Chapter 44. Cue Sheets and Scoring Long Scenes#

44.1 Music changes with emotional sections, not with shots#

A forty-second conversation may cut ten times while the music stays one continuous bed. A cue should change when strategy, power or the question changes — not every time the picture cuts.

44.2 The six stages of a cue#

The entry establishes the emotion. The development adds or removes layers. The drop clears space for a key line. The rise creates imminence. The hit confirms the turn. The tail lets the result linger. Not every cue contains all six.

Figure 44-1 Dialogue, narration, music, ambience, Foley and subtitles share one timing clock. Music reduces density under dialogue and drops out before the proof appears. Ambience continues across cuts. Foley locks to action phases. Subtitles follow the final speech rather than the script text.

Use this as the working template for a spotting session. Lock dialogue and narration first, then judge room tone and action sounds. Music enters last and yields frequency and dynamic space before key language and proof. The red verticals are not positions where a hit is required; they are narrative reveals — and sometimes the most effective treatment is music leaving entirely, with only paper, breath, or a room going suddenly quiet.

cue:
  id: CUE_E001_BOARDROOM
  source: MUS_LINYUN_TRUTH_INVESTIGATIVE
  in: 00:04.200
  development: 00:14 low pulse enters
  duck: 00:24.500 Lin Yun starts recording
  drop: 00:31.800 melody removed before the authorization title appears
  hit: 00:40.100 the signature is confirmed
  out: 00:57.000 into the father's account mystery

44.3 Entries#

Music can enter after a scene has begun, letting ambience establish reality first. Place the entry on a decision, a door closing, a prop appearing, or a new strategy starting. Filling from the first frame loses hierarchy.

44.4 Drops#

Before a key reveal, reduce the melody, the low end, or all music, exposing breath and room tone. The drop must come early enough to register and must not signal the reveal too obviously.

Music stopping is itself an event. Used repeatedly it becomes a formula.

44.5 Ducking#

Music lowers automatically or manually when dialogue arrives. A practical starting point is music roughly 12–18 dB below the voice, adjusted for orchestration, voice and device. Sidechain reduction should be smooth, so the music does not pump on every word.

The point is not a fixed number. It is that every word is clear on a phone speaker while the music still carries emotion.

44.6 Phrases and cut points#

Change on musical phrases, sustains or the relationship to a pulse wherever possible. When a picture reversal lands at the wrong musical position, nudge the cue, use stems, or re-cut a few frames — do not make the dialogue chase the music.

Music serves the performance. Performers must not accelerate unnaturally to hit a beat.

44.7 Bridging scenes#

When the same question continues, music can L-cut across scenes. When time or emotion resets, end it and let the new room tone enter. A cross-episode cut point can let the hit's tail run into black, and the next episode's opening should be designed fresh rather than simply repeating it.

44.8 Scoring long narration#

Use a low-information-density bed and change stems by thought segment. Reduce melody near narration stresses, add a layer at a turn in the paragraph, leave a short tail after a landing. Do not give every sentence its own musical gesture.

44.9 Cue versions#

The episode, the IAA version, ads and compilations all cut differently. A cue keeps its master and recombinable stems and is versioned per use, with sources recorded. Never crudely trim the episode's music into a fifteen-second ad.

44.10 The cue sheet#

The full sheet records timecode, music asset, version, entry method, stems, ducking, hits, fades, rights and the narrative reason. Timecodes come from a locked or near-locked cut.

44.11 The spotting session#

Before scoring, run a spotting session: director, editor and sound lead play the near-locked picture together and mark on the timeline what the music should know, when it starts knowing it, and when it must be silent. The discussion does not begin with which piece goes here; it begins with what the audience needs to feel, understand or await at this moment.

Each mark records in point, out point, narrative function, point of view, intensity, the lines it must avoid, and whether it may cross scenes. A cue whose narrative function cannot be stated gets no music yet. Spotting also exposes scenes that lack rhythm of their own — music must not cover for an editing problem.

44.12 The grammar of entering and leaving#

A hard entry suits decisions, impact and state changes. A fade suits noticing, approaching and emotion gradually surfacing. Morphing in out of ambience suits mystery and subjective experience. A hard exit creates rupture; a sustained tail preserves aftermath; a layered exit removes the melody while the pulse continues, saying the question is not finished.

Do not use one grammar indiscriminately within an episode. The entry method itself creates expectation: if every piece of evidence arrives on a hard low-frequency entry, the music spoils the reveal before the evidence is legible.

44.13 Cues and picture-lock changes#

Music editing depends on picture timing. Adding or removing twelve frames after picture lock can displace the drop, the hit and the phrasing together. Tag every change with a cue impact: slide only, rearrange with stems, re-edit required, or recompose required. Never stretch a whole piece until artifacts or tempo problems appear.

Reserve elastic regions for high-risk scenes: sustains, drum-free beds, loopable pulses and separate tails. Then adjusting a dialogue pause need not break the thematic phrasing.

