skills/ai-film-pipeline-master/references/phase-03-shooting-script/sound-as-visual.md

# Sound as a Visual Tool

26 cards. These decide a shot: how it moves, how it cuts, and how it sits in time.

Every card carries what it is, what it does to the audience, where it works best, what to avoid, and the prompt written out in full.

> **On the `editor:` and `mix:` lines in this file, and the numbers in them.** These cards describe
> **decisions about the cut** — where a sound starts relative to a picture, which side of a cut it
> lands on, what the silence is doing — and that is why they live in the shot-planning phase rather
> than the audio one. Their example lines are written in an editor's voice because that is who
> executes them.
>
> **The figures in those examples are illustrations, not settings.** Phase 4 states the boundary for
> the whole package — *the phase specifies, the edit decides* — and it holds here: a frame count, a
> decibel figure or a millisecond value in one of these examples shows the *shape* of the move at a
> plausible size, and the real value is set on a timeline against the actual material. Take the
> decision from the card and the number from the edit. Where a card seems to hand you a measured
> value to obey, it is showing you what the move looks like when someone has already made it.

---

### Which Side of the Cut

**Also called:** the sound-as-visual selector, side of the cut, who may hear the move  
**What it is:** The card at the head of the `sound.` shelf. It does not describe a sound. It asks the three questions this shelf's own pair graph keeps asking — **does the sound cross a cut or stay inside the shot, may the audience catch the move being made, and what is the sound made of** — and hands back a named shortlist for each answer. The axes were read off the `Yields to:` conditions that touch this shelf, in both directions, and off its `Avoid when:` boundaries; the conditions that repeated became the questions. Every card cited here is cited by id.  
**Effect on the audience:** None. This card never reaches the track. It exists so that the sound decision written into a shot was chosen against the cut, the register of the whole piece and the material actually to hand, rather than reached for because it is the move everybody already knows.  
**Used for and where it works best:** Answer the three questions **in order and out loud, one sentence each, before you open the shelf**. Question one is a fork, not a filter: the two halves of this shelf never send a reader to each other, so answering it puts you on one half and closes the other. Question two is the one that decides whether a move plays as craft or as a cheap effect, and it is answered by the piece, not by the moment — settle the register of the whole piece before you answer it for one shot. Question three rarely narrows the list; it names the door you will leave by if the card is not enough — a music move ends in the score, a world move ends in the mix, a voice move ends back in the writing.  
**Best in:** formats: all | genres: all  
**Avoid when:** **You are cutting a reel, a trailer or an advert and following this shelf's own pairs.** Every edge that lands on this shelf from a louder card is a retreat — *reads cheap*, *soap-worn*, *breaks authenticity*, *must be inaudible* — and not one condition anywhere in the package sends a reader toward a more audible move for the audience's sake. On a piece whose grammar is the audible device, the graph under this map points the wrong way every time, and question two's third value has to be taken on the format's word, not the graph's. Also avoid it for music you are composing: all three music cards on this shelf leave it — two for the score, one for the optional rights card — and a composed stop, pulse or switch is decided in the audio phase, not here. And **distrust question two whenever the piece changes register inside itself**: the same whoosh is grammar in the reel cut and a cheap library effect in the documentary cut of the same material, and the axis has one answer per piece, not per shot. The frame counts and seconds in these cards are illustrations; the map returns a move, never a length. And **do not let this map settle the silence before a cut.** `sound.silence_before` drops the audio *"down to the room tone, never to digital zero"*, while `sound.room_tone` names an authored vacuum under `sound.silence_before` as its one exception, where *"its finite digital zero is the effect"*. The two cards disagree about the same beat; the map routes to both, and the scene decides whether its silence is a believable room or an editorial void.  
**Prompt:** `[Selector — produces no prompt. It returns card ids. The chosen card carries the prompt.]`  

**The axes**

**Axis 1 — Does the sound cross a cut, or stay inside the shot?**
Asked first because this shelf answers it with its structure before its words. The shelf's six internal edges all stay on one side of this question — not one of them joins a join card to an inside card — and the two halves leave by different doors: the join half for `cut.`, `pacing.`, `rights.`, `sub.` and two `foley.` cards; the inside half for `score.`, `soundscape.`, `mix.` and `dial.`. Not one destination family is shared. The conditions say it in words as well: *"New sound violent and surprising; leading burns the surprise"* (`sound.j_cut`), *"Fast comedy, where trailing audio drags the rhythm"* (`sound.l_cut`), *"After an emotional scene, which needs silence"* (`sound.prelap`), *"The vacuum precedes an impact, not a cut"* (`sound.silence_before`), *"No single identifiable sound spans both scenes"* (`sound.bridge`), and from outside the shelf, *"When the silence must arrive before the cut"* (`cut.to_silence`).

**Axis 2 — May the audience catch the move being made?**
The strongest repeat in the `Yields to:` field — eight of the conditions that touch this shelf — and every one of them is a statement about the piece, not about a feeling: *"Drama and cinematic documentary, where a whoosh reads cheap"* (`sound.whoosh`), *"Serious drama, where an edit hit reads cheap"* (`sound.sting`), *"Serious drama, where lifted texture sounds artificial"* (`sound.asmr`), *"Documentary, where a muffled POV reads as manipulation"* (`sound.subjective`), *"Quiet scenes, where any transition sound breaks the mood"* (`sound.audio_whip`), *"Audiences now predict the plain trailer riser"* (`sound.riser`), *"Soap-worn repetition; carry the line plainly instead"* (`sound.echo_last_word`), and from outside, *"When the join must be inaudible, one real sound"* (`cut.sound_match`). The `Avoid when:` lines ration the middle value in so many words — *more than twice in a film* (`sound.silence_before`), *once in an entire piece* (`sound.echo_last_word`), *twice in the same scene* (`sound.music_stop`). And the internal edges obey it: every one that crosses a tier crosses the way its condition says — `sound.echo_last_word` down to `sound.postlap` on *soap-worn*, `sound.postlap` up to `sound.echo_last_word` on *obsession*, `sound.whoosh` down to `sound.bridge` on *reads cheap*.

