A finished mix has no neutral home. Every platform you might deliver to listens with a different ruler in its hand, and most of those rulers disagree by thirteen loudness units or more. Hand the same master to YouTube and to Netflix and one of them will reject you on intake.
This is not a quirk. It is the structural condition of finishing in 2026. There are at least five widely-enforced loudness standards in active use, and they are not converging. The studios that pretend otherwise are quietly redoing masters in the week after delivery.
The five rulers, in plain numbers
Music streaming — Spotify, YouTube Music, Tidal, Amazon Music — normalize to about minus fourteen LUFS integrated. Loud, broad, modern. Apple Music sits at minus sixteen LUFS via Sound Check. A little more conservative because Apple cares about playlist consistency more than peak loudness.
European broadcast follows EBU R128 at minus twenty-three LUFS integrated, with a tolerance of half a loudness unit. United States broadcast lives under the FCC's CALM Act at minus twenty-four LKFS, true-peak ceiling minus two dBTP. Both targets exist to protect viewers from the commercial-break loudness wars of the late nineties. Both are non-negotiable for cable and over-the-air carriage.
Then video streaming. Netflix, Amazon Prime Video, Disney Plus, and HBO Max all target minus twenty-seven LKFS, measured dialog-gated. Thirteen LU quieter than Spotify. Quieter than European broadcast. Quieter than anything else in the building.
“A master is not a number. It is a destination.”
Why Netflix went so much lower
The minus twenty-seven target sounds wrong on first read. It is not. Netflix is protecting cinematic dynamic range — the gap between a whisper and an explosion that defines feature-film sound design. If you normalize a film to music-streaming loudness, you flatten that gap to nothing. The whisper rises. The explosion does not. The mix loses its argument.
Dialog-gated measurement compounds this. Instead of integrating loudness across the whole timeline, the gate measures only moments where speech is present. Music passages, ambience, and effects are excluded from the headline number. A documentary with quiet voiceover and loud archive music will measure very differently under EBU R128 than under Netflix's spec, even when nothing about the file has changed.
What this means for budgeting
There is no single deliverable. There is a target list. The honest version of a quote acknowledges that and prices for it.
- A web master at minus sixteen LUFS for client social cuts
- A broadcast master at minus twenty-three or minus twenty-four LKFS, depending on territory
- A streaming master at minus twenty-seven LKFS dialog-gated, with a Dolby Atmos pass if the platform requires it
- Optional: a YouTube-specific master at minus fourteen LUFS for music-driven content
- Optional: a stem package — dialog, music, effects — that lets the client conform any of the above without re-bouncing the mix
Most studios bury this on a deliverables sheet at the back of the SOW. We put it on the front page. Not because we like talking specs to clients — most of them do not need to care — but because the price you quote should reflect the work you actually have to do. Five masters is not one master with five names.
If your post-production partner quoted you a single loudness number, ask which one. If the answer is silence, you are paying for one master and inheriting four problems.
