A reference track can make a first call productive. It can also turn every demo into a failed imitation. The difference is not whether a director uses references. It is whether each reference answers a defined question.
When I receive “something like this,” I still need to know what “this” means. Is it the pace, restraint, instrumentation, emotional distance, build or room left for dialogue? Copying the surface would avoid that conversation, not solve it.
I treat references as evidence of an intention. They are not a specification for notes. That approach gives the director useful control while leaving the score room to develop its own identity.
Start with the dramatic job
The most useful note describes what the cue changes. “The audience should sense the threat before the character does” is actionable. “Use strings like the temp” names a technique before identifying the job.
I ask what the audience knows at the start, what should change by the end and what the music must not reveal. Those three answers are often more valuable than a long genre description.
Directors do not need technical language. Words such as withheld, unstable, intimate or mechanical can work well when they are tied to a specific beat. “More cinematic” is harder to use because it does not define an effect. I would rather discuss the scene than make the director translate it into orchestration terms.
Label every reference
An unlabeled playlist asks the composer to reverse-engineer the brief. The tracks may share a palette, a pulse or only a feeling the director remembers from another film.
I recommend one sentence and one timecode for every reference. “I like the restraint in the opening, not the large final build” prevents the wrong half of a track from becoming the model. “The pulse supports the edit without hitting every cut” identifies a relationship to picture rather than a sound to reproduce.
Negative references can be useful too. A track may demonstrate what feels too sentimental, too heroic or too busy under dialogue. I keep those examples separate from positive references so that the approval team does not mistake a warning for a target.
Diagnose temp love before writing around it
A temp track can gain authority through repetition. Editors, directors and producers may watch a scene against the same recording for weeks. Its accents can begin to feel inevitable even though the music was written for a different story.
I do not begin by arguing against the temp. I ask what breaks when it is muted. The answer might be momentum across a slow camera move, a lift before a reveal, or continuity through an awkward transition. An original cue can preserve that function without tracing the source track.
Watching the scene without music is equally valuable. It may show that the edit already carries the tension. It may reveal that the temp has been hiding a pacing issue. The original score should not automatically inherit every task performed by a temporary piece.
Do not brief toward an allegedly safe imitation
“How different does it need to be?” is the wrong creative and legal test. Copyright disputes do not come with a universal safe number of changed notes, seconds or instruments. The relevant rights and analysis also vary by territory and by what material is being used.
For a UK or US production, the agreement may require the composer to deliver original work and provide warranties or indemnities. Those clauses should not be treated as permission for the production to request a close copy. A warranty gives a contractual assurance, while an indemnity can allocate financial consequences if a claim arises.
I bring the discussion back to story and develop a musical logic that belongs to the project. If the production wants to use protected material it does not already control, music supervision and clearance need to enter the conversation. Composition and recording rights must be considered separately. That is a different budget, schedule and rights path.
Keep the reference set small and ranked
Twenty tracks rarely create twenty times more clarity. They often create contradictory targets. A compact, ranked group is easier to use: perhaps one reference for space, one for energy and one for emotional trajectory.
The team should mark what is essential, what is flexible and what belongs only to the current edit. If references conflict, the composer needs an approver, not a guessing game.
I also establish who can give final notes. A director, editor and producer may each have a different playlist. Their disagreement is a production decision. It should not appear late as three incompatible rounds of feedback.
Build a project palette before chasing cues
I often test a limited palette before scoring individual scenes. A few timbres, a rhythmic behaviour, a density range and a short list of exclusions can move the conversation away from the temp. The team can assess whether the film has begun to acquire its own sound.
That material can start from the script, locations or characters before picture lock. Early themes or textures may even give editorial useful material without relying on commercial temp tracks. If the editor needs flexibility, selected stems can be supplied with clear status and version labels.
The palette is not a prison. It is a shared grammar. It should remain recognisable when a quiet dialogue scene and a large sequence demand different scale.
Make the first review answer one question
An early demo should have a declared purpose. It may test the central colour, a character point of view or how the score behaves around dialogue. Asking the same file to prove final orchestration, mix, timing and theme can make notes impossible to interpret.
Useful feedback connects a timecode to an outcome. “At 01:12 the score reveals the turn too early” gives me a problem to solve. “The temp still works better” does not explain whether the issue is pace, harmony, density or familiarity.
Every review also needs a picture-version label. A temp may have shaped one cut and remain emotionally fixed after the edit changes. Without version control, the production can compare music written for different structures and call the mismatch a composition problem.
A workable first email is short
For each important scene, I want a compact note: dramatic function, point of view, trajectory, elements to preserve, elements to avoid and annotated references. Then I need the production facts: current cut, dialogue status, picture lock target, schedule, deliverables, stems and approval chain.
That is enough to begin a useful conversation. References do not remove the need for trust or judgement. They give the team concrete material for discussing an experience that can otherwise be hard to name.
The objective is not to pretend that influence disappears. It is to understand why a track works, then build a score with its own internal reasons. If you are preparing a first email or spotting conversation, I can help turn a reference playlist into a precise, production-ready brief.