LUFS & Loudness Explained: The Complete Guide for Musicians

What LUFS actually measures, why it replaced the old peak-meter approach to loudness, and how loud your track really should be — for your mix, and for every streaming platform.

If you've ever exported a mix that sounded huge in your studio and then felt thin and quiet the moment you played it next to a released track, you've run into the gap between peak level and perceived loudness. LUFS is the measurement that closes that gap — and once you understand it, a lot of confusing mixing and mastering advice starts making sense.

What LUFS actually measures

LUFS stands for Loudness Units Full Scale. Unlike a simple peak meter — which just tells you the loudest single sample in your file — LUFS is modeled on how the human ear actually perceives loudness over time. It applies frequency weighting (we hear midrange frequencies as louder than deep bass or extreme highs at the same volume) and averages the signal across a window of time, rather than reacting to a single instantaneous spike.

That's why two tracks can have the exact same peak level (say, -1 dB) and sound wildly different in loudness. A track that's heavily compressed and dense will read much louder in LUFS than a sparse, dynamic track with the same peak — because LUFS is measuring the sustained energy your ear actually perceives, not just the tallest spike in the waveform.

The three LUFS numbers you'll see

• Integrated Loudness — the average loudness across the entire track, from start to finish. This is the number streaming platforms use to normalize your track.
• Momentary Max — the loudest moment, measured over a very short 400ms window. Useful for spotting the single loudest section of a song (a chorus hit, a drop).
• Short-Term Max — loudness averaged over a slightly longer 3-second window. Useful for judging how loud a whole section (verse, chorus) feels, rather than one instant.

Why streaming platforms care about LUFS

Before loudness normalization existed, the "loudness war" rewarded whoever mastered their track the loudest — because on shuffle-play, the loudest track simply felt more impactful, regardless of how it actually sounded. Streaming platforms ended that arms race by normalizing every track to a target LUFS level, so a quiet, dynamic track and a brick-walled loud one play back at roughly the same perceived volume.

That means mastering your track far louder than a platform's target doesn't make it sound louder to the listener — it just means the platform turns it down more, and you lose dynamic range for nothing.

So how loud should your mix actually be?

There's no single correct number — it depends on genre and how much dynamic range you want to preserve. But a few practical guidelines hold up well in practice:

• Mastering to roughly -14 to -9 integrated LUFS covers most modern genre norms without requiring extreme limiting.
• Very dense, aggressive genres (some EDM, trap, hard rock) often sit toward the louder end of that range and rely more on arrangement density than raw dynamic range for impact.
• Acoustic, jazz, and classical material generally sits quieter, sometimes well below -14 LUFS, because preserving dynamic range is part of the genre's character.
• Mastering dramatically louder than a platform's target mostly just costs you dynamic range at no perceptual benefit — the platform will turn it back down anyway.

A common mistake: chasing loudness at the expense of dynamics

Because LUFS rewards sustained energy, it's tempting to just slam a limiter on the master bus until the number looks "competitive." The problem is that this usually comes at the direct cost of dynamic range — the difference between the quiet and loud parts of your track — which is a big part of what makes a mix feel alive rather than flat.

A healthier approach: get the arrangement, balance, and mix decisions right first, then use loudness processing to reach a reasonable target — not to manufacture excitement that the mix itself isn't providing.