An approved cue is not yet a delivered cue. The composer still has to turn the creative decision into files that editorial and the re-recording mixer can place, identify and control without opening the original music session.
I like to define that handoff before the first polished mockup. Waiting until the final week creates avoidable questions. Does the mix need stems? Should every file start at the same timecode? Is a no-melody alternate expected? Who decides which revision is current? None of this is glamorous. Neither is discovering the wrong “final” in the dub.
Ask who will receive the score
The director may approve the music, but a different person often has to use it. On an independent feature, that may be the picture editor before the files reach a sound team. On a larger post-production, a music editor, supervising sound editor or re-recording mixer may specify the delivery.
I ask for a technical contact and the date when dependable files are required. A short delivery specification can then cover file type, sample rate, bit depth, start point, stem layout, naming and transfer. The production should provide its established requirements where they exist. The composer should not invent a house standard on its behalf.
This conversation also reveals the real deadline. “Music approved” and “music ready for the final mix” are separate milestones. Recording, editing, music mixing, exports and checks all sit between them.
Keep the approved full mix
The approved full mix, in the agreed channel format, is the clearest record of the musical balance. It should remain part of the delivery even when the production requests stems. Editorial can use it as a reference, and the dub can compare the recombined material against it.
A full mix has limits. If dialogue loses clarity under one busy musical layer, the mixer cannot lower that layer alone. Turning down the entire cue may remove the emotional support along with the conflict. Stems create controlled options without requiring a new composition.
The right layout follows the score. Strings, brass, percussion, synths, soloists and rhythmic elements are common groups, but they are not a mandatory template. A sparse acoustic cue may need very little separation. A dense hybrid cue may benefit from more. I want each split to answer a plausible mix decision.
Stems should rebuild the master
My basic quality-control test is simple: place all stems at unity and confirm that they reproduce the approved master closely. Missing shared effects, duplicated processing, different limiting or an accidental routing change can make the sum behave differently.
That difference matters. If the stems rebuild a louder, drier or less coherent mix, the re-recording mixer is not adjusting the approved score. They are starting a second music mix under schedule pressure.
More stems do not automatically solve the problem. A folder of tiny instrument groups can move every musical balance decision into the dub. That adds complexity without necessarily adding useful control. I would rather isolate a featured motif, a potentially competing percussion layer or a source-like element for a reason than export every track because the session contains it.
Give every file a dependable sync reference
The file has to land on the picture as reliably as it sounds. I normally deliver all elements of a cue with a common start, including silence before a later musical entrance. Editorial can place the files at one known timecode rather than rebuilding offsets by ear.
The exact convention depends on the post workflow. A production may use a reel-based approach, a continuous timeline or another established reference. Consistency is the important part. The start point, frame rate and picture version should be documented and should agree with the exports.
I include a technical cue list showing the cue identifier, title, revision, duration, start reference and associated files. This is not the same document as a performing-rights cue sheet. Its job is operational: someone who has never seen my DAW session should understand what is in the delivery.
Separate alternates from stems
A stem is one component of the approved mix. An alternate changes the creative version. A no-drums pass, button ending, shorter opening or reduced melody may be valuable, but the production should identify why it needs that option and whether it requires picture approval.
Keeping those categories distinct prevents a familiar mess. If every stem, experiment and requested alternate sits in one folder, editorial has many files but no clear authority. I organise the hierarchy so the approved full mix, its stems and any approved alternates remain obvious.
File names should carry that logic. Cue number, concise title, version and revision are usually more useful than a string of adjectives. I avoid “final,” “final final” and “use this one.” Those names describe somebody’s optimism, not the state of post-production.
Picture lock still needs a version number
Picture lock is a production state, not a force field. A late trim can shift a hit point, expose a tail or change the pace of an entire cue. Before final exports, I confirm the picture reference used for each approval.
If the cut moves, the team needs to decide whether to edit, conform or rewrite the music. Editorial can make practical adjustments, especially during temp work, but a silent recut of approved score can alter phrasing and structure. I prefer to receive the changed picture, agree on the fix and make sure the delivered version matches what the director heard.
This becomes particularly important when the composer is working remotely and the post team is moving quickly. A precise version trail replaces the informal knowledge that would otherwise live in a room.
Plan for downstream uses without exporting everything
The feature mix may not be the last time the production needs the score. Festivals, international versions, trailers, teasers and recut credits can create later requests. A structured archive of approved masters, stems and sync information makes those requests manageable.
That does not mean producing every imaginable cutdown in advance. I ask which uses are actually planned. Deliverables should support the project, not become a warehouse of hypothetical files.
Audio exports are also a safer long-term record than relying only on the original DAW session. Software versions, sample libraries and licences change. A session may no longer open exactly as it did. Clearly labelled audio preserves the approved work even when the production environment has moved on.
Put one practical question in the hiring process
Before hiring a composer, ask: “What exactly will our post team receive, and how do you quality-check it?” A useful answer should connect creative approval to technical delivery. It should mention the master, the planned stems, sync, revisions and the recipient’s specification without pretending one template fits every film.
Strong deliverables cannot rescue the wrong score. But a strong score delivered ambiguously can lose impact at the final mix. Agreeing the handoff early protects the music when dialogue, effects and picture finally meet.
If you are building a post schedule, I can help define a lean deliverables list for the film before the first cue is written.