**Axis 3 — What is the sound made of?**
Nine conditions turn on the material, and they are exactly the ones that send a reader off the shelf: *"The drop is composed into the cue before a climax"* (`sound.music_stop`), *"A played instrument taken over by the orchestra"* (`sound.diegetic_switch`), *"The accelerating pulse is scored percussion"* (`sound.heartbeat`), *"The bed should be pitched underscore, not felt sub"* (`sound.drone_bed`), *"The location's own tone is unusable"* (`sound.room_tone`), *"A real named space, matched by impulse response"* (`sound.perspective`), *"Narration that would only describe the picture"* (`sound.vo_bridge`), *"The piece ships subtitled or dubbed"* (`sound.overlap_dialogue`), and from outside, *"Distance and shot size, not a room, set the space"* (`mix.convolution_ir_matching`). Asked last because it seldom narrows the shortlist; it tells you which department finishes the job.

**The map**

**Axis 1 — which side of the cut**

| Answer | The shortlist, by id |
| :--- | :--- |
| **Across a cut — the join is the event** (twelve) | `sound.j_cut` · `sound.l_cut` · `sound.bridge` · `sound.prelap` · `sound.postlap` · `sound.echo_last_word` · `sound.vo_bridge` · `sound.silence_before` · `sound.needle_drop` · `sound.riser` · `sound.whoosh` · `sound.audio_whip` |
| **Inside the shot — the scene is the event** (twelve) | `sound.music_stop` · `sound.sting` · `sound.diegetic_switch` · `sound.drone_bed` · `sound.heartbeat` · `sound.offscreen` · `sound.ambience_swap` · `sound.subjective` · `sound.room_tone` · `sound.perspective` · `sound.overlap_dialogue` · `sound.asmr` |
| **Nowhere — the viewer will not hear it at all** (one) | `sound.sound_off` |

**Axis 2 — may the audience catch it**

| Answer | The shortlist, by id |
| :--- | :--- |
| **Never — if it is noticed, it has failed** (twelve) | `sound.j_cut` · `sound.l_cut` · `sound.bridge` · `sound.prelap` · `sound.postlap` · `sound.vo_bridge` · `sound.overlap_dialogue` · `sound.offscreen` · `sound.room_tone` · `sound.perspective` · `sound.drone_bed` · `sound.heartbeat` |
| **Once, and the moment has to pay for it** (seven) | `sound.silence_before` · `sound.music_stop` · `sound.needle_drop` · `sound.diegetic_switch` · `sound.echo_last_word` · `sound.ambience_swap` · `sound.subjective` |
| **Always — the move belongs to the format, not the moment** (six) | `sound.sting` · `sound.riser` · `sound.whoosh` · `sound.audio_whip` · `sound.asmr` · `sound.sound_off` |

**Axis 3 — what it is made of**

| Answer | The shortlist, by id |
| :--- | :--- |
| **A voice** (five) | `sound.prelap` · `sound.postlap` · `sound.echo_last_word` · `sound.vo_bridge` · `sound.overlap_dialogue` |
| **The music** (three) | `sound.needle_drop` · `sound.music_stop` · `sound.diegetic_switch` |
| **The world's own sound** (nine) | `sound.j_cut` · `sound.l_cut` · `sound.bridge` · `sound.offscreen` · `sound.ambience_swap` · `sound.subjective` · `sound.room_tone` · `sound.perspective` · `sound.asmr` |
| **A sound nothing in the world makes** (six) | `sound.sting` · `sound.riser` · `sound.whoosh` · `sound.audio_whip` · `sound.drone_bed` · `sound.heartbeat` |
| **The absence of sound** (two) | `sound.silence_before` · `sound.sound_off` |

*A card appears once per axis. The answer is the intersection of the three shortlists, and it is normally one to three cards wide.*

**The default, and the one against it**

**The default is `sound.j_cut` at the join and `sound.room_tone` under the shot — and it is argued, because the measure names a door, not a fallback.** The most-reached card on the shelf by edges is `sound.bridge`, sent to from two families — by `sound.whoosh` (*"where a whoosh reads cheap"*) and by `cut.sound_match` (*"When the join must be inaudible"*) — and both fire on one condition: the join must not be heard. That is the exit from the loud tier, not the place the shelf falls back to. Counting distinct conditions instead, `sound.subjective` leads on two — *"Detachment played from inside the character's head"* from `sound.ambience_swap` and *"Period documentary, where a glitch edit breaks authenticity"* from `foley.stutter_audio_impact` — but the second lands a documentary reader on a card whose own `Avoid when:` says it *reads as manipulation* in documentary, and which yields straight on to `sound.ambience_swap` for that reason. One valid condition ties with every other card here. So the shelf has no measured default, and the two named above are chosen on the cards' own words: `sound.j_cut` calls itself *the default in most professional scene transitions* and leaves the shelf on one narrow condition only, a violent surprise that goes to `cut.smash`; `sound.room_tone` is the only card here whose `Best in:` reads all and all, whose body says *every scene without exception*, and which leaves only when the location's own tone is unusable.

**The one against it is `sound.offscreen`.** Nothing routes to it and it routes nowhere — it is the only card on this shelf outside the pair graph altogether. The default makes a promise and keeps it inside a second or two: the sound arrives first and the picture it belongs to follows. `sound.offscreen` makes the same promise and never keeps it — the knock is heard, the camera holds on the listening face, and the door is never shown. It is the one move on the shelf that makes the world larger than the frame instead of joining two frames together. Its cost is stated in its own `Avoid when:` and it is the honest one: *if you never reveal the source in a scene that demands an answer the viewer feels cheated*, so it has to be chosen as deliberate mystery or as delayed information, and the card makes you say which. Its near cousin against the grain is `sound.ambience_swap`, the only card that changes the place and refuses to change the picture.

