A Post-Production Workflow That Survives a Busy Month

Post-production workflows are usually documented as a sequence of steps and fail for a different reason entirely: the steps that protect a project are boring, invisible and easy to skip when a deadline is close. A workflow that works is one whose dull parts happen automatically rather than when there is time.

The order below is conventional. What matters is which parts are non-negotiable.

Ingest and Backup, Before Anything

Material comes off the cards to two separate destinations, and the copy is verified rather than assumed. Only then are the cards released for reuse.

This is the single step that cannot be recovered later. Every other mistake in post-production is repairable; lost source material is not, and a card that was reformatted after a single copy is a project with one point of failure.

The practical form is a consistent destination structure and a habit rather than a decision. Where a project runs over multiple shoot days, it happens at the end of each one, not at the end of the week.

Naming and Structure

A folder structure and a naming convention, applied the same way every time.

The reason is not tidiness. It is that a project revisited in six months is unnavigable without it, and that the most common real-world post-production error is exporting the wrong file because several had similar names.

A convention worth using includes the project, the stage and the version in every export name, and never uses “final” as a version identifier — because there is always one after it.

Synchronisation and Preparation

Separately recorded audio aligned to picture, multiple cameras aligned to each other, and everything labelled.

For interview-led projects a transcript belongs in this stage. Text is faster to search than video and a structure is much easier to argue about in words than in a timeline, which makes the transcript the most efficient tool in the whole workflow for that kind of material.

Proxies belong here too where the source material is heavy relative to the machine. The decision has to be made at ingest rather than retrofitted mid-edit, and the discipline that prevents the classic proxy error is a final check that the delivered export matches the source resolution.

Selects

Watching everything and marking what is usable. This is the stage most often skipped under pressure and the one whose absence shows up as a finished piece missing its best moment, because nobody knew it existed.

Assembly, Rough Cut, Fine Cut

The assembly puts everything usable in intended order and is deliberately unpolished. Its purpose is to reveal what the project actually has rather than what the plan said it would.

The rough cut establishes structure — watchable start to finish, too long, blunt transitions, temporary music. This is the version that goes to review, and reviewing at this stage rather than later is what prevents an expensive structural change after the finish has begun.

The fine cut is where frames are argued about, and where everything that does not advance the piece is removed.

Lock

A named point after which the structure and timing do not change.

It exists because colour, sound mixing, graphics and titling all depend on a fixed timeline. A structural change after lock means the graded shots no longer align, the mix no longer matches, and the titles have to be repositioned — three departments redoing work.

Announcing the lock explicitly, to the client as well as internally, is what makes it real. A project without a stated lock point finishes repeatedly.

Finishing

Graphics and titles, sound mixing, colour grading, in whatever order the project’s dependencies require.

Audio deserves a specific note because it is the finishing step most often abbreviated. Level consistency across a piece, dialogue that sits above the music, and delivery at the loudness standard the destination requires are what make a piece sound professional, and they are entirely separate from the edit.

Delivery

Export to the destination’s published specification: codec, container, resolution, frame rate, audio configuration, loudness, caption format.

Then check the export itself rather than the timeline. Watching the delivered file from beginning to end, including the last frame, catches the errors that a timeline preview does not: a render fault, an audio dropout, a missing graphic, a proxy-resolution export.

Multiple deliverables each get their own export and their own check. A family of twelve versions is twelve opportunities for one to be wrong.

Archive

The source material, the project file, the assets, the finished exports, and a short note stating what was delivered, in which version, on what date.

The asset part is the one people omit. A project archived without its fonts, graphics, music and source files is an archive that cannot be re-edited from, which defeats the purpose of keeping it.

What Makes a Workflow Survive Volume

One project can be completed with any workflow, including none. Several concurrent projects under deadline expose whichever step was being done by memory.

The properties that hold up: the backup happens without a decision, the naming convention is identical every time, the review stages are defined in the agreement rather than negotiated per project, and the delivery specification is obtained at the start rather than at the end.

This reference publishes no software recommendations or settings, since both differ by application and change with each version. What is public and dated instead: the delivery specifications published by the platform, broadcaster or client receiving the piece, including the loudness standard they cite — which is the document the entire finishing stage answers to.