What the letters stand for

LUFS = Loudness Units relative to Full Scale.

• Loudness Units — a number that represents perceived loudness, not raw signal level
• relative to Full Scale — measured against the digital ceiling of your file (0 dBFS)

The "perceived loudness" part is what makes LUFS useful. Your DAW's peak meter shows the loudest single sample in a file. Your ear doesn't hear single samples — it hears averages, weighted toward frequencies that the human ear is more sensitive to. LUFS models that.

Why LUFS exists when we already had dB

dB measurements (dBFS in digital audio) show signal level. A track that peaks at -3 dBFS has a maximum sample that's 3 dB below the digital ceiling. That tells you nothing about how loud the song sounds.

Two songs can both peak at -3 dBFS. One can sound twice as loud as the other if the average level is higher and the dynamic range is smaller. dB alone can't tell you that. LUFS can.

LUFS was standardized so loudness could be measured the same way across rooms, monitors, and engineers. Before LUFS, every studio had its own "I think this sounds about right" approach to loudness — and the streaming era needed something more consistent than that.

The three LUFS readings you'll see

Your loudness meter shows three numbers. They mean different things and answer different questions.

Momentary LUFS

A 400ms snapshot — the loudness of the last 0.4 seconds of audio. Bounces around as the song plays. Useful for spotting loud moments (a chorus hit, a snare crack) but not for judging the song overall.

Short-term LUFS

A 3-second sliding average. Smoother than momentary, still responsive enough to track section-to-section changes. Useful for hearing whether your chorus is meaningfully louder than your verse, or whether the bridge dropped too far.

Integrated LUFS

A single number that represents the loudness of the entire song from start to finish, gated to exclude silence and quieter background sections. This is the number that matters for streaming. When Spotify says "we normalize to -14 LUFS," they mean integrated LUFS.

If you can only look at one number, look at integrated LUFS.

How LUFS measures perceived loudness

The trick that makes LUFS more useful than raw dB: it applies a frequency-weighting filter called K-weighting before measuring. K-weighting models how the human ear hears loudness across frequencies:

• Bass below ~60 Hz gets slightly de-emphasized (we hear sub-bass as less loud than its actual signal level)
• Treble above ~2 kHz gets slightly boosted (we hear that range as more prominent than its signal level)

The K-weighted signal then gets averaged over time (with two-stage gating to exclude silence and very quiet sections). The result is a loudness reading that correlates with how loud a human listener perceives the track.

This is also why a track full of low-frequency bass can have high peak levels but a relatively quiet LUFS reading — your ear doesn't perceive the sub the same way the meter sees the dB. The K-weighting accounts for that.

(The full standard is EBU R 128 plus ITU-R BS.1770-4 for the underlying loudness measurement algorithm.)

What integrated LUFS numbers mean for your music

A rough mental map:

Integrated LUFS	Practical feel	Where you'll hear it
-23 LUFS	Broadcast spec — dynamic, conservative	EBU broadcast standard, classical recordings
-16 LUFS	Quiet streaming master	Apple Sound Check default, classical/ambient on streaming
-14 LUFS	Spotify, YouTube, TIDAL and Amazon normalization target	Streaming-aware indie and singer-songwriter
-12 LUFS	Comfortably loud streaming master	Most indie rock, electronic, pop in 2026
-10 LUFS	Loud streaming master	Modern pop, contemporary hip-hop
-8 LUFS	Aggressive loudness	Trap, hyper-pop, EDM with heavy limiting — accepts streaming penalty for genre character
-6 LUFS and below	Loudness-war territory	Old CD-era masters, deliberately destructive limiting

There's no "right" LUFS for your song. There's only what works for the genre, the platform, and what you want the song to sound like after the streaming platform normalizes it.

True peak is different from LUFS

A common confusion: LUFS and True Peak measure different things and you need to watch both.

• LUFS — how loud the average is
• True Peak (dBTP) — the highest momentary peak in the file, including inter-sample peaks the streaming codec might add

A master at -10 LUFS with peaks at -0.1 dBFS sounds loud and brick-walled. A master at -10 LUFS with peaks at -1.0 dBTP sounds loud but has codec-safe headroom. Same LUFS, different true peak, different result on Spotify.

I wrote more about this in True peak vs sample peak — what your meter isn't telling you.

How to read your own track's LUFS

The free LUFS Mix Analyzer reads any audio file in the browser — drop the file, get integrated LUFS, short-term LUFS, and true peak in a few seconds. It's the same EBU R 128 / ITU-R BS.1770-4 implementation that pro mastering tools use, validated against iZotope Insight 2.

If you want to compare two of your own masters at different LUFS — say, hear what your song sounds like at -14 vs -10 — the Streaming Preview tool plays them at the level each platform would, so the comparison is fair.

The 30-second version

• LUFS measures perceived loudness, not signal level
• Integrated LUFS is the one number that matters for streaming
• Spotify, YouTube, TIDAL and Amazon Music normalize to about -14 LUFS; Apple Sound Check normalizes to -16
• Higher LUFS (louder masters) gets the streaming penalty — the platform lowers playback, but you don't get the dynamics back
• True peak is a separate measurement; both LUFS and TP matter

LUFS isn't complicated once you stop thinking of it as a fancier dB. It's a different measurement that answers a different question: how loud does this sound to a human?