How to Prep Audio for Film/TV

Film and TV audio delivery is not flexible. The specs are not suggestions. If you send the wrong sample rate, the wrong track layout, or audio that hasn’t been loudness-normalized, it gets kicked back. Or worse — it gets accepted, goes to air, and sounds wrong on every platform.

The requirements aren’t arbitrary either. They exist because the downstream systems — broadcast encoders, streaming ingest pipelines, re-recording stages — are built around specific expectations. Give them what they expect and everything works. Don’t, and you’re the person who held up the delivery.

Here’s what you need to know before you hand anything off.

Understanding the deliverable requirements

Before you format a single file, get the spec sheet from whoever is receiving your work. This might be a broadcast network, a streaming platform, a post-production supervisor, or a finishing house. Every one of them has a technical delivery specification document. Ask for it. Read it.

There’s no single universal standard. The EBU and ATSC have published loudness standards that most broadcasters follow. Netflix, Amazon, Apple TV+, and Disney+ each have their own technical specs. What’s required for a US network drama is different from what you need for a UK broadcaster, which is different from a Netflix original.

That said, there are common patterns and conventions across most film and TV delivery contexts. Understanding those gives you a foundation to work from regardless of what the spec sheet says.

Track layout standards (dialogue, music, effects)

The D/M/E format — Dialogue, Music, Effects — is the industry-standard layout for delivering international versions of a project. It’s how broadcasters and distributors prepare content for dubbing, remixing, and localization.

The standard track layout for a stereo D/M/E is:

Tracks 1–2: Full mix stereo (L/R)
Tracks 3–4: Dialogue only (L/R or M/S)
Tracks 5–6: Music only (L/R)
Tracks 7–8: Effects only (L/R)

For a 5.1 D/M/E:

Tracks 1–6: Full 5.1 mix (L, R, C, LFE, Ls, Rs)
Tracks 7–12: Dialogue 5.1
Tracks 13–18: Music 5.1
Tracks 19–24: Effects 5.1

Some deliverables require an M&E only (music and effects combined, no dialogue) rather than separate M and E tracks. Others require a textless version or separate stem groups for music. Confirm the exact layout before you build your session.

If you’re delivering to a re-recording stage, they may also want individual stems: production dialogue, ADR, group/walla, foley, hard effects, backgrounds/ambience, and music — each as a discrete stem. This is common for big-budget content.

Within each category, keep your internal track order consistent. Dialogue tracks should be consistently ordered: lead character first, then supporting, then group. This makes the re-recording mixer’s job faster and reduces errors.

Sample rate, bit depth, and format specs

Sample rate: 48 kHz is the standard for all film and television. 44.1 kHz is for music distribution (CD, streaming). If you’re delivering audio for picture, it’s 48 kHz. No exceptions.

Bit depth: 24-bit for delivery of uncompressed audio. Some facilities accept 32-bit float for archival, but 24-bit is the broadcast and post-production standard. Do not deliver 16-bit.

File format: Broadcast WAV (BWF) is the professional standard. It embeds timecode and metadata in the file header, which is essential for sync. A plain WAV works for some deliveries, but BWF is preferred wherever sync matters — which in film and TV is everywhere.

If you’re delivering a mix as a Dolby Atmos master, the format is ADM BWF (Audio Definition Model Broadcast Wave Format). This is a multichannel BWF with Atmos object metadata embedded.

Timecode: Your files need to sync. This means timecode embedded in the BWF header or a separate EDL or AAF, depending on what the facility needs. Confirm the frame rate — 23.976, 24, 25, 29.97 DF/NDF — and make sure your session and your files match it. A timecode mismatch discovered at the re-recording stage is a very bad day.

Codec: Uncompressed BWF for delivery to a post-production facility or broadcaster. Some streaming platforms accept compressed formats (like AAC for certain Apple deliverables), but check the spec. When in doubt, deliver uncompressed.

Mono vs. stereo vs. surround: what goes where

Dialogue is mono. Almost always. A single person speaking is a point source, and it belongs on a mono track. If you’re delivering a D/M/E with dialogue on a stereo track pair, that dialogue should be the same signal on both channels, positioned center — or it should just be a mono file on a mono track, depending on the layout.

Music can be stereo (most of the time) or mono (if it was recorded or mixed that way, or if it needs to be compatible with a mono playback context).

