What "Deliverables" Means in Audio Post
Deliverables are the set of audio files that constitute the finished product to the destination. The set has three layers.
The masters. The full mix in its final channel configuration: stereo, 5.1, 7.1, Atmos, IMAX 6-track. The masters are what plays to the audience.
The stems. Component versions of the mix that allow the destination to remix or replace elements. Standard stems are dialogue (DX), music (MX), and effects (FX), with optional Foley, ADR, backgrounds, and other splits depending on the project. Stems exist for re-mixing, international localization, and platform-specific adaptations.
The deliverables documentation. The conform documents (EDL, change list), the QC report, the loudness measurement report, and the production audio archive. These are not audio files but they are part of the deliverable set, and they are required for sign-off.
The delivery checklist sits at the end of the audio post production workflow. Everything upstream (AAF prep, dialogue edit, mix) feeds into it. Mistakes upstream surface as deliverable failures, which is why a strong delivery checklist is also a backstop on the rest of the workflow.
Loudness Specifications
Three loudness regimes cover most professional audio post deliveries. The applicable regime is set by the destination, not by the project.
Broadcast US (ATSC A/85). Loudness target -24 LKFS, tolerance ±2 LU. True peak ceiling -2 dBTP is the conservative practical norm. ATSC A/85 is enforced by the CALM Act in the United States and is the operational standard for terrestrial and cable broadcast.
Broadcast Europe (EBU R 128). Loudness target -23 LUFS, tolerance ±1 LU. Maximum true peak -1 dBTP. The Loudness Range (LRA) is also measured, with practical guidelines that vary by program type. EBU R 128 is the standard across European broadcasters.
Streaming. Each platform publishes its own loudness specification, and the targets differ from broadcast. Netflix, Disney+, Apple TV+, Amazon Prime Video, and the major audiobook and music platforms each maintain delivery documents that specify the loudness target, true peak ceiling, and channel layout requirements for content destined for that platform. The platform's current delivery specification is the source of truth; loudness targets do change between document versions, and a recent version is the only safe reference.
Theatrical. Theatrical mixes are not measured to a loudness target in the broadcast sense. The mix is calibrated to a sound-pressure level reference (typically 85 dB SPL at the mix position) and the final loudness is an artistic decision within the dynamic range the format supports. Theatrical deliverables specify channel layout and format (DCP audio standards), not LUFS targets.
The loudness measurement is taken on the final master, not on an intermediate. For a 5.1 deliverable, the measurement runs across all six channels per the ITU-R BS.1770 algorithm that ATSC A/85 and EBU R 128 both use. The QC report documenting the measurement accompanies the master at delivery.
Stem Requirements
Stems separate the mix into component layers so the destination can remix or substitute elements. The standard stems for a professional delivery:
Dialogue (DX) stem. Production dialogue and ADR mixed at delivery level. For projects requiring an ADR-only sub-stem, that sub-stem is delivered as a separate file (ADR stem) alongside the full DX stem.
Music (MX) stem. Score and source music. For projects with separate license-cleared and original music, those may split into separate sub-stems.
Effects (FX) stem. Sound design, hard effects, and (in some delivery conventions) Foley. The exact split varies by platform; some specify a separate Foley stem.
Backgrounds (BG) or Ambience stem. Atmos beds, room tone, and ambient continuity. Sometimes folded into the FX stem; other deliveries require it separate.
M&E stem. Music and Effects combined, free of dialogue. Used by international localization teams to dub other languages; covered in detail below.
Each stem is delivered in the same channel layout as the master (stereo, 5.1, 7.1, Atmos as applicable). For a 5.1 master, the DX stem, MX stem, and FX stem are each 5.1 files, summing to the master when added back together. The stem export automation covers how stems are printed from a Pro Tools session and what to verify on the print.
M&E (Music and Effects) Requirements
The M&E stem is the deliverable that allows international versions to be created without remixing from the original session. It contains music, effects, Foley, and backgrounds. It does not contain dialogue.
The discipline that makes a usable M&E:
Foley is included in M&E. Footsteps, prop handling, cloth movement, and other Foley sit in the M&E because the international version still needs them. A footstep does not change between languages.
No dialogue at all. Production dialogue, ADR, walla (background dialogue), and any other voice element is absent. This is the property that makes the M&E useful for dubbing.
Hard effects without dialogue mask. Effects that play under dialogue in the original mix (a door slam, a phone ring) sit at full level in the M&E because the dub will replace the dialogue and may need the effect at full presence.
Ambient continuity. The room tone, ambience beds, and atmospheric layers carry through the M&E so the dubbed version has the same acoustic environment as the original.
The M&E is built during the mix, not bolted on at delivery. A mixer building toward a clean M&E avoids dialogue-keyed sidechain decisions, keeps Foley at independent levels, and prints the M&E as a parallel deliverable while the dialogue stem is also being printed. M&E deliverables that get assembled after the fact (by muting the DX stem in the final mix) often miss the under-dialogue effects and produce a stem that international dubs cannot use cleanly.
File Naming and Metadata
Delivery file names follow a structured convention that lets the receiving platform identify, sort, and ingest the files correctly. A common pattern:
{Project}_{Episode}_{Format}_{Channel}_{Version}_{Date}.wav
A specific example:
ShowTitle_S01E01_Master_5.1_v02_20260506.wav
ShowTitle_S01E01_DX_5.1_v02_20260506.wav
ShowTitle_S01E01_MX_5.1_v02_20260506.wav
ShowTitle_S01E01_FX_5.1_v02_20260506.wav
ShowTitle_S01E01_M&E_5.1_v02_20260506.wav
The exact convention is destination-specific. Netflix, Disney, Apple, and broadcast networks each publish file naming requirements that include episode codes, channel layout markers, language codes, and version identifiers. The platform's delivery specification is the source of truth; the convention shown above is a generic baseline.