**The unasked**

**Sixteen of the twenty-five cards this map reaches are the target of no `Yields to:` field anywhere in the package.** They can be chosen but nothing ever sends you to them. By id:

`sound.j_cut` · `sound.l_cut` · `sound.music_stop` · `sound.needle_drop` · `sound.sting` · `sound.whoosh` · `sound.audio_whip` · `sound.drone_bed` · `sound.heartbeat` · `sound.diegetic_switch` · `sound.vo_bridge` · `sound.offscreen` · `sound.room_tone` · `sound.overlap_dialogue` · `sound.sound_off` · `sound.asmr`

**And the shape of that list is the finding, not its length.** Three things are on it.

**First: every music card is on it, and every music card leaves.** `sound.needle_drop`, `sound.music_stop` and `sound.diegetic_switch` are reached by nothing, and each yields off the shelf — to the optional rights card, to the score's own vacuum, to the score's own transformation. Music on this shelf is never chosen between; it is only exited. The shelf decides *where* music enters and stops relative to a cut, and hands *what* it is to somebody else.

**Second, and it is the larger one: the pair field on this shelf can say *quieter* and it can say *deeper*, and it has never once said *louder*.** Of the six cards whose move belongs to the format, five are reached by nothing, and the sixth, `sound.riser`, is reached only by its own twin in the audio phase, `foley.shepard_tone_riser` — the two send readers to each other and nothing else sends a reader to either. Of the six sounds nothing in the world makes, the same. The two edges on this shelf that climb a tier — `sound.postlap` to `sound.echo_last_word`, `sound.ambience_swap` to `sound.subjective` — both climb into a character's head, never out toward the audience. **The readers who wrote these fields were each guarding a serious piece against a cheap one**, and the field has no way to say the opposite: *this reel dies in the first second unless the move is heard*. The card this shelf wants is the one that says when the piece needs the device heard and what that costs — the scroll-stop, the logo hit, the drop on the beat — and until it exists, the six cards that would answer it are reached only by being retreated from.

**Third: the declared default is on it.** `sound.j_cut` names itself the default and nothing on or off the shelf ever sends a reader to it; `sound.room_tone`, the other half of the default, is reached by nothing either. The two moves every piece makes are the two nobody routes to — which is what a floor looks like in a pair graph, and why the default above is argued rather than counted.

`sound.which_side_of_the_cut`

---

### J-Cut

**Also called:** sound leads picture, split edit  
**What it is:** The sound of the new scene arrives before its picture: you hear the market noise or the next character's voice while you are still watching the old scene.  
**Effect on the audience:** It pulls the viewer forward and creates anticipation. The mind starts building the new place before you show it, so when the cut comes it feels inevitable.  
**Used for and where it works best:** It is the default in most professional scene transitions: start the new audio half a second to two seconds before the picture. In drama and series it removes the dead time between scenes. In documentary, let the location sound come up under the last line of the interview.  
**Best in:** formats: Feature Film, Drama Series Episode, Limited Series, Documentary Feature, Documentary Series, Heritage / Historical Documentary, Short Film, Video Essay, Commercial / TVC, Trailer / Teaser | genres: Drama, Thriller / Suspense, Crime / Procedural, Mystery / Whodunit, Historical / Biopic, Heritage (civilisation-focused), War, Political Thriller, Romance  
**Avoid when:** If the new sound is violent and surprising, leading with it burns the surprise. In that case use a smash cut instead of a J-cut.  
**Prompt:** `editor: bring the next scene's audio up one second before the picture cut, keeping the outgoing image on screen through it`  
**Yields to:** `cut.smash` — New sound violent and surprising; leading burns the surprise.  
`sound.j_cut`

### L-Cut

**Also called:** sound lags picture, trailing audio  
**What it is:** The sound of the old scene continues over the picture of the new one: you see the new place while still hearing the shouting, the music or the rain from before.  
**Effect on the audience:** It makes the old scene follow the viewer. Residue, regret, longing, or a shock the character cannot shake off.  
**Used for and where it works best:** After a parting, after a death, after a battle, and at the end of a memory. The rule is that the longer the old sound lingers over the new picture, the heavier what was lost becomes. In drama, war and tragedy it is the most important transition after the J-cut.  
**Best in:** formats: Feature Film, Drama Series Episode, Limited Series, Short Film, Documentary Feature, Heritage / Historical Documentary, Documentary Series, Music Video, Anime (Japanese 2D) | genres: Drama, Tragedy, War, Melodrama, Romance, Historical / Biopic, Heritage (civilisation-focused), Psychological / Arthouse, Crime / Procedural  
**Avoid when:** If the old sound runs on too long the viewer gets confused and thinks the scene has not changed. In fast comedy it drags the rhythm down.  
**Prompt:** `editor: hold the outgoing scene's audio for one to three seconds over the new picture, then fade it under the new ambience`  
**Yields to:** `sound.prelap` — Fast comedy, where trailing audio drags the rhythm; and `sound.j_cut` — the new scene must pull forward, not the old one linger.  
`sound.l_cut`

### Sound Bridge

**Also called:** audio bridge, linking sound  
**What it is:** A single sound, music, a scream, a bell, a call to prayer, wind, runs across the cut and covers both scenes, linking two places or two times.  
**Effect on the audience:** The ear hears continuity so the mind assumes a relationship. It links two things with no visual connection and makes the contrast between them read louder.  
**Used for and where it works best:** Chapter transitions, passing time, and tying two story lines together. In heritage work the wind over the ruins carries on over the city when it still stood. In documentary the sound of the craftsman working continues over shots of the finished object.  
**Best in:** formats: Feature Film, Documentary Feature, Heritage / Historical Documentary, Documentary Series, Drama Series Episode, Video Essay, Music Video, Short Film, Trailer / Teaser, Limited Series | genres: Heritage (civilisation-focused), Historical / Biopic, Drama, War, Mythology / Ancient Epic, Tragedy, Political Thriller, Faith / Biblical Epic, True Crime  
**Avoid when:** If the sound is generic with no identity, library music for example, it links nothing and becomes filler. Choose a sound that means something in the story.  
**Prompt:** `editor: run one identifiable sound at constant level straight across the cut, and change only the picture underneath it`  
**Yields to:** `cut.sound_match` — No single identifiable sound spans both scenes.  
`sound.bridge`

