Loudness Targets by Platform: A Reference Table

Last checked: 5 October 2026

Every platform normalises. You cannot win by mastering louder - the platform simply turns you down, and all you have left is the dynamic range you crushed to get there. The only thing worth doing is hitting the target cleanly and keeping the headroom.

This is the table for loudness targets by platform we work from. Where a delivery document exists, the delivery document wins.

Music Streaming (Loudness Targets by Platform)

Platform	Integrated Loudness	True Peak
Spotify	-14 LUFS	-1 dBTP
Apple Music	-16 LUFS	-1 dBTP
YouTube	-14 LUFS	-1 dBTP
Amazon Music	-14 LUFS	-1 dBTP
TIDAL	-14 LUFS	-1 dBTP
Deezer	-15 LUFS	-1 dBTP

Apple sitting 2 LU quieter than everyone else is the one that catches people out. If you master to -14 and it goes to Apple Music, it is turned down 2 dB more than you expected relative to the rest of the playlist.

Dolby Atmos Music (Loudness Targets by Platform)

Spec	Requirement
Integrated Loudness	-18 LUFS
True Peak	-1 dBTP or lower
Master Format	ADM BWF, 48 kHz / 24-bit
Also Required	A stereo fallback master

Measure on the Dolby Atmos Renderer's own meter, not your DAW's. They do not always agree, and the renderer is the one the distributor's QC will use. The -18 figure is the single most common delivery failure we see. An Atmos mix sent at streaming-stereo loudness is 4 LU hot and gets rejected or squashed.

Film and Television (Loudness Targets by Platform)

Standard	Loudness	True Peak
Netflix	-27 LKFS plus or minus 2, dialogue-gated	-2 dBTP
EBU R128, Europe	-23 LUFS plus or minus 0.5, whole programme	-1 dBTP
ATSC A/85, US	-24 LKFS, whole programme	-2 dBTP

The Netflix number is the one people misread. -27 LKFS is dialogue-gated - it measures only the dialogue, not the programme. A mix that reads -27 overall is far too quiet. Gate to dialogue, hit -27 there, and let the rest of the mix sit where it belongs around it.

Theatrical Release

Cinema is the exception on this page, and the exception matters: there is no loudness normalisation in a cinema. Every other platform here measures your file and turns it up or down. A cinema does not. The room is calibrated instead, and the fader sits at reference. So there is no LUFS target for a feature. What there is, is a calibrated monitoring chain - and if you mix in a room aligned to the standard, the level you hear is the level the audience gets.

Parameter	Value
Calibration Signal	Pink Noise at -20 dBFS
Screen Channels, L C R	85 dB SPL, C-weighted, Slow
LFE	Plus 10 dB relative, so 95 dB from the same signal
Peak Headroom, Main Channels	105 dB
Peak Headroom, LFE	115 dB
Playback Fader	Reference, 7.0 on a Cinema Processor

Surround arrays are aligned to the screen channels, but the per-speaker figure depends on how many speakers the array has. That is a room calibration question, not a delivery spec.

What this means in practice. If your mix room is calibrated to 85 dB and the cinema is calibrated to 85 dB, the translation is exact. That is the entire point of the standard, and it is why a feature mix has no loudness number attached to it. Mix it to sound right at reference and it plays right.

The corollary is the trap: a mix made in an uncalibrated room, or monitored quietly, almost always comes out too compressed and too loud. There is no normaliser downstream to rescue it.

What to do with this

Ask for the delivery document first. Every broadcaster, streamer and distributor publishes one, and the numbers in it beat anything you remember or read in a forum. The table above is a starting point, not a substitute.

Measure integrated, not momentary. Integrated loudness is the figure every spec is written against. A momentary reading tells you about one moment; it tells you nothing about whether the programme passes.

Leave the true peak ceiling alone. The ceiling is there because lossy codecs overshoot on decode. A file that peaks at -0.1 dBTP can clip after it has been encoded to AAC or Opus, on hardware you will never hear.

One master, many folddowns. Do not remix for each platform. Mix once to the strictest target in the chain, check the folddowns, and let normalisation do its job. Chasing each number separately produces a set of masters that do not match each other.

Mixing to a spec you have not worked with before? Send it over with the material and we will tell you what it actually requires.