Streaming Loudness Targets: Every Platform, Sourced

Spotify normalizes playback to −14 LUFS in both directions; Apple Music targets −16 LUFS with Sound Check off by default; YouTube, Amazon, Tidal and Deezer only turn loud tracks down. No social platform publishes a target. Master to −1 dBTP true peak and let each platform level. Every number on this page carries its source.

Hi, I'm Trina with BeamTracer.com — this one's the loudness chart itself: what every streaming platform really does to your master at playback, with receipts.

Every "LUFS chart for 2026" you've seen blends a handful of officially published numbers with a large amount of folklore, and almost none of them tell you which is which. This page does. Every number below carries a source class — OFFICIAL when it comes from the platform or standards body's own publication, COMMUNITY when it's measured or consensus behaviour the platform could change tomorrow — and a per-row verified date. Changes land in the changelog at the bottom, and the whole dataset is downloadable as JSON.

Two tools live on this page. The first answers "what happens to a master at level X?" — drag the sliders. The second answers "what happens to my master?" — drop the file; it never leaves your browser.

What happens to your track at any loudness level?

Set an integrated loudness and a true peak below, and every platform row recalculates: how much gain that platform applies, in which direction, and where the track lands. A loud master is turned down by the difference; a quiet one is turned up only on the platforms that boost, and only as far as its true-peak headroom allows.

The widget starts at a deliberately instructive setting: −20 LUFS integrated, peaking at −5 dBTP. Most loudness calculators will tell you Spotify plays that track at −14. It doesn't, and the reason is the single most under-explained rule in this whole subject:

A platform will boost a quiet track only as far as its true-peak headroom allows. Spotify's ceiling is −1 dBTP. A track peaking at −5 dBTP has 4 dB of headroom, so a −20 LUFS master gets +4 dB and lands at −16 LUFS — two decibels short of the target, permanently. Slide the true peak up and down and watch the cap move; this is also why turning your limiter's output down "for safety" on a quiet master quietly costs you playback level.

Notice what the other rows do with the same quiet master: YouTube, Amazon, Tidal and Deezer simply refuse to boost at all. Up-normalization is the exception, not the rule.

Is −14 LUFS a target I should master to?

No — it's a playback reference, not a mastering target, and the distinction matters. Normalization is a gain offset applied at the listener's end; it doesn't reward you for arriving at the reference level, and nothing bad happens when you don't. Analyses of large commercial catalogs bear this out: RoEx's analysis of 200,000+ tracks across 30 genres (David Ronan, AES Show) found roughly 79% of masters exceed Spotify's −14 and 91–92% exceed Apple's −16 — essentially the entire industry masters past the reference and gets turned down, and none of those records broke.

Master the music at the loudness it wants. The one number that behaves like a rule is true peak: stay at or below −1 dBTP (−2 for very loud masters), because lossy encoders overshoot — more on that below.

Does Spotify turn quiet tracks up?

Yes — Spotify is the notable platform that normalizes in both directions, and it publishes the behaviour: −14 LUFS by default, with listener-selectable Loud (−11) and Quiet (−19) modes. Three fine points from its own documentation: the boost is capped by true-peak headroom (the widget above models this), a limiter engages when boosting dynamic material — so a quiet, dynamic master played on Loud gets squeezed in ways you never approved — and normalization is not applied on the web player or most third-party devices. Albums played straight through are normalized on album gain, so quiet interludes stay quiet relative to the singles.

Turning down, by contrast, is a clean negative gain. Spotify states plainly that no additional distortion occurs — the "Spotify will destroy your loud master" claim has the physics backwards.

What does YouTube actually do?

YouTube publishes no loudness documentation at all — every number you've seen for it is measured, not sourced. The measured behaviour: uploads are analyzed server-side and louder-than-reference content is attenuated toward roughly −14 LUFS. It's always on, with no user toggle on the main platform. You can see it on any video: right-click → Stats for nerds → the "Volume / Normalized" line, with a content-loudness figure showing exactly how much your upload is being pulled down.

Two wrinkles worth knowing. Stable Volume is a separate, playback-time leveler: on by default, it "balances the range between quiet and loud" within a video as it plays, so quiet passages do come up on the listener's device. It does not change the upload-time rule — a quiet upload is still never boosted as a whole — and YouTube turns it off for official music videos and YouTube Music, so for music the old "never turns anything up" answer still holds. And YouTube Music appears to use a far louder reference (independent measurements put it near −7 LUFS), which is why most music passes through it untouched and people wrongly conclude it doesn't normalize. That conflicts with every chart listing YT Music at −14 — we flag the conflict rather than pick a side silently.

Is Apple Music really −16 LUFS?

Mostly, with two asterisks the charts leave out. Sound Check — Apple's normalization — is off by default on iOS and macOS, so a large share of Apple Music listeners hear no normalization at all. And the −16 figure itself is community-measured: Apple's own Digital Masters documentation emphasizes clipping and inter-sample-peak control, not a single integrated target. The official numbers Apple does publish are for Atmos (integrated not above −18 LKFS) and Apple Podcasts (−16 LKFS ±1, true peak ≤ −1) — which is where the podcast convention on this page comes from.