### Pre-Lap

**Also called:** dialogue pre-lap, advance line  
**What it is:** A line of dialogue from the next scene is heard over the final frames of the current one, so the character speaks before we see her.  
**Effect on the audience:** It removes dead time completely and interlocks the scenes. The viewer feels the show never pauses to breathe between scenes.  
**Used for and where it works best:** Fast modern television and dialogue driven drama, and it became widespread in Aaron Sorkin's series. It is strongest when the incoming line answers or contradicts the question of the outgoing scene. In long form advertising the next speaker's voice pulls you into the following block.  
**Best in:** formats: Drama Series Episode, Limited Series, Feature Film, Soap Opera / Telenovela, Vertical Micro-Drama, Short Film, Corporate / Training, Documentary Series, Commercial / TVC | genres: Drama, Legal Drama, Political Thriller, Medical Drama, Comedy, Crime / Procedural, Rom-Com, Mystery / Whodunit, Melodrama  
**Avoid when:** Between every two scenes it leaves the show no room to breathe and exhausts the viewer. Emotional scenes need silence after them, not a line pushing in over the top.  
**Prompt:** `editor: lay the first line of the next scene over the last twenty four frames of the current shot, then cut on the second word`  
**Yields to:** `pacing.pause_act_break` — After an emotional scene, which needs silence.  
`sound.prelap`

### Post-Lap

**Also called:** trailing dialogue, voice carry over  
**What it is:** A line from the scene that just ended continues over the opening shots of the new one, so we see one place and hear speech from another.  
**Effect on the audience:** It makes the words follow the character. The viewer understands the line never left her, it is still turning in her head.  
**Used for and where it works best:** After a threat, a promise or an accusation: we see her at home and still hear his voice from the previous scene. In psychological drama and crime it functions as memory. In documentary, the witness's line continues over a shot of the place he described.  
**Best in:** formats: Feature Film, Drama Series Episode, Limited Series, Documentary Feature, Heritage / Historical Documentary, Short Film, Documentary Series, Soap Opera / Telenovela, Anime (Japanese 2D) | genres: Psychological / Arthouse, Drama, Crime / Procedural, Thriller / Suspense, Tragedy, Mystery / Whodunit, True Crime, Melodrama, War  
**Avoid when:** If the line runs too long over the new scene the focus breaks and we lose our place. Keep it to one short clear sentence.  
**Prompt:** `editor: carry the last line of the previous scene over the first two seconds of the new picture, then drop into the new ambience`  
**Yields to:** `sound.echo_last_word` — The line should read as obsession, not continuation.  
`sound.postlap`

### The Silence Before the Cut

**Also called:** pre-cut silence, audio drop before the cut  
**What it is:** All the audio drops shortly before the cut — down to the room tone, never to digital zero, which `sound.room_tone` says reads as a broken file — so the shot finishes in silence and only then does the cut happen. How long the silence runs is set on the timeline in the edit, against the shot it is emptying.  
**Effect on the audience:** The strongest free tension tool in cinema. The ear notices the vacuum instantly and senses that something is about to happen.  
**Used for and where it works best:** Before the explosion, before the news, before the death, before the reveal. In horror and suspense it is the equation for the beat before the hit. In trailers, a full silence before the final shot pulls the audience forward in their seats. The difference from a hard cut to silence: here the silence arrives before the cut, there it arrives with it on the same frame. Its two neighbours in Phase 4 are separate moves with separate lengths: `foley.acoustic_vacuum_drop` puts the vacuum before an **impact** rather than a cut, and `score.pre_climax_score_vacuum` removes only the score and leaves the world running.  
**Best in:** formats: Trailer / Teaser, Feature Film, Drama Series Episode, Short Film, Anime (Japanese 2D), Vertical Short-Form (Reels, Shorts, TikTok), Music Video, Game Cinematic / Gameplay, AI-Generated Clip | genres: Horror, Thriller / Suspense, War, Psychological Horror, Action, Drama, Supernatural Horror, Disaster, Slasher  
**Avoid when:** Use it more than twice in a film and the viewer learns it and predicts the hit. And already quiet scenes gain nothing from it; silence needs noise before it to be felt.  
**Prompt:** `editor: drop every audio track except the room tone to zero before the cut, leave the picture running silent, then cut — set the length of the silence against the shot`  
**Yields to:** `foley.acoustic_vacuum_drop` — The vacuum precedes an impact, not a cut.  
`sound.silence_before`

### The Music Stops

**Also called:** needle off, music cut out  
**What it is:** The music that was playing stops dead on one frame, and the real room sound is left on its own.  
**Effect on the audience:** It changes the mood completely in an instant. Everything that felt fun or heroic becomes real and heavy, and the viewer is dragged back to reality.  
**Used for and where it works best:** The moment a scene turns: the party ends with news, the hero realises he lost, the joke becomes serious. In dark comedy and drama it is the beat that flips the film. In music video, stopping the track before the final chorus doubles the impact of its return.  
**Best in:** formats: Feature Film, Drama Series Episode, Music Video, Short Film, Trailer / Teaser, Vertical Short-Form (Reels, Shorts, TikTok), Commercial / TVC, Anime (Japanese 2D), Limited Series | genres: Dark Comedy / Satire, Drama, Thriller / Suspense, Tragedy, Crime / Procedural, War, Musical, Horror, Comedy  
**Avoid when:** If the music was not loud and present beforehand the stop is not felt. Twice in the same scene it usually loses its meaning, because the first stop already spent the silence; the exception is when the second stop answers the first and the scene is built on the pair.  
**Prompt:** `editor: cut the music to zero on one frame with no fade, leaving only the real location sound running underneath`  
**Yields to:** `score.pre_climax_score_vacuum` — The drop is composed into the cue before a climax.  
`sound.music_stop`