Effects depend entirely on the content. An ambience bed is usually stereo or surround. A discrete hard effect — a gunshot, a specific door — should be mono. A moving vehicle pass might be stereo with panning automation, or handled as a Dolby Atmos object if you’re delivering in that format.

Backgrounds and room tone are typically stereo, sometimes surround.

The short version: mono for point sources, stereo for width and ambience, surround for immersive environments.

Dealing with dual mono files before handoff

Dual mono is one of those problems that hides in plain sight. A dual mono file is a two-channel interleaved file where both channels contain the same audio — or near-identical audio with slight differences from hardware or conversion artifacts.

It shows up constantly in film and TV production. Camera audio is often recorded mono and output as dual mono. Field recorders sometimes write dual mono. Sound libraries frequently deliver effects as interleaved stereo even when the source material is mono.

If you hand off a session with dual mono files sitting on stereo tracks, you’re creating problems for everyone downstream. A stereo dialogue track that’s actually dual mono will be processed through a stereo chain unnecessarily. It’ll take up double the storage. And if the two channels have any phase or level difference — even a tiny one — you’ll have imaging issues in the mix and in downmixes.

The right approach: audit every stereo file in your session before delivery. Listen to both channels separately. Check the waveforms. Anything that’s truly mono content should be a mono file on a mono track.

If you have a large session with hundreds of files — a common situation with SFX libraries and foley — doing this manually is brutal. MonoMate was built for exactly this situation. Drop your files in, it identifies every dual mono file instantly, and converts them to true mono in one pass. Worth running before any major handoff.

Loudness standards: LKFS targets for broadcast and streaming

Loudness normalization is mandatory for broadcast and most streaming delivery. If your mix isn’t hitting the right target, it either gets rejected or gets automatically attenuated or boosted by the platform’s normalization system — and automatic gain changes can damage your dynamic range and transient character.

Know your targets before you mix to them.

EBU R128 (European broadcast standard):

Integrated loudness: -23 LUFS (±1 LU tolerance)
True peak: -1 dBTP
Loudness range (LRA): typically 6–18 LU, context-dependent

ATSC A/85 (US broadcast standard):

Integrated loudness: -24 LKFS (±2 LU tolerance)
True peak: -2 dBTP

Netflix:

Integrated loudness: -27 LKFS (±2 LU tolerance)
True peak: -2 dBTP
Note: Netflix’s target is quieter than most broadcasters, which means you have more headroom for dynamics

Amazon Prime Video:

Integrated loudness: -24 LKFS
True peak: -2 dBTP

Apple TV+:

Integrated loudness: -27 LUFS (same as Netflix)
True peak: -1 dBTP

Disney+:

Integrated loudness: -24 LKFS
True peak: -2 dBTP

Theatrical (cinema):

Reference level: 85 dB SPL / -20 dBFS per channel
No integrated loudness target in the same sense — loudness is managed by calibration, not normalization

Measure your integrated loudness with a compliant meter (Nugen VisLM, Waves WLM Plus, iZotope Insight, etc.). Don’t eyeball it. And measure the final mix, not stems — the full program integrated loudness is what matters for QC.

True peak is not the same as peak. True peak accounts for inter-sample peaks that occur during D/A conversion. Use a true peak meter, and leave appropriate headroom below your platform’s TP limit.

Checklist

Run through this before delivery.

Format and technical

Sample rate: 48 kHz confirmed
Bit depth: 24-bit confirmed
File format: BWF (with timecode embedded)
Frame rate confirmed and consistent between session and files
No sample rate conversion artifacts — all source files at 48 kHz

Track layout

D/M/E layout confirmed against spec sheet
Track order matches delivery spec exactly
All tracks labeled clearly
Dialogue on mono tracks (or confirmed dual channel center per spec)
Music and effects stems correctly separated

Audio content

All dual mono files identified and converted to true mono
Stereo files verified as genuinely stereo
No clipping on any track or bus
No phase cancellation in mono downmix check

Loudness

Integrated loudness measured and hits target (±tolerance)
True peak within spec
Loudness range appropriate for content type
Full-program loudness measured (not stems only)

Sync and metadata

Timecode embedded in BWF headers
Files sync correctly against picture
ISRC or metadata embedded if required by spec

Delivery

Spec sheet reviewed against every item above
Delivery method confirmed (upload portal, hard drive, SFTP)
Naming convention matches spec
MD5 checksums generated if required