What about TikTok, Instagram and the other social platforms?

No social platform publishes a loudness spec. None. The "−10 to −12 LUFS for TikTok and Instagram" numbers that circulate in blog posts have no traceable source — they're folklore copied between articles. What's actually documented: Meta uses xHE-AAC with dynamic loudness management that adapts to playback context (no single number), and independent measurements of TikTok's in-feed behaviour conflict — some engineers measure normalization near −14, others find none at all. We ship no number for these platforms because there is no number to ship.

Why does every platform ask for −1 dBTP?

Because of what happens after you upload. Lossy encoders — AAC, Ogg, Opus, MP3 — reconstruct the waveform and can push peaks roughly a decibel above the original, worst at low bitrates. A master that touches 0 dBFS on a sample-peak meter can clip audibly on the listener's DAC after transcoding. The −1 dBTP ceiling is the standard margin for that overshoot; Amazon asks for −2, and −2 is the common recommendation for masters well above −14 because louder masters overshoot harder. This is also why the standard measures true peak — with 4× oversampling between samples — rather than sample peak: a sample meter can under-read the reconstructed waveform by up to about 0.7 dB.

Why does mono measure 3 dB louder than stereo?

Because BS.1770 sums channels. The "same" audio delivered as one mono channel measures about 3 dB hotter than as two stereo channels, which is the entire explanation for the podcast convention pair: −16 LUFS stereo ≈ −19 LUFS mono. They're one loudness written down two ways, not two opinions. If your podcast host asks for −16 and your file is mono, deliver −19 — matching "−16" on a mono file makes you 3 dB louder than every stereo show on the platform.

Does loudness affect royalties or the algorithm?

No. Normalization is a playback gain applied after every recommendation and accounting decision has been made; a −8 master and a −14 master with the same play count earn the same. The related belief — that louder masters perform better on streaming — has no credible evidence behind it either: playback is level-matched, and listener studies find no reliable preference for the more compressed version at matched loudness. The loudness war's incentive is gone; what survives normalization is dynamics, which is a whole post of its own.

How do you measure your own master's loudness?

Drop the file into the analyzer below. It decodes and measures the whole track on your machine — integrated loudness, true peak, loudness range — and reports what each platform will do with those numbers. The file never leaves your browser.

The measurement follows the same ITU-R BS.1770-5 and EBU R128 standards as BeamTracer's Loudness meter, with a ±0.1 LU measurement target. For the full instrument — live metering, five views, history, counters, session export, audition — open the Loudness tab; it's free.

Which streaming loudness claims are false?

These specific claims circulate widely enough to correct by name. Each was checked against the platform's own documentation.

"Spotify turning your track down can cause clipping or distortion." (Disc Makers, November 2024.) Backwards. Attenuation is clean negative gain — Spotify's documentation states no additional distortion occurs. The clipping risk lives on the boost side (limiter engagement) and in lossy transcoding of over-hot masters.

"Target −11 to −13 LUFS for YouTube." (Hollyland, and others.) No basis — YouTube's video platform is attenuate-only, so mastering above the reference just buys attenuation. The same article's "YouTube never turns quiet audio up" is still true of the upload normalization; only the playback-time Stable Volume leveler lifts quiet passages, and it is off for music videos.

"Instagram and TikTok favor −10 to −12 LUFS; Facebook targets −13." (OpusClip, among many.) Fabricated precision — none of these platforms has ever published a loudness target.

"Google Podcasts: −14 LUFS." (SONE's 2026 standards roundup.) Google Podcasts was discontinued in 2024. A 2026 chart still listing it is a chart copied from an old template — treat the rest of it accordingly.

"YouTube only adjusts volume if the user enabled normalization." (VI-Control forum, widely repeated.) False — it's server-side and always on; Stats for nerds shows it acting on any video.

Can you download or cite this loudness dataset?

Yes — the whole dataset is published under CC BY 4.0. Download the target references, source classes, and verification dates shown on this page:

• /loudness-targets.json — every row with its target, ceiling, direction, defaults, source class and verified date. CC BY 4.0; attribution link appreciated.
• Cite this page: BeamTracer, "Streaming Loudness Targets," verified 2026-08-22, beamtracer.com/blog/streaming-loudness-targets.

@misc{beamtracer-loudness-targets,
  title        = {Streaming Loudness Targets: Every Platform, Sourced},
  author       = {{BeamTracer}},
  year         = {2026},
  howpublished = {\url{https://www.beamtracer.com/blog/streaming-loudness-targets}},
  note         = {Dataset verified 2026-08-22}
}

Changelog

Every substantive change to the numbers on this page lands here, dated. A re-verification that finds nothing changed updates the per-row verified dates only.

Next

• LUFS in plain language — how to read every number this page throws around
• What Spotify actually does to your track — the flagship walkthrough of the boost cap and audition preview
• How dynamic is your master? PSR, PLR and LRA — what survives after everyone plays at the same loudness
• Level a multi-person livestream in the browser — apply the same loudness standards to live voices

That's the whole chart. beamtracer.com — I'll see you in the next one.