Quick Reference
| Parameter | Requirement | Notes |
|---|---|---|
| Program loudness | -24 LKFS ± 2 LU | Pass range -26 to -22 LKFS, measured per ITU-R BS.1770-3 |
| Gating | Program (ungated) | No dialogue gating - different from Netflix |
| Max true peak | -2.0 dBTP | Oversampled true peak, hard requirement |
| Sample rate | 48 kHz | Required |
| File format | WAV or MOV (QuickTime) | MOV is accepted, unlike most broadcast specs |
| Encoding | PCM only | No lossy or bitstream deliveries |
| Channel formats | Stereo up to 5.1 | Loudness target applies across mix formats |
How Prime Video Measures These Values
Program loudness, not dialogue loudness
The -24 LKFS ±2 LU target is measured over the full program per ITU-R BS.1770-3, without dialogue gating. Music cues, effects and silence all count toward the integrated value. A mix leveled by dialogue anchor (the Netflix method) will read differently here, typically quieter, and can miss the -26 LKFS floor.
Broadcast headroom on the peaks
The -2.0 dBTP ceiling is a hard requirement measured as oversampled true peak. Limiting at -2.3 dBTP in the mastering chain keeps meter tolerance from ever deciding the outcome. Sample-peak metering under-reads inter-sample activity and is not sufficient for this check.
PCM in WAV or MOV
Delivery is 48 kHz PCM, packaged as WAV or QuickTime MOV. The MOV option is a Prime Video particularity: most broadcast and streaming specs take WAV or BWAV only. No lossy codecs, no AC-3 bitstreams in the delivery master.
One target across mix formats
Stereo and 5.1 deliveries measure against the same loudness target and peak ceiling. If you deliver multiple mix formats for the same title, each file has to pass on its own - a passing 5.1 mix does not certify the stereo fold-down.
FAQ
Is Prime Video measured like Netflix?
No, and this is the most common cross-platform mistake. Netflix measures dialogue-gated loudness at -27 LKFS; Prime Video measures program loudness (the full mix, no dialogue gating) at -24 LKFS ±2 LU per ITU-R BS.1770-3. The same master can pass one and fail the other. Measure each delivery against its own spec.
Where does the -24 LKFS target come from?
It is the ATSC A/85 target from US broadcast television, written into law by the CALM Act. Amazon adapted the broadcast standard for OTT delivery, which is why Prime Video and US broadcast share the same numbers: -24 LKFS ±2 LU with a -2.0 dBTP ceiling.
Is LKFS different from LUFS?
No. LKFS (ATSC terminology) and LUFS (EBU terminology) are the same unit: K-weighted loudness per ITU-R BS.1770. A -24 LKFS requirement and a -24 LUFS requirement describe an identical measurement. The gating mode is what actually differs between specs, not the unit.
Why is the true peak limit -2 dBTP instead of -1?
Broadcast-derived specs keep more headroom because downstream transcodes and legacy playout chains clip earlier than music streaming stacks. Best practice is limiting around -2.3 dBTP so measurement tolerance never pushes the reading over the -2.0 limit.
Can I deliver the same file to Prime Video and Netflix?
Usually not without re-levelling. A dialogue-heavy mix at Netflix's -27 LKFS dialogue-gated target typically measures quieter than -26 LKFS program loudness and fails the Prime Video floor. Check the file against both specs before assuming one master covers both platforms.