### The Needle Drop

**Also called:** song drop, music cue entrance, licensed track drop, song entry  
**What it is:** A recognisable song enters on one precise moment, usually on a strong shot or at the start of a montage, and takes command of the scene.  
**Effect on the audience:** The song brings its entire history and the audience's feelings about it, so you get ready made emotion in a single second.  
**Used for and where it works best:** A character's entrance, the start of a montage, the end of the film, and trailers. The rule is to let the song enter on a movement or on a cut, never in the middle of a shot, and in a trailer to bring it in after a full second of silence or after a line, on the first big image. In advertising and reels a recognisable track is what stops the scroll, and contradiction beats agreement: a happy song over a brutal scene.  
**Best in:** formats: Feature Film, Trailer / Teaser, Music Video, Vertical Short-Form (Reels, Shorts, TikTok), Commercial / TVC, Drama Series Episode, Short Film, Meme / Comedy Edit, Sports Broadcast / Sports Film, Title Sequence / Opening Credits, Game Cinematic / Gameplay, Documentary Series | genres: Crime / Procedural, Gangster, Comedy, Dark Comedy / Satire, Action, Romance, Coming of Age, Sports Drama, Heist, Drama, Thriller / Suspense, War  
**Avoid when:** If the lyrics describe the scene literally it becomes naive, if they fight the narration both are lost, and a famous song used for no reason makes the piece feel derivative.  
**Prompt:** `editor: drop the song in on the frame of the cut, not mid shot, and let the first beat land on the strongest image`  
**Example:** Goodfellas, the Layla piano coda entering over the discovery of the bodies.  
**Yields to:** `rights.cleared_before_locked` — Only where the brief itself raises the track's rights; optional craft, never a step.  
`sound.needle_drop`

### The Sting / The Hit

**Also called:** stinger, musical hit, jump scare hit  
**What it is:** A short sharp sound, a string stab, a drum hit, a metal clang, a breath, placed on the exact frame of the shocking image at a level above everything before it.  
**Effect on the audience:** It doubles the shock and makes the body jump. Most jump scares in horror are sound, not picture.  
**Used for and where it works best:** Horror on the appearance of the thing, trailers on the logo, advertising on the entrance of the product. In documentary and true crime, a soft hit on the reveal of evidence. The rule is to pull everything down half a second before it so the hit feels louder rather than just measuring louder.  
**Best in:** formats: Trailer / Teaser, Feature Film, Drama Series Episode, Commercial / TVC, Vertical Short-Form (Reels, Shorts, TikTok), Game Cinematic / Gameplay, Anime (Japanese 2D), Documentary Series, Music Video | genres: Horror, Thriller / Suspense, Slasher, Supernatural Horror, True Crime, Action, Psychological Horror, Mystery / Whodunit  
**Avoid when:** Too many hits and the body habituates and stops jumping, and the sting turns comic. In serious drama it reads as cheap manipulation.  
**Prompt:** `editor: duck everything half a second before, then land one short sharp musical hit on the exact frame of the reveal`  
**Example:** Psycho, the shrieking violins landing with the shower stabs.  
**Yields to:** `score.revelation_stinger_impact` — Serious drama, where an edit hit reads cheap.  
`sound.sting`

### The Riser

**Also called:** build up, whoosh riser, tension build  
**What it is:** A sound that rises in pitch and level over two to ten seconds until it peaks, then stops on the cut or on the hit.  
**Effect on the audience:** It builds involuntary physical tension. The ear anticipates the peak and braces, and whatever follows feels bigger.  
**Used for and where it works best:** Trailers, reels and advertising before the reveal, and in film before an explosion or an arrival. The rule is that the rise must end exactly on the intended frame; one second over and it is ruined. In documentary use it sparingly before an important fact.  
**Best in:** formats: Trailer / Teaser, Vertical Short-Form (Reels, Shorts, TikTok), Commercial / TVC, Feature Film, Game Cinematic / Gameplay, Short Film, Drama Series Episode, Music Video, Documentary Series, Product Video / Motion Design | genres: Action, Thriller / Suspense, Horror, Sci-Fi, Superhero, Disaster, War, True Crime, Mystery / Whodunit  
**Avoid when:** Trailers have used it so heavily that audiences now know exactly what is coming. And if there is no real peak after it, the viewer feels cheated.  
**Prompt:** `editor: build a rising tone that peaks exactly on the cut frame, then cut it to zero on that same frame`  
**Yields to:** `foley.shepard_tone_riser` — Audiences now predict the plain trailer riser.  
`sound.riser`

### The Whoosh

**Also called:** swoosh, transition whoosh  
**What it is:** A rush of air or fast movement laid over the transition, starting slightly before it and ending slightly after, so it feels like something swept past.  
**Effect on the audience:** It makes the transition feel deliberate and designed, and links the two shots with an audible movement even when the images share none.  
**Used for and where it works best:** Reels, advertising, corporate video and title cards, and with every motion transition, whip pans, object wipes, pushes. The rule is to start the sound five frames before the transition so it leads the eye instead of chasing it.  
**Best in:** formats: Vertical Short-Form (Reels, Shorts, TikTok), Product Video / Motion Design, Corporate / Training, Explainer, Trailer / Teaser, Vlog / YouTube Long-Form, Commercial / TVC, Kids Content, Meme / Comedy Edit | genres: Comedy, Action, Science / Educational, Family, Musical, Sports Drama, Adventure  
**Avoid when:** On every transition it turns the piece into a loud slide show. In drama and cinematic documentary it reads as a cheap library effect.  
**Prompt:** `editor: lay a short air whoosh starting five frames before the transition and tailing off five frames after it`  
**Yields to:** `sound.bridge` — Drama and cinematic documentary, where a whoosh reads cheap.  
`sound.whoosh`

### Audio Led Whip

**Also called:** sound leading the transition, audio driven cut  
**What it is:** Sound performs the entire transition: a sweep, a hit or a moving effect travelling from one ear to the other, with the picture cutting in the middle of it.  
**Effect on the audience:** The ear travels and the eye follows. You can join two shots with no visual relationship at all and the transition still convinces.  
**Used for and where it works best:** Reels, music video and advertising, and in documentary between two segments. The practical trick is to pan the sound across the stereo field in the same direction as the movement in the picture. With generated clips it also hides differences in colour and style between them.  
**Best in:** formats: Vertical Short-Form (Reels, Shorts, TikTok), Music Video, Commercial / TVC, AI-Generated Clip, Documentary Feature, Video Essay, Trailer / Teaser, Product Video / Motion Design, Explainer, Vlog / YouTube Long-Form | genres: Musical, Action, Science / Educational, Thriller / Suspense, Comedy, Heritage (civilisation-focused), Adventure, Political Thriller  
**Avoid when:** If the effect is too loud it steals attention from the picture. And in quiet scenes any transition sound breaks the mood.  
**Prompt:** `editor: pan a moving sound across the stereo field in the direction of the movement and place the picture cut at its centre`  
**Yields to:** `cut.hard` — Quiet scenes, where any transition sound breaks the mood.  
`sound.audio_whip`