44.14 The boardroom cue in Backlit Takeover#

The first six seconds carry no music — only air conditioning, chairs and Lin Yun's footsteps. As Lin Wei announces the dismissal, a very low humiliation bed enters. Lin Yun does not argue, and the bed continues with the high end removed. Seven hundred milliseconds before she places the authorization on the table, the music leaves entirely, with only heavy paper landing. Once the title is legible, Lin Yun's three-note theme enters for the first time — as two notes. When Zhou Lan recognizes the signature there is no hit; the low pulse simply stops, leaving room for her inhale.

The four sound changes correspond to the public frame, the protagonist's strategy, the appearance of evidence and the antagonist's realization — not to each reverse angle. The cue sheet is therefore a map of power in the scene, not a music playlist.

44.15 SOP for scoring#

First, mark the scene's emotional and strategic changes. Second, choose the theme state. Third, decide where there will be no music. Fourth, design entry, development, drop, rise, hit and tail. Fifth, use stems to make room for dialogue. Sixth, fine-tune against phrasing. Seventh, check on a phone speaker. Eighth, save separate cues for distribution versions.

44.16 Fault tree#

Symptom: the music jumps at every cut. It was edited to picture rather than to emotional sections. Restore continuous cues.

Symptom: dialogue is always buried. Only the overall level was lowered while the melody still occupies the band. Pull stems, EQ and duck.

Symptom: reversals sound like a cheap sound effect. The same hit is used every time. Let proof, performance and silence land first.

Symptom: the narration is tiring to listen to. The music's information density is high and unvarying. Use beds and sectional layers.

44.17 Checklist, exercises and deliverables#

Check that cues follow emotional sections; that entries and exits have reasons; that key lines are cleared; that ducking is smooth; that phrasing supports the cut points; that long narration uses low density; and that distribution versions are produced separately.

Exercise one: write a complete cue sheet for E001. Exercise two: produce full, dialogue-bed and suspense versions of one cue. Exercise three: design only three musical changes across forty seconds of narration.

Deliverables for this chapter: the cue sheet, stem automation, the ducking map, the music edit, versioned cues, and the phone listening record.

44.18 The spotting decision tree: prove why music is needed#

Every candidate entry answers first: does the scene already have clear performance and acoustic space; is the music supplying information the audience does not have, or repeating an emotion; whose point of view does it belong to; and what understanding would be lost if it were deleted. A cue that cannot answer stays empty. Silence is not unfinished work; it is a decided sound state.

When music is needed, choose its job: establishing latent threat, maintaining momentum across cuts, bridging an elision, marking a change of knowledge, extending an aftermath, or forming a commercial cut point. One cue carries at most one or two primary jobs. A brief demanding sadness, suspense, romance, propulsion, reversal and brand feel at once usually produces generic music.

The spotting session outputs decisions and open questions rather than auditioning endlessly on the spot. The director owns narrative point of view, the music lead proposes grammar, the editor supplies timing, the mixer protects dialogue, and the producer confirms rights and cost. Anyone may suggest; responsibility for approving a cue must be singular.

44.19 Cue elasticity and structured re-edits#

Commercial microdrama frequently changes duration for campaigns and platform versions. Design cues with fixed points and elastic regions in advance: the theme's first appearance, the key drop and the closing hit are fixed; drum-free beds, sustains and loopable pulses can stretch. Edit only inside the elastic regions rather than dragging a melodic climax into the middle of a sentence.

cue_elasticity:
  cue: CUE_E001_AUTH_REVEAL
  fixed_sync:
    - {event: file_hits_table, time: 00:41.280}
    - {event: zhoulan_recognizes_signature, time: 00:47.900}
  elastic_regions:
    - {from: 00:33.000, to: 00:40.800, material: pulse_bed}
    - {from: 00:48.400, to: 00:52.000, material: sustained_tail}
  forbidden_edits:
    - time_stretch_full_mix_over_4_percent

When re-versioning, move stems and phrase modules first, re-edit if necessary, and consider slight time stretching last. Each distribution version keeps its own cue version and source master, so a temporary sixty-second cut cannot overwrite the episode's music.

44.20 Cue acceptance covers story and technique together#

Story acceptance asks whether the music reveals too early, whether it steals the performer's judgment, whether it makes every reversal identical, and whether it leaves room for the payoff. Technical acceptance asks whether entries and exits click, whether loops are audible, whether stems are phase coherent, whether ducking sounds natural, whether the function survives on a phone speaker, and whether the licence covers the distribution use.

Review in order: the version without music, then with music, then music and ambience alone. Listening with music first is easy to be persuaded by, which hides whether the scene holds on its own. Listening to music alone reveals whether cues are over-fragmented or restarting every ten seconds.

The approval record includes the picture version hash. When picture lock changes, the system marks cues needs_sync_review automatically — a cue cannot keep its passing status merely because the audio file did not change.

A note on sources#

Musical continuity comes mainly from cue structure and stems, not from regenerating background music per shot. Scoring has to collaborate with dialogue, narration and ambience on one sound timeline.