Loudness targets · Film and series
Netflix loudness spec: −27 LKFS dialogue
Netflix asks for −27 LKFS ±2, with true peak under −2 dBTP. The important part is what gets measured: dialogue, not the whole mix. A standard full-mix meter, including MixOutpost's, measures something different, and this page is honest about what that means.
What Netflix asks for
Netflix's sound delivery specs set programme loudness at −27 LKFS ±2 with a maximum true peak of −2 dBTP. On paper that looks like a broadcast spec a few dB quieter than ATSC A/85. The difference is in how the −27 is measured.
Broadcast specs like EBU R128 and ATSC A/85 are usually measured across the whole programme. Netflix measures loudness dialogue-gated: only the parts where someone is speaking count. The reading reflects how loud the dialogue is, not the average of dialogue, music and effects.
Why dialogue gating exists
With a full-programme measurement, a film with a huge score and long battle scenes has to pull everything down to hit the average, dialogue included. A quiet two-hander can play its dialogue much hotter. Viewers then ride the remote between titles.
Anchoring the measurement on dialogue keeps speech at a consistent level from one title to the next, and leaves the mixer free to make the action loud and the quiet scenes quiet. That's also why −27 is lower than broadcast numbers: dialogue sits under the peaks of a cinematic mix, and the spec leaves room above it.
What a full-mix meter tells you
MixOutpost's meter, like most loudness meters, measures the whole mix to ITU-R BS.1770. It has no dialogue detection. So the Netflix target in the app carries a note: Netflix measures dialogue only; this is the whole mix.
- Action-heavy content: the full-mix reading tends to come out louder than the dialogue-gated figure, because music and effects add energy outside the dialogue.
- Sparse, dialogue-led content: the two readings tend to be closer, since most of the programme is speech.
- Long quiet ambiences: can pull a full-mix reading down while the dialogue-gated figure doesn't move.
Use MixOutpost's Netflix verdict as a guide while you mix, and for spotting scenes that jump. Confirm the final number with a dialogue-gated meter before delivery. If your session has a dialogue bus, putting a second MixOutpost plugin on it gives you a speech-only reading to compare, though it still includes the pauses between lines, so it isn't the same as dialogue gating.
What to aim for
- Set dialogue level first and build music and effects around it. It's what the spec anchors to.
- Keep true peak under −2 dBTP across all channels, with a true-peak limiter on the print master.
- Watch loudness range. A cinematic LRA is fine on Netflix, but viewers on TV speakers and phones will still reach for the remote if the explosions are far above the dialogue.
- Work from Netflix's current specs. They cover much more than loudness: formats, channel layouts, stems and deliverables.
Specs change
Netflix updates its partner delivery specifications. The figures here match the Netflix target built into MixOutpost at the time of writing. The spec document supplied for your project is the one to follow.
Checking a Netflix mix with MixOutpost
MixOutpost measures on your iPhone from exactly what your Mac sends, so you can check while you mix and listen on the phone the way many viewers do.
- 01
Stream the print master
Insert the MixOutpost plugin last on the master bus and press Start Sharing. Pair your iPhone once with a QR code.
- 02
Open Loudness and pick Netflix
Under TV and film delivery, choose Netflix. The app shows within spec, or too loud or too quiet by how many dB, with the note that this is a whole-mix reading.
- 03
Reset per reel or episode
Play from the top and read the result at the end. Readings are taken before Listen as, Dim, Mute or phone volume. Treat the verdict as a guide, and confirm with a dialogue-gated meter.
Netflix compared with other delivery specs
| Spec | Integrated | Tolerance | True peak |
|---|---|---|---|
| EBU R128 (Europe broadcast) | −23 LUFS | ±1 LU | −1 dBTP |
| ATSC A/85 (US broadcast) | −24 LKFS | ±2 dB | −2 dBTP |
| Netflix (dialogue-gated) | −27 LKFS | ±2 dB | −2 dBTP |
See every target, including music streaming and podcasts, on the loudness targets page.
Questions
What is the Netflix loudness spec?
−27 LKFS ±2, dialogue-gated, with a maximum true peak of −2 dBTP. That's the figure MixOutpost uses. Netflix publishes detailed sound delivery specifications for its partners; always work from the current version.
What does dialogue-gated mean?
The loudness is measured only over the parts of the programme where someone is speaking. Music and effects between lines don't count towards the reading. The aim is consistent dialogue level from title to title, while leaving the action free to be loud or quiet.
Can MixOutpost measure dialogue-gated loudness?
No. MixOutpost measures the whole mix, and the Netflix target in the app says so. A full-mix reading is a useful guide, but it isn't the number a dialogue-gated measurement will give. Confirm the final mix with a dialogue-gated meter.
Will my full-mix reading be higher or lower than the dialogue reading?
It depends on the content. A loud score or big effects push a full-mix reading up relative to dialogue; long stretches of quiet ambience pull it down. That's exactly why the two can disagree by several dB.
Hear your mix on your phone
Free Mac plugin (AU, VST3) and iPhone app. Lossless, bit-exact and encrypted.