### The Drone Bed

**Also called:** low drone, sub bass bed, tension bed  
**What it is:** A very low continuous layer under the scene, barely audible as a sound but felt in the chest, running without change.  
**Effect on the audience:** It generates permanent unease without the viewer knowing where it comes from. The body senses threat and the mind cannot name it.  
**Used for and where it works best:** Horror, suspense, science fiction and crime documentary, and any long scene you want tense without obvious music. In heritage work a low bed under temple shots creates awe instead of another epic score. Cut it suddenly when you want relief.  
**Best in:** formats: Feature Film, Drama Series Episode, Documentary Feature, Documentary Series, Heritage / Historical Documentary, Game Cinematic / Gameplay, Trailer / Teaser, Short Film, 360 / VR, AI-Generated Clip | genres: Horror, Thriller / Suspense, Psychological Horror, Sci-Fi, True Crime, Mystery / Whodunit, War, Dystopia / Post-Apocalyptic, Heritage (civilisation-focused)  
**Avoid when:** If the bed runs through the whole film the ear adapts and the effect dies. It has to stop somewhere for its presence to be felt.  
**Prompt:** `mix: run a continuous low frequency bed under the scene at a level felt rather than heard, and cut it dead on the release`  
**Yields to:** `soundscape.score_minimal_drone` — The bed should be pitched underscore, not felt sub.  
`sound.drone_bed`

### The Heartbeat Under the Scene

**Also called:** heartbeat cut, pulse bed  
**What it is:** A heartbeat or a steady pulse runs under the picture, and sometimes the cuts themselves land on each beat.  
**Effect on the audience:** It makes the viewer share the character physiologically: their pulse syncs with what they hear, and speeding it up genuinely raises their heart rate.  
**Used for and where it works best:** Fear, waiting, physical effort or dying, and sports before the decisive kick. Accelerate the pulse gradually while shortening the shots toward the climax. In medical documentary it works literally against the image of the heart.  
**Best in:** formats: Feature Film, Drama Series Episode, Trailer / Teaser, Sports Broadcast / Sports Film, Vertical Short-Form (Reels, Shorts, TikTok), Game Cinematic / Gameplay, Anime (Japanese 2D), Documentary Feature, Short Film | genres: Horror, Thriller / Suspense, Sports Drama, War, Survival, Medical Drama, Psychological Horror, Action, Disaster  
**Avoid when:** It is overused, so a loud obvious heartbeat plays as cliche. Keep it low under everything so it is felt rather than heard.  
**Prompt:** `mix: place a slow heartbeat under the scene, accelerate it as the sequence tightens, and stop it dead on the resolution`  
**Yields to:** `score.war_drum_acceleration_pulse` — The accelerating pulse is scored percussion.  
`sound.heartbeat`

### Diegetic to Non-Diegetic Switch

**Also called:** source to score switch, in world to out of world  
**What it is:** Music starts from a source inside the scene, a radio, a player, a phone, and then becomes the film's own score at full level and quality, or the reverse.  
**Effect on the audience:** It opens the scene from reality into feeling. The moment the music changes state is the moment the scene turns from information into emotion.  
**Used for and where it works best:** The start of a montage, the dance beat, the character stepping out into the street, the ending. In music video it is the natural crossing from performance into the imagined world. The reverse, from score down into a radio inside the room, drops the viewer back into reality hard.  
**Best in:** formats: Feature Film, Music Video, Drama Series Episode, Short Film, Commercial / TVC, Limited Series, Anime (Japanese 2D), Trailer / Teaser, Dance Film | genres: Drama, Romance, Musical, Coming of Age, Crime / Procedural, Comedy, Melodrama, Gangster, Tragedy  
**Avoid when:** If the switch happens without a dramatic moment to justify it, it sounds like a mixing error. And the audio quality must change clearly so the viewer feels the crossing.  
**Prompt:** `mix: start the track thin and filtered as if from a speaker in the room, then open it to full stereo score on the emotional turn`  
**Yields to:** `score.diegetic_to_score_transmutation` — A played instrument taken over by the orchestra.  
`sound.diegetic_switch`

### Voice-Over Bridge

**Also called:** narration bridge, commentary bridge  
**What it is:** The narrator's voice continues across several shots and scenes and binds them, so the images change beneath one unbroken line of speech.  
**Effect on the audience:** The speech becomes a thread and the images become beads on it. The viewer accepts any jump in place or time because the voice never stopped.  
**Used for and where it works best:** Documentary, heritage work, video essays and narrated fiction. The golden rule is that the picture must add to the words rather than repeat them. In advertising, one voice across ten shots builds a story in thirty seconds.  
**Best in:** formats: Documentary Feature, Heritage / Historical Documentary, Documentary Series, Video Essay, Commercial / TVC, News Report / Reportage, Explainer, Feature Film, Corporate / Training, Trailer / Teaser | genres: Heritage (civilisation-focused), Historical / Biopic, Science / Educational, True Crime, Nature / Wildlife, Political Thriller, Drama, War, Faith / Biblical Epic  
**Avoid when:** If the narration explains everything in the picture the viewer stops looking. And a voice with no gaps tires the ear and the meaning evaporates.  
**Prompt:** `editor: cut the pictures under one continuous narration line, changing image on the phrase breaks and never on mid word`  
**Yields to:** `sub.voiceover_counterpoint` — Narration that would only describe the picture.  
`sound.vo_bridge`

