Walk into any mixing forum and ask what loudness target you should master to, and the answer comes back fast: -14 LUFS, that is what Spotify wants. The number has become so familiar that it is treated like a universal constant. It is not. Each major platform has its own rules, and the rules are asymmetric in ways that change what your master actually sounds like in someone else's earbuds.
The short version: most platforms turn loud tracks down. Some boost quiet tracks up, but with limits. One major service sits 2 LU below the rest. And the only reason -14 LUFS feels universal is that nobody has had a strong commercial reason to deviate from it, even though the AES committee that studies loudness recommends a different number.
What Each Platform Actually Does
Spotify targets -14 LUFS integrated by default. Loud masters get attenuated to hit the target. Quiet masters get boosted toward the target, but the boost is capped so the true peak stays at or below -1 dBFS. Spotify documents this explicitly: a master at -20 LUFS with only -1 dB of headroom will only be boosted to -16 LUFS, not all the way to -14, because the extra gain would push peaks past the ceiling. Users can switch to a Loud setting (-11 LUFS, with a limiter applied to soft material to reach target) or a Quiet setting (-19 LUFS, attenuation only). The limiter only ever runs in the Loud mode.
Apple Music uses Sound Check, which targets approximately -16 LUFS. This is the only major service aligned with the AES recommended practice for streaming distribution. Sound Check has been on by default on iOS since around 2022. There is a wrinkle: published sources disagree on whether Sound Check boosts quiet content or only attenuates loud content. The cautious assumption is that it behaves like Spotify in spirit, with attenuation as the dominant case and only conservative boosts on the way up.
YouTube targets -14 LUFS, attenuation only. Quiet tracks play at their original level. A master well below the target will sound noticeably softer than the surrounding content in autoplay.
Tidal targets -14 LUFS and normalizes per album, not per track. This matters for record listening: the relative levels you set between tracks on an album survive playback intact, where on Spotify track-mode playback would flatten them. Tidal also only turns loud tracks down.
Amazon Music targets -14 LUFS but reportedly requires -2 dBTP of true peak headroom rather than the -1 dBTP common elsewhere, because the Echo and Alexa playback chain is unusually prone to inter-sample clipping. The -2 dBTP figure comes from a single industry source, so treat it as a strong recommendation rather than gospel until Amazon publishes its own delivery spec.
Deezer is reported to sit at around -15 LUFS, between Apple and the rest. As with the Amazon headroom number, this comes from one published table rather than a Deezer-published source.
The Cap That Quiet Masters Hit
The asymmetry that surprises people most often is Spotify's boost cap, because it inverts a piece of advice that has circulated for years. The folk wisdom says: do not chase loudness, master with dynamics intact, Spotify will turn quiet tracks up. The first two parts are still good advice. The third is wrong in the way it is usually told.
Spotify will boost a quiet master, but only up to the point where the loudest true peak hits -1 dBFS. A -22 LUFS master with -3 dBTP gets boosted by 2 dB and lands at -20 LUFS in playback, not -14. A -16 LUFS master with -3 dBTP gets boosted by 2 dB and reaches -14, hitting the target cleanly. The platform never tells you which case you are in. There is no public dashboard for an artist to look at the gain Spotify decided to apply.
In practice this means a deliberately conservative master, the kind a careful engineer would describe as having room to breathe, can sit several LU below where the producer thought it would. It still has its dynamics, the transients are intact, the mix is undamaged. It is just permanently quieter than the surrounding catalog.
The -16 LUFS That Almost Was
The AES Recommended Practice AES77-2023, ratified in July 2023 and built on the earlier TD1008 document, recommends -16 LUFS for distribution loudness of popular music and -18 LUFS for speech content. Apple's choice of -16 LUFS for Sound Check aligns with this recommendation. Almost nobody else does.
The reasoning behind -16 LUFS is straightforward: it leaves more headroom for transient preservation, reduces the loudness pressure on dynamic genres, and gives mastering engineers a target that does not require as aggressive limiting. The reason -14 LUFS won the market is harder to articulate cleanly. Spotify chose it first, became the volume leader nobody wanted to be quieter than, and the rest of the industry followed. The 2 LU gap between Apple and everyone else is the visible residue of that history.
One Master, Not Five
The platform-by-platform breakdown can make it sound like mastering for streaming requires custom delivery for each service. It does not. The asymmetries cancel out in a useful way.
If you master to roughly -14 LUFS integrated with peaks at or below -1 dBTP, every platform behaves reasonably. Spotify accepts it without attenuation and without much boost. YouTube and Tidal accept it as is. Apple Music turns it down by 2 dB, which is inaudible in any context where the user is not A/B comparing to another service. Amazon may apply a hair more attenuation if the -2 dBTP guideline is real. There is no engineering case for delivering separate masters per platform.
What does matter:
- Mastering louder than -14 LUFS gains nothing on the loud platforms. Spotify, YouTube, and Tidal will simply attenuate the playback. The dynamics are gone, the platform takes back the loudness, and the listener hears a quieter, more squashed version of what you made.
- Keep at least 1 dB of true peak headroom, 2 dB if you went hot. Lossy codecs amplify inter-sample peaks. A file that reads -0.3 dBTP before encoding can read +1.0 dBTP after AAC at streaming bitrates.
- Master to the loudest track on the album. Tidal and album-mode Spotify both preserve inter-track relationships. Mastering each track to a uniform integrated loudness destroys the dynamic shape of a record.
- For Spotify QA, listen on Normal. That is what most casual listeners use. Loud is the only mode where Spotify applies its own limiter, and it is the right mode to listen on for stress-testing how your dynamics react to platform-side processing.
The Number Was Never the Point
LUFS targets describe playback level, not creative intent. A platform that normalizes to -14 LUFS guarantees that listeners will hear your track at roughly the same perceived volume as the track before and after it in a playlist. It does not guarantee that your master will play at -14 LUFS. It does not guarantee that the dynamics you preserved will survive intact. And it does not promise that the target itself is the right artistic choice for your record.
The useful frame is not what number to chase. It is what each platform will do to the file you upload, and what choices that leaves you. The answer in 2026 is: master with the headroom and the dynamics you want, watch your true peaks, do not bother optimizing per platform, and accept that the platforms have homogenized loudness in ways that mostly help listeners and rarely hurt records that were not already chasing loudness in the first place.