LUFS for Spotify, Apple Music and YouTube: the 2026 guide

Every time you upload a song to Spotify, Apple Music or YouTube, the platform measures its volume and adjusts it so it plays at the same level as the rest of the catalog. The standard they all use has a name: LUFS. Understanding what it means and how it affects your music is an essential part of the mastering process.

This guide explains what LUFS are, why they matter, what targets each platform uses and —more importantly— why it makes no sense to master thinking only about the number.

What is LUFS?

LUFS stands for Loudness Units Full Scale. It's a standardized measure of perceptual loudness: it doesn't measure the instantaneous peak of a signal, but how the human ear perceives the volume of audio over time. The scale is logarithmic and has 0 LUFS as its absolute reference point, representing the loudest possible signal without distortion. In practice, loudness values are expressed as negatives: the closer to 0, the higher the level.

There are three LUFS measures worth distinguishing:

• Integrated LUFS. Measure the average loudness of the whole song from start to finish. It's the value streaming platforms use for normalization. When people talk about "Spotify's target", they're talking about integrated LUFS.
• Short-term LUFS. Measure loudness in three-second windows. Useful for detecting especially loud or quiet sections within a track.
• Momentary LUFS. A 400-millisecond window. Useful for monitoring loudness in real time during mastering.

It's important not to confuse LUFS with dBFS. dBFS (decibels Full Scale) measure the peak level of the digital signal, that is, the maximum value the waveform reaches at a specific instant. A file can have high peaks in dBFS and at the same time a low integrated loudness, if those peaks are brief transients with a lot of silence around them. They're two different metrics measuring different things.

Related to dBFS is true peak, expressed in dBTP. Unlike the standard digital peak, true peak calculates the peaks that appear between samples when the audio is converted to analog in the player. True peak can exceed 0 dBFS if uncontrolled, causing distortion on playback. That's why all platforms recommend a true peak limit of approximately -1 dBTP to avoid clipping on decoding.

Loudness normalization per platform

The idea behind normalization is simple: when moving from one song to another in a playlist, the listener shouldn't have to adjust the volume. To achieve this, each platform measures the integrated LUFS of each track and adjusts the playback gain until it gets close to its internal reference.

The key thing to understand varies by platform and playback mode. Tracks above the reference are always turned down in all modes. Those below depend on the case: Spotify has three modes —Normal (the default in many contexts), Loud and Quiet— and in Normal and Loud it does apply positive gain to raise tracks that don't reach its -14 LUFS reference, also engaging a limiter in Loud mode to control peaks; only in Quiet does it not raise the volume of quieter tracks. Spotify, for example, will normalize a track mastered to -8 LUFS by turning it down several dB on playback in any mode; Apple Music and other platforms also normalize toward their own reference.

The practical effect is clear: mastering well above the platform's target gives you no advantage. Your track won't sound louder than the others because the platform will turn it down. The only thing you lose by mastering excessively loud is dynamic range: to reach that high loudness, the limiter has to crush the differences between the loud and quiet moments, removing the punch that makes a song hit.

Spotify has normalization on by default, although the user can disable it from the app settings. Apple Music normalizes similarly with its Sound Check feature, and YouTube applies its own normalization algorithm to all videos. In every case, the principle is the same: the platform's loudness target marks the real ceiling of your playback.

Loudness target table per platform

The following values are indicative references for 2026. Platforms can adjust their algorithms at any time and the exact values may vary slightly depending on playback mode or region. Treat them as guides, not absolute rules.

Indicative values for 2026. They may vary with platform updates. Use them as a working reference, not a fixed technical specification.

Apple Music deserves a separate mention: its Apple Digital Masters program recommends delivering files in high resolution (24-bit / 96 kHz or higher) so its encoder gets the best possible result. It's not a different loudness requirement, but a delivery-quality recommendation to get the best possible result from its encoding process.

The "loudness war" and why normalization made it counterproductive

For decades, the music industry experienced what's known as the loudness war: the race to make each release sound louder than the competition's. The logic was simple and well documented: in a direct comparison, louder audio sounds subjectively better for a few seconds. Labels pushed for masters to be louder and louder, and engineers delivered them by crushing the dynamics with ever more aggressive limiters.

The result was an era of hyper-compressed masters where the dynamic range —the difference between soft and loud moments— shrank to a minimum. Entire albums that sound flat, fatiguing, without the natural punch of the instruments. Drums with no attack, guitars with no air, the vocal pinned to the ceiling at all times.

Streaming platform normalization changed the rules of the game. Now, mastering everything to the max no longer gives an advantage because the platform levels the field. But it has a real cost: if you crush the dynamics to reach -8 LUFS in the master, when the platform turns it down to -14 LUFS on playback, your song will sound just as flat and compressed. The listener will hear a track with no punch next to one mastered more carefully at the same playback volume. The loudness war, in the streaming context, is a race where you only lose.

Why you should NOT master "for the algorithm"

A common question is: "exactly how many LUFS do I have to master to for Spotify?" The question starts from a wrong premise: that chasing the exact number is the goal of mastering.

The goal of mastering is for your song to sound as good as possible in all the contexts where it will be heard: cheap headphones, reference speakers, the phone speaker, the car system, the monitors of a club. Platform normalization is a consequence of the distribution system, not a creative parameter you should optimize.

A well-made master has tonal balance, dynamic range appropriate to the genre, and an output level consistent with market references. If those elements are in place, the resulting loudness naturally lands in the range platforms handle without issue. You don't need to measure the number and adjust it: good technical judgment takes you to the target without chasing it.

Besides, each genre has its own conventions. An electronic dance track can work with more compressed dynamics than an acoustic piece or a ballad, and that's perfectly legitimate. What matters isn't that they all reach exactly the same number, but that the result translates well across all systems and that playback normalization doesn't penalize the work done.

If you have doubts about how to prepare your mix so it reaches the mastering process well, the guide on what differentiates mixing from mastering explains what the mastering engineer expects when receiving the file.

Frequently asked questions