### Off-Screen Sound

**Also called:** acousmatic sound, unseen source  
**What it is:** A sound whose source is not in frame: a knock, footsteps, a scream, breathing, an animal. We hear it and never see it.  
**Effect on the audience:** Imagination builds the source, and anything imagination builds beats anything you can show. It also enlarges the world: the film becomes bigger than the frame.  
**Used for and where it works best:** Horror and suspense before the reveal, drama to build a living world behind the walls, and documentary to give a sense of place. With generated clips it is a golden tool: crowd or battle sound makes a single shot feel like a whole world.  
**Best in:** formats: Feature Film, Drama Series Episode, Short Film, AI-Generated Clip, Documentary Feature, Heritage / Historical Documentary, Game Cinematic / Gameplay, 360 / VR, Anime (Japanese 2D), Vertical Micro-Drama | genres: Horror, Thriller / Suspense, Psychological Horror, Drama, War, Mystery / Whodunit, Supernatural Horror, Heritage (civilisation-focused), Survival  
**Avoid when:** If you never reveal the source in a scene that demands an answer the viewer feels cheated. Decide whether this is deliberate mystery or delayed information.  
**Prompt:** `mix: place the key sound clearly off screen and hold the camera on the listening face, never cutting to the source`  
`sound.offscreen`

### Ambience Swap

**Also called:** atmos change, room tone switch  
**What it is:** You change the ambience layer alone with no change in picture: the street noise goes and interior silence, wind or a reverb takes its place.  
**Effect on the audience:** The viewer feels the place changed, or that the character has detached from reality, with no visual device at all.  
**Used for and where it works best:** The moment of mental detachment inside a crowd, the entrance to a memory, the loss of consciousness. In documentary and heritage work, swapping the modern ambience for the sound of the ancient era over the same picture makes the viewer travel with no cut at all. In editing it is a silent signal that the scene has turned internal.  
**Best in:** formats: Feature Film, Heritage / Historical Documentary, Documentary Feature, Drama Series Episode, Short Film, AI-Generated Clip, Anime (Japanese 2D), Docudrama / Reenactment, 360 / VR | genres: Drama, Heritage (civilisation-focused), Historical / Biopic, Psychological / Arthouse, War, Psychological Horror, Mythology / Ancient Epic, Tragedy, Faith / Biblical Epic  
**Avoid when:** A sudden swap with no dramatic reason sounds like a mixing error. Make the change take half a second to two seconds so it reads as a decision.  
**Prompt:** `mix: cross fade the ambience layer to a different world over one second while the picture stays exactly the same`  
**Yields to:** `sound.subjective` — Detachment played from inside the character's head.  
`sound.ambience_swap`

### Echo of the Last Word

**Also called:** word echo, repeated line echo  
**What it is:** The last word or line of a scene returns as an echo or a faint repetition over the following scene.  
**Effect on the audience:** It turns the line into a fixed idea in the head of the character and of the viewer. The word is no longer dialogue, it is a wound.  
**Used for and where it works best:** After an insult, an accusation or a broken promise, and in memory and trauma scenes. In psychological drama and melodrama it is the standard language of remembering. In heritage work a line from a prophecy or a curse returns in echo over a battle scene.  
**Best in:** formats: Feature Film, Drama Series Episode, Soap Opera / Telenovela, Short Film, Anime (Japanese 2D), Music Video, Vertical Micro-Drama, Heritage / Historical Documentary, Docudrama / Reenactment | genres: Melodrama, Psychological / Arthouse, Drama, Psychological Horror, Tragedy, Romance, Mythology / Ancient Epic, Horror, Supernatural Horror  
**Avoid when:** Daily soaps wore it out until it became material for parody. Use it once in an entire piece, and with a light echo rather than a heavy one.  
**Prompt:** `mix: repeat the final line once at lower level with a long reverb tail over the opening of the next scene`  
**Yields to:** `sound.postlap` — Soap-worn repetition; carry the line plainly instead.  
`sound.echo_last_word`

### Subjective Sound

**Also called:** POV sound, muffled interior sound  
**What it is:** All the sound is heard from inside the character's head: muffled, with tinnitus, or underwater, with her breath and pulse louder than the entire world.  
**Effect on the audience:** It moves the viewer from observer to participant in an instant. It is the strongest subjective effect in cinema and cheaper than any visual device.  
**Used for and where it works best:** After a blast, underwater, during a panic attack, at the point of fainting. It suits war, sports and psychological drama. The method is to pull every detail down, leave breath, pulse and a high ringing tone, and bring the world back gradually as the character returns.  
**Best in:** formats: Feature Film, Drama Series Episode, Short Film, Game Cinematic / Gameplay, 360 / VR, Anime (Japanese 2D), AI-Generated Clip, Sports Broadcast / Sports Film, Trailer / Teaser | genres: War, Psychological Horror, Thriller / Suspense, Drama, Survival, Action, Sports Drama, Medical Drama, Horror  
**Avoid when:** If the muffled state runs too long the ear tires and the scene loses its information. In documentary it reads as manipulation when the event is real.  
**Prompt:** `mix: duck all exteriors, leave breath and pulse with a high ringing tone, then restore the full world over three seconds`  
**Example:** Saving Private Ryan, the muffled ringing after the blast on the beach.  
**Yields to:** `sound.ambience_swap` — Documentary, where a muffled POV reads as manipulation.  
`sound.subjective`

### Room Tone

**Also called:** presence, atmos floor  
**What it is:** The real emptiness of a space: a very low layer recorded on location and laid under everything, so that silence is never dead digital silence.  
**Effect on the audience:** It makes the scene feel real without anyone noticing. Its absence makes the viewer feel something is wrong without being able to name it.  
**Used for and where it works best:** Record the location's own tone before leaving and use it beneath ordinary dialogue trims, pauses and quiet coverage. A deliberately authored acoustic vacuum under `sound.silence_before` is the exception: its finite digital zero is the effect, so room tone returns only after that beat ends.  
**Best in:** formats: all | genres: all  
**Avoid when:** Absolute digital silence reads as a broken file. Borrowing the layer from another location usually gives the scene away, because the ear knows the difference even when the viewer cannot explain it; the exception is a location whose own tone is unusable and a close match recorded in a room like it.  
**Prompt:** `mix: record thirty seconds of the location's own tone and lay it under every cut in the scene, including the silences`  
**Yields to:** `soundscape.stem_ambience_bed` — The location's own tone is unusable.  
`sound.room_tone`

