What LUFS Actually Measures#
LUFS stands for Loudness Units Full Scale. It's a standardized way of measuring perceived loudness, defined by the ITU-R BS.1770 specification — the same reference every major streaming platform uses. Unlike a peak meter, which only tells you how close the waveform gets to clipping, LUFS approximates how loud the music actually sounds to a human ear. It's the first meter I look at and the last one I trust before a bounce.
It does this with a K-weighting filter that emphasizes the frequencies we're most sensitive to (roughly the upper midrange) and de-emphasizes deep sub-bass, then averages the energy over time. One loudness unit (LU) is effectively one decibel, so turning a track down 3 dB drops its integrated loudness by about 3 LU. Simple, honest math — the kind I wish more metering involved.
This is why I retired the old RMS-and-hope approach years ago. A track can hit 0 dBFS peaks and still feel quiet, or sit at -6 dB peaks and feel enormous. LUFS measures the thing listeners actually perceive. If the difference between the level of the signal and the level you hear is still fuzzy, my free lesson on Volume vs Gain clears it up in ten minutes.
Platform Targets in 2026#
Here's what each major platform actually does with your master in 2026 — I keep a version of this table taped near my monitor. The important column is the last one: what happens to a track that arrives louder than the target.
| Platform | Normalization Target | True Peak Ceiling | Loud Masters Get… |
|---|---|---|---|
| Spotify | ≈ -14 LUFS integrated | -1 dBTP | Turned down |
| YouTube | ≈ -14 LUFS integrated | -1 dBTP | Turned down |
| Apple Music | ≈ -16 LUFS integrated | -1 dBTP | Turned down |
| Beatport / club | None (DJ playback) | -0.3 to -1 dBTP | Played as-is |
Notice the pattern: the streaming services all normalize, and they only ever turn loud tracks down — they don't boost quiet ones past their peaks. Beatport and club playback are the exception: there's no normalization between the file and the DJ's channel fader, which is why club masters still live in the -8 to -6 LUFS range. I play out regularly, so most of my releases target both worlds — and it's exactly why I built the Mixing & Mastering Bundle: it gives me the metering and dynamics tools to print a streaming version and a club version from the same session.
"Master for the song, not for the meter. The meter works for you, not the other way around." — Monakai
My practical takeaway after years of printing both versions: -14 LUFS is a reference point, not a requirement. Master to whatever loudness serves the music, keep true peaks under -1 dBTP, and let normalization handle the rest.
Why Louder Is Not Better Anymore#
I lived through the CD era, and louder genuinely won back then — a louder disc jumped out of the carousel. Streaming killed that advantage. When Spotify normalizes your -6 LUFS master down by 8 dB to match everything else, the loudness is gone but the damage you did to get there stays: flattened transients, pumping limiter artifacts, a kick that hits like a damp sponge.
Meanwhile, the track it plays next to — mastered at a sane -10 LUFS with real dynamics — keeps all of its punch at the exact same playback level. The loudness war didn't end in a treaty. It ended because normalization made the prize worthless.
"If your master only works loud, it doesn't work." — Monakai
Chasing loudness isn't wrong; chasing it by smashing a limiter is. A loudness maximizer that protects transients lets you reach competitive club loudness without the sponge. That's exactly why I built LOUD by Monakai — I wanted something that pushes integrated loudness up while preserving the transient detail that normalization-era listeners (and algorithms) reward.
Short-Term vs Integrated vs True Peak#
Any decent loudness meter shows at least three numbers — and each answers a different question I ask during every mastering session:
- Momentary (400 ms window): How loud is it right now? Useful for spotting whether a drop actually feels bigger than the build.
- Short-term (3 s window): The loudness of the current phrase. Watch this during choruses — if your chorus short-term is barely above the verse, your arrangement has a dynamics problem no limiter will fix.
- Integrated: The gated average across the whole track, with silences ignored. This is the number Spotify's -14 refers to, and the number on your final bounce.
- True peak: The peak level of the waveform between samples, which is what a DAC actually reconstructs. Intersample peaks above 0 dBTP distort on playback, so hold your ceiling at -1 dBTP.
In my mixes, an integrated reading of -9 LUFS with choruses peaking at -6 short-term and verses sitting at -14 means healthy dynamics. If short-term and integrated are within 2 LU of each other for the whole song, you've made a very loud sausage. I've made a few myself — they paid rent, but I'm not proud of them. For the gain-staging fundamentals underneath all of this, the Mixing Basics lesson is the right starting point, and the metering section of the Mixing & Mastering Bundle puts all four numbers on one screen — which, frankly, is why I built it.
Try It Yourself in Your DAW#
Open your current project and run this check before your next export:
- Insert a loudness meter last on the master bus. Use a BS.1770-compliant meter (free options like Youlean Loudness Meter work in Ableton, FL Studio and Reaper — see the DAW compatibility guide for install paths). Reset it, then play the track from the top.
- Read the integrated value after a full playthrough. Note where it lands. Anything above -10 LUFS integrated is club-master territory; -14 to -12 is a comfortable streaming zone.
- Check the true peak readout. If it exceeds -1 dBTP, your converter will clip on playback even if your DAW meters look clean.
- Set your final limiter's ceiling to -1.0 dBTP and enable oversampling (4x or higher) so the limiter actually catches intersample peaks.
- Raise loudness with transient-safe processing. Drive LOUD by Monakai until integrated loudness hits your target, then pull back 1 dB. If you can hear the kick losing its click, you have gone too far.
- Simulate normalization. Turn your master down by the difference between your integrated LUFS and -14 (e.g. -8 LUFS master → pull it down 6 dB) and A/B against a reference at matched level. This is what Spotify listeners will actually hear.
- Bounce and re-import. Drag the exported file back into the session and meter the file itself. Trust the printed file, not the plugin chain.
Do this once per release and loudness stops being a superstition and becomes a checkbox. More workflow guides live on the Monakai Audio music production blog, and the full plugin lineup is in the catalog.