Most producers ask this question once the schedule is already tight. By then the answer has largely been made for you.
Below are the four entry points, and what each one actually does to your post-production schedule. I have worked at all four, and the differences are not subtle.
Before the shoot
One case makes it unavoidable: a musical performance filmed in sync. A character singing on camera, a band playing in shot, a musical number. The track has to exist before you roll, usually recorded, so the cast can perform to playback. Other diegetic music, a radio in the background, a track heard through a wall, can be handled later.
But there is a second case, less known, and I was wrong about it for years. A score can be written before the shoot to give the film its pace. Much of the music for Minari was composed ahead of filming, and some scenes were then shaped around it. That inverts the usual order, and it produces results the standard sequence does not.
It is a directing decision, taken early, and it changes how the film is made. Not an add-on.
During the shoot
The most underused window relative to what it returns.
The point is not to write the final score. It is to produce a handful of sketches, thematic directions, rough demos, while the film is being shot. Two reasons, and the second is the one that costs money later.
The first is obvious: three months gives you options that three weeks does not. Rewriting, trying a wrong idea, discarding it.
The second is your temp track.
The temp track problem
Your editor will most likely cut to existing music. It helps find the rhythm, and there is nothing wrong with it.
Then the director watches the film fifty times with that music. The producer sees it that way. Test screenings hear it. After three months, that track and your film have fused in everyone’s head.
When the original score finally arrives, it is measured against something that had a six-month head start and that the edit has quietly grown around. It starts at a disadvantage that has nothing to do with its quality. The industry calls this temp love, and most composers have met it.
It is not inevitable. Your options:
- Rotate the temp through the edit, so nothing settles.
- Cut without music on the sequences where score will carry the scene.
- Actually license the temp if it has become the film. Sometimes the honest answer, often outside the budget.
- Bring in a music editor, for whom handling this is part of a wider job covering music editing, sync and tracking.
- Bring the composer in early enough that the temp is his.
The last one is my preference, and not only because it suits me. Demos produced during the shoot become the edit’s temp, the team attaches to them, and the film never goes through the switch. It is composing time, so it is a budget line, but it is spent up front instead of on repair.
At first assembly
The best compromise, and the one I would pick if I could only have one.
The film exists and its structure is readable, but nothing is locked. The composer works from something real rather than a script, proposes directions before the cut crystallises, and the editor can still give a sequence the eight seconds a phrase needs to land.
That back-and-forth between cut and score is one of the few places a film genuinely gains something. It narrows as the cut hardens, and is gone once picture is locked.
The cost: things will move, and you will rewrite. It is a modest price.
After picture lock
Locked cut, nothing moves, the composer writes to something final. In my experience this is the most common configuration, and the most efficient on paper.
Two consequences worth knowing before you choose it.
First, the time left is the time left. If delivery is fixed and the edit ran three weeks over, those weeks come out of post, and music absorbs its share alongside sound, grade and VFX. Music is also the one department the final mix waits on, which limits how far its delivery can slip.
Second, the film is what it is. A sequence that would have gained from a few more seconds will not get them.
This entry point is perfectly workable, on one condition: that the schedule genuinely reserves the time, rather than treating music as the variable that flexes.
What to do with this
Three decisions worth taking early:
Budget a demo phase during the shoot. A modest line against what it prevents, and it solves the temp track problem before it exists.
Decide your temp track deliberately. Cutting to a famous cue creates a comparison your score will have to survive. Not forbidden, but it should be a choice rather than a default.
Ring-fence the music time in the post schedule. When it starts only at picture lock, it has little slack, and the mix is waiting behind it.
I work both ways, from early demos during a shoot to a tight post-production window. If you are planning a project and want to talk it through, get in touch.