it changes nothing.
It is not a compressor: the gain is the same for the whole file, so quiet passages stay quiet relative to loud ones. If a target needs several decibels of limiting, the page says so — that much limiting can sound squashed, and a lower target is often better.
Limitations
- Lossy encoders (MP3, AAC, Opus) can move peaks by a few tenths of a decibel; the page measures the finished file and says if a peak ended up above the ceiling. A ceiling of −2 dBTP leaves room for that.
- One gain for the whole file: it does not even out a recording whose parts differ in level (that needs compression or manual editing).
- Ogg Vorbis and Opus files cannot be written here, so they are saved as MP3 (or another format you choose); a video’s sound is always re-encoded.
- One file at a time. Copy-protected files (DRM) and codecs browsers cannot decode cannot be normalised.
Privacy
Everything happens in your browser. What you enter or open here is not uploaded or stored by MySmartCoPilot. The audio engine and the encoders download from MySmartCoPilot the first time you open a file. Your audio and video are never uploaded.
Frequently asked questions
What do I get without a pass?
Without a pass, Audio Normalizer (LUFS & Peak) shows the first half of the normalised sound (up to 30 seconds) with a short MySmartCoPilot chime every 10 seconds; the measured loudness, peaks and gain stay hidden. Until you unlock it, the result can’t be downloaded or copied. A Pro, Premium or Ultimate pass, a one-time payment that never renews, unlocks the full result. The pricing page lists the passes and their prices.
What does “normalise to −14 LUFS” mean?
It means changing the level so that the file’s integrated loudness — its average loudness over the whole programme, measured as ITU-R BS.1770 describes — is −14 LUFS. Two files at −14 LUFS sound about equally loud, whatever their peaks.
What is the difference between peak and loudness normalisation?
Peak normalisation sets the highest sample (or true peak) to a level, for example −1 dBFS; a spiky recording then still sounds quieter than a dense one. Loudness normalisation sets the perceived loudness in LUFS, which is what streaming services, podcast apps and broadcasters use.
Which loudness should I use?
For podcasts, −16 LUFS with true peaks at most −1 dBTP, as Apple recommends. For music on streaming services, −14 LUFS, the level Spotify plays at. For broadcast, −23 LUFS (EBU R 128). For anything else, use what the platform or client asks for: Custom takes any value from −40 to −5 LUFS.
Why is the result a little away from the target?
The page measures the finished file, after encoding. Heavy limiting lowers the loudness slightly, and lossy encoders can change it a little. Results within ±0.5 LU are shown as on target; the page explains anything further off.
Does normalising a video re-encode the picture?
No. The picture is copied packet for packet, so it keeps its quality and timing; only the sound is decoded, normalised and encoded again (AAC for MP4, MOV and MKV, Opus for WebM), and it stays in time with the picture: the AAC encoder’s start-up delay is hidden in the file’s edit list, as players expect.
Is my file uploaded?
No. Measuring, limiting and encoding happen in your browser; only the page’s code is downloaded.