What is LUFS? Loudness Normalisation Explained for Voice Actors

LUFS stands for Loudness Units Full Scale. It's the measurement standard that every major audio platform — YouTube, Spotify, Apple Podcasts, ACX, broadcast networks — uses to normalise loudness. If you're delivering voiceover professionally and don't understand LUFS, you're guessing at one of the most important technical specs in your workflow.

This guide explains what LUFS actually measures, how it differs from other loudness metrics, and what targets you need to hit for every platform.

Why loudness standards exist

Before loudness normalisation became standard, the "loudness war" in music meant that louder masters simply sounded better on playback because they hit the listener harder. The same problem existed in broadcasting: louder advertisements would interrupt quieter programming and jolt listeners.

Streaming platforms solved this by normalising all content to a standard loudness target on playback. Audio that's louder than the target gets turned down; audio that's quieter gets turned up. The result is consistent perceived volume across all content on the platform — whether you're listening to a quiet audiobook or a heavily compressed pop song.

For voice actors and audio producers, this means delivering your audio at the platform's target loudness gives you exactly the listening experience you intend. Delivering louder than the target means the platform turns you down — you don't get a loudness advantage, you just get normalised anyway, often with less favourable results.

What LUFS actually measures

LUFS is a perceptual loudness measurement. Unlike peak metering (which measures the maximum amplitude of a waveform) or RMS metering (which measures average power), LUFS is weighted to how human hearing works — more sensitive to midrange frequencies (where voices sit), less sensitive to very low and very high frequencies.

There are three variants you'll see:

• Integrated LUFS — loudness measured across the entire file, from start to finish. This is the number that matters for platform delivery. When ACX says "−18 to −23 LUFS," they mean integrated LUFS.
• Short-term LUFS — loudness measured over a rolling 3-second window. Useful for monitoring loudness during a session but not used for delivery spec.
• Momentary LUFS — loudness over a 400ms window. The most responsive reading, useful for identifying peaks in a signal but not for delivery.

When someone says "normalise to −14 LUFS," they always mean integrated LUFS unless otherwise specified.

Platform LUFS targets: the complete reference

Each platform has a different normalisation target. Here are the ones that matter for voiceover work:

• YouTube: −14 LUFS integrated. The most common target for YouTube narrators, documentary VO, and e-learning content.
• Spotify: −16 LUFS for podcast and spoken word content.
• Apple Podcasts & most podcast platforms: −16 LUFS integrated.
• ACX (Audible/Amazon): −18 to −23 LUFS integrated. This is a range — you must land within the window, not just below a ceiling. This is stricter than most other platforms.
• EBU R128 (European broadcast standard): −23 LUFS integrated. Used by the BBC, European broadcasters, and most corporate video production houses.
• ATSC A/85 (North American broadcast): −24 LUFS integrated.

LUFS vs dBFS: what's the difference?

dBFS (decibels Full Scale) measures the peak amplitude of an audio signal — the highest point the waveform reaches before clipping. LUFS measures perceived loudness. These are different things and solve different problems.

A voice recording can have the same peak level (say, −6 dBFS) but very different LUFS measurements depending on how much dynamic range the recording has, how long the pauses are, and how compressed or processed it is. A voiceover file with long silences between sentences will have a lower integrated LUFS than the same voice with tight silence trimming and no gaps — even at identical peak levels.

This is why silence trimming and LUFS normalisation interact: removing excess silence raises your integrated LUFS measurement, which means you need less gain increase to hit your target loudness. Doing silence removal before loudness normalisation produces cleaner results than doing it afterwards.

True peak limiting: the other spec you can't ignore

Most platforms also specify a maximum true peak level alongside LUFS. True peak differs from standard peak metering because it accounts for inter-sample peaks — distortion that occurs between digital samples when audio is converted to analogue. ACX requires −3 dBTP maximum; most streaming platforms require −1 dBTP.

A true peak limiter at the end of your processing chain is non-negotiable for professional delivery. Without it, loudness normalisation can inadvertently push inter-sample peaks above the ceiling, causing distortion on playback that you wouldn't hear in your DAW but listeners absolutely will.

How to hit LUFS targets without a DAW

The manual workflow — import file, run loudness analysis, apply gain adjustment, check true peak, export — takes 5–10 minutes per file and requires a loudness metering plugin. Across a 30-chapter audiobook, that's real time.

CleanCut VO normalises to your chosen platform target in a single automated step. Select YouTube, ACX, Spotify, podcast, or broadcast from a dropdown — the output hits the correct integrated LUFS target with true peak limiting applied. No metering plugin, no iterative renders, no guesswork.