Who splits the poly WAVs, the assistant editor or the sound editor?

Either, as long as it happens before the dialogue edit and the track names are preserved. Many productions do it at offload so the whole team works with named stems.

What sample rate should the whole project use?

48 kHz for anything going to picture. Higher rates are fine for effects recording, but they get converted to 48 kHz for the session.

Do I need surround for a short film?

No. A well made stereo mix travels better through festivals and online platforms than a rushed 5.1.

8 min read

The audio post production workflow, from card offload to final mix

Most audio post problems are handover problems. Here is the whole chain, and the file format decision that matters at every step.

Audio post production is a sequence of handovers. Each stage receives files from the previous one, and almost every delay in a schedule comes from a file that arrived in the wrong shape rather than from creative disagreement.

1. Offload and verify

Sound cards are copied with checksum verification to at least two destinations. Nothing is renamed, nothing is converted. The recorder's poly WAVs and its sound report are the master documents for the rest of the process.

2. Sync and conform

Audio is married to picture by timecode where dual system was shot, or by waveform matching where it was not. This is where mismatched sample rates surface: a 48 kHz recorder against a camera running 44.1 kHz internal audio produces drift that grows across a long take.

3. Dialogue edit

The dialogue editor needs mono stems, named by microphone, at the original rate and depth. They choose between boom and lav per line, remove noises, patch with room tone and prepare ADR cues. Poly files usually get split here if the assistant editor has not already done it.

4. Sound design and effects

Library and recorded effects arrive in every format imaginable: 96 kHz masters, 44.1 kHz stock downloads, MP3 references from a director. They get converted to the session rate before they hit the timeline, not after.

5. Foley and backgrounds

Recorded to picture, usually mono, usually at the session rate. Backgrounds may be stereo or surround, and a 5.1 background sometimes needs a stereo fold down for a web deliverable later.

6. Pre dubs and the final mix

Dialogue, effects and music are balanced separately, then mixed together against loudness targets. Deliverables come out as stems plus a printed mix, and a stereo version is folded down from the surround master for online use.

The file rules that keep it moving

One sample rate for the whole project, decided before the shoot.
24 bit WAV everywhere, MP3 only for review copies.
Mono stems with meaningful names, never numbered channels.
Every conversion done deliberately, never as a silent side effect of an import.

Splitting, merging, converting and inspecting all of that can be done locally in a browser, which matters when the material is embargoed and cloud converters are not an option.