Embedded metadata in BWF audio files carries delivery-relevant information that survives the file transfer: timecode, sample rate, bit depth, channel layout assignment. Every deliverable file is exported as BWF (Broadcast Wave Format) with the timecode anchor matching the project start, not zero-based.
Platform-Specific Deltas
Each major destination publishes its own delivery specification. The following are the categories of difference to anticipate; the specific values change between specification versions, so the platform's current delivery document is the operating reference.
Netflix. Multiple deliverable variants per title (5.1, 7.1, Atmos when applicable, near-field 2.0). Specific stem split requirements that may include separate Foley and ADR stems. Loudness target and true peak ceiling specified in the Netflix delivery document.
Disney+ and Disney-owned platforms. Atmos deliverables required for Marvel, Star Wars, and Pixar feature releases. Streaming-tier loudness target. Specific channel layout requirements per platform (Disney+, Hulu, Star+).
Apple TV+ and Apple Originals. Atmos-first delivery model for original content. Loudness specified in the Apple delivery document. File naming includes Apple-specific identifiers.
BBC and EBU broadcasters. EBU R 128 loudness compliance is mandatory. Specific archive format requirements (BWF with full metadata population). M&E required for any program with international distribution.
Theatrical (DCP). Digital Cinema Package audio standards: 5.1 minimum, 7.1 and Atmos for premium releases, IMAX 6-track for IMAX deliveries. No LUFS target; the mix is calibrated against the dub stage at a specified SPL reference.
For any specific delivery, the platform's current delivery specification document is the file to work from. Loudness targets, stem split requirements, and file naming conventions update between versions, and a project that delivered cleanly six months ago may need different settings on the next title.
Pre-Delivery QC
The QC pass before delivery catches the issues that block sign-off. Five checks.
Loudness measurement. Run the integrated loudness measurement on the final master. Confirm the value falls within the target tolerance for the destination. The measurement is also generated for every stem; stems do not have a target loudness, but their measurements are part of the QC documentation.
True peak measurement. Confirm the true peak ceiling is within the destination's specification. True peak is measured with an oversampled true-peak meter, not a sample-peak meter; the difference matters for some streaming targets.
Stem sum check. Sum the DX, MX, and FX stems together in a clean session. The sum should match the final master within numerical noise. A mismatch indicates a routing problem somewhere in the mix that has to be resolved before delivery.
M&E dialogue check. Solo the M&E stem and confirm there is no dialogue audible at any point. Spot-check three points across the timeline; if any dialogue is present, the M&E was not built clean and has to be re-printed.
File naming and metadata audit. Confirm every deliverable file follows the destination's naming convention and that the embedded BWF timecode matches the project start. A correct file with a wrong name fails ingestion at the destination as surely as a wrong file.
The full pre-delivery audit is documented in the QC report that ships with the deliverables. The report is part of the deliverable set, not a separate artifact.
How Upstream Session Discipline Shapes Delivery
A clean delivery is the output of clean upstream prep. A session that started with the facility template applied, dialogue and music and effects routed to their dedicated sub-buses, ADR on its own track group, and a properly populated AAF metadata inheritance from picture editorial is a session where the stems print clean and the M&E builds without dialogue contamination.
fPost handles the AAF prep step at the start of audio post: AI content classification at the dialogue, SFX, or music level, template routing applied, and Import Session Data preserved. The downstream effect on delivery is that the routing the deliverables process depends on is in place from the first day of dialogue editing rather than rebuilt at the last day before delivery.
Frequently Asked Questions
What is the difference between LKFS and LUFS?
LKFS (Loudness, K-weighted, relative to Full Scale) and LUFS (Loudness Units relative to Full Scale) are the same measurement under different names. ATSC A/85 uses LKFS; EBU R 128 uses LUFS. The values are numerically identical for the same audio material.
What loudness target does Netflix specify?
Netflix maintains a current sound delivery specification document with the loudness target, true peak ceiling, and stem requirements for content delivered to the platform. The document is updated periodically. Always reference the current Netflix delivery specification at the time of delivery rather than a number from a previous project.
What is included in an M&E stem?
Music, effects, Foley, and backgrounds. No dialogue (production, ADR, or walla). The M&E lets international localization teams dub the project into other languages without remixing from the original session.
Why does my final stem sum not match the master?
The most common causes are dialogue-keyed sidechain processing applied to a non-dialogue stem (which depends on the dialogue stem playing simultaneously), a stem missing a routing leg, or a master pass that has processing applied at the master output beyond what the stems carry. A correct delivery has DX + MX + FX summing to the master within numerical noise.
Does fPost handle the delivery stage?
No. fPost handles the AAF import and session organization step at the start of audio post. The delivery process happens after the mix is locked, runs through the facility's stem-print and QC pipeline, and is independent of fPost. A session that started with fPost-applied template structure has the routing in place that delivery depends on, but fPost is not a delivery tool.
fPost automates AAF import and session organization for Pro Tools, including AI content classification at the dialogue, SFX, or music level, template routing, stereo and mono correction, and Import Session Data preservation. Demo at forte-ai.com/demo.