### Sound Perspective

**Also called:** sonic perspective, matching sound to shot size  
**What it is:** The size and reverb of the sound change with the size of the shot: in a wide the voice is distant with room in it, in a close up it is near and dry.  
**Effect on the audience:** It makes the viewer feel their ear is with the camera. When the perspective breaks the scene feels fake and the viewer cannot say why.  
**Used for and where it works best:** Any scene with changing sizes: dialogue, a speech, a temple, a hall. In heritage and epic work it is what conveys the scale of a space, a long reverb tail on the wide of the temple. In advertising the opposite applies: a close dry voice builds intimacy with the product.  
**Best in:** formats: Feature Film, Heritage / Historical Documentary, Documentary Feature, Drama Series Episode, Commercial / TVC, 3D CG Animation, Anime (Japanese 2D), Game Cinematic / Gameplay, 360 / VR, AI-Generated Clip | genres: Heritage (civilisation-focused), Historical / Biopic, Drama, Mythology / Ancient Epic, War, Faith / Biblical Epic, Horror, Fantasy / Epic, Sci-Fi  
**Avoid when:** If every line sits at the same closeness while the sizes change the scene feels flat and studio bound. And too much reverb makes the dialogue unintelligible.  
**Prompt:** `mix: match reverb and distance to the shot size, dry and close on the close ups, distant with room tail on the wides`  
**Yields to:** `mix.convolution_ir_matching` — A real named space, matched by impulse response.  
`sound.perspective`

### Overlapping Dialogue

**Also called:** overlapping lines, talking over  
**What it is:** Characters talk over each other and interrupt instead of waiting their turn, and the mix decides who is heard most clearly.  
**Effect on the audience:** It delivers real life, chaos and pressure. The viewer has to work to follow, so they pay more attention rather than less.  
**Used for and where it works best:** Family scenes, arguments, offices, emergency rooms, meetings. In medical and legal drama and naturalistic comedy it is what makes a scene feel unwritten. The rule is to leave the one important line clear of everyone else so it does not get lost.  
**Best in:** formats: Feature Film, Drama Series Episode, Sitcom / Multi-Cam Comedy, Limited Series, Short Film, Soap Opera / Telenovela, Documentary Feature, Mockumentary, Vertical Micro-Drama | genres: Comedy, Drama, Medical Drama, Legal Drama, Family, Dark Comedy / Satire, Crime / Procedural, Melodrama, Verite / Observational  
**Avoid when:** If necessary information is lost in the overlap you have lost the scene. And in dubbing and subtitling overlaps are a nightmare, so plan for the translated version.  
**Prompt:** `mix: overlap the secondary lines under the main one, keeping the story line clear and the rest as textured background`  
**Yields to:** `dial.subtitle_constraint` — The piece ships subtitled or dubbed.  
`sound.overlap_dialogue`

### Sound-Off Design

**Also called:** muted viewing design, silent scroll design  
**What it is:** Designing the video so it reads completely with the sound off: on screen text, clear facial expression, self explaining action, and a visible rhythm instead of an audible one.  
**Effect on the audience:** It guarantees the message lands, because a large share of platform viewing happens with the sound muted.  
**Used for and where it works best:** Reels, Shorts, TikTok and social advertising. The rule is that the first three seconds must read from the picture alone, and captions or text are part of the design rather than decoration. In vertical micro drama the burned in text is what carries the story.  
**Best in:** formats: Vertical Short-Form (Reels, Shorts, TikTok), Vertical Micro-Drama, Commercial / TVC, Product Video / Motion Design, Explainer, Meme / Comedy Edit, News Report / Reportage, Vlog / YouTube Long-Form, Corporate / Training, Kids Content | genres: Comedy, Drama, Science / Educational, Family, Romance, Thriller / Suspense, Political Thriller, Slice of Life  
**Avoid when:** If the joke or the fact lives entirely in the audio you lose most of your viewers. And text in the lower third is usually lost, because the platform interface sits on top of it; the exception is a horizontal player where that band stays clear.  
**Prompt:** `editor: design the cut to read fully muted, with accessibility captions fixed inside the current central safe band and one clear visual idea every three seconds; preview on the destination interface`  
**Yields to:** `promo.ugc_caption_carries` — Social ads, where sequential captions carry the points.  
`sound.sound_off`

### Close Textured Sound / ASMR Style

**Also called:** ASMR sound design, hyper close foley  
**What it is:** Recording and amplifying very small close sounds: rustling paper, bread cracking, wood being scraped, water pouring, a step on gravel.  
**Effect on the audience:** It creates a sense of touch. The viewer feels the texture in their body, and that stops the scroll harder than any music.  
**Used for and where it works best:** Food and product advertising, reels, and craft and heritage documentary, a chisel on stone, wet clay, a loom. In heritage work it is the strongest way to make the past feel real with no expensive shot at all. Record very close, lift the detail, and drop the music out or keep it very low.  
**Best in:** formats: Commercial / TVC, Product Video / Motion Design, Vertical Short-Form (Reels, Shorts, TikTok), Documentary Feature, Heritage / Historical Documentary, Vlog / YouTube Long-Form, Kids Content, Explainer, Travel / Real Estate / Hospitality, Corporate / Training | genres: Science / Educational, Heritage (civilisation-focused), Nature / Wildlife, Family, Slice of Life, Historical / Biopic, Romance, Preschool / Early Learning  
**Avoid when:** Lifting the detail above the music in an emotional scene becomes irritating and steals focus. And in serious drama over amplified texture sounds artificial.  
**Prompt:** `mix: record the texture very close, lift the fine detail above the music bed, and let one sound carry the whole shot`  
**Yields to:** `mix.transient_shaping_foley` — Serious drama, where lifted texture sounds artificial.  
`sound.asmr`