skills/ai-film-pipeline-master/references/phase-05-song-generation/cat-lyric-3.md

# Lyric, Verse, Refrain and Song Structure — Part 3

Category `lyric`, continued. Parts 1 and 2 are in `cat-lyric.md` and `cat-lyric-2.md` — read them first; their 66 cards cover song form from the verse-chorus spine outwards, hook siting and the title line, prosody and rhyme, the main song types, two-voice writing, the working process, and eight named failures. Nothing here repeats them.

**Part 3 of 4.** 10 cards across 3 sections: music as a written decision, placement inside a musical, and the album as a structure.

These cards close gaps the database reported about itself: each one was reached for by a kit and could not be cited because no card existed. Parts 1 and 2 treated a song as a thing being written. Part 3 treats it as a thing being *placed* — in a scene that has to be cleared and paid for, in an act that has to end, in a running order that someone will hear once, in order, on a Friday.

---

<!-- GENERATED:toc — do not edit by hand. Generated in the source package -->

## Contents

- [Music as a Written Decision](#music-as-a-written-decision)
- [Placement Inside a Musical](#placement-inside-a-musical)
- [The Album as a Structure](#the-album-as-a-structure)
- [Coverage](#coverage)
- [Sources](#sources)

<!-- /GENERATED:toc -->

## Music as a Written Decision

### Music Clearance as a Writing Constraint

**Also called:** the two-licence problem, sync and master clearance, rights-led song choice, writing the alternates, music supervision as a writing input
**What it is:** Every use of an existing recording needs two separate permissions that come from two different owners — a synchronisation licence for the composition, from its publisher, and a master use licence for that particular recording, from whoever owns it — and either can refuse for any reason, including not liking the scene. This card is optional craft for a brief that itself names an existing commercial recording and raises its rights: it shows how to write a scene that survives a substitution. It is never a required step, and this skill runs no clearance (root `SKILL.md` §0).
**Effect on the audience:** Invisible when it works. When it fails they meet the scene rebuilt around a substitute it was never designed for, and the substitution is usually audible: a moment that means slightly less than it was written to mean, cut to a rhythm the replacement does not have.
**Used for and where it works best:** Where the brief raises rights, apply it the moment a specific recording enters a draft, not at the music-supervision stage. Four variables set the price — **term, territory, media and usage** — and a fifth, prominence, multiplies it: a featured drop costs several times the same song used as background, and a licence bought for festivals does not cover streaming. 2026 practice runs roughly from nothing on a student film, through $0–1,500 all-in on an indie short and about $1,000 for festival-only rights, to around $45,000 for an indie feature in distribution, $3,000–45,000 for a network television episode, $50,000 and up for national advertising, and $100,000–250,000 and far beyond for a famous recording in a global campaign or a studio trailer. Most deals go out on Most Favoured Nations parity, so publishing and master are quoted equal and a hard negotiation with one side has to be won twice. The craft that follows: **write the alternates into the script, so the scene's meaning does not hang on one recording's lyric.** Name a second and a third choice in the scene description, and build the beat on what the character *does* with the music — leaves it on, turns it up, cannot turn it off — because behaviour survives a substitution and a lyric reference does not. See `lyric.needle_drop_lyric` for what the words then do inside the scene, and `emo.score_does_the_work` for the failure a famous recording most invites.
**Best in:** formats: Feature Film, Series, Trailer, Advertising, Documentary, Brand Film | genres: Drama, Crime, Coming of Age, Heritage, Journalism, Popular Music
**Shelf life:** dated — review 2027-09. The two-licence structure and the four variables are durable; the fee bands are 2026 practice and will move, and they move fastest at the advertising end.
**Avoid when:** The brief does not raise rights — the skill does not raise them for it. And when the music is commissioned, or drawn from a production library, or written for the project — one owner, one licence, and the whole constraint collapses into a fee. It also does not apply to a composition in the public domain *performed newly for you*; it does apply the moment you reach for someone else's recording of it, which is the trap, because the song being old says nothing about who owns the record. The real failure mode is not refusal, it is timing: clearance runs weeks to months, estates are slower than labels, and a scene built on one record in the second draft is a scene that gets rewritten in the edit.
**Example:**
> *A scene as most writers first draft it:*
> INT. TAXI — NIGHT. The radio is playing "Ya Zaman" — the 1961 recording, the famous one. She listens to the last verse and cries.
>
> *The same scene written to survive clearance:*
> INT. TAXI — NIGHT
> The radio is already playing when the door shuts. She reaches for the dial, and doesn't.
> She lets it run all the way to the end of the street.
> `[MUSIC — 1st choice: "Ya Zaman," 1961 master. Publishing and master both sit with the family; approached 12 March, no answer yet. 2nd: "The Long Bridge," same year, same room, label has cleared for us before. 3rd: original composition, same tempo, same language. The beat is her not turning it off. No line of dialogue refers to the words.]`

**Yields to:** `lyric.writing_to_brief` — Music is commissioned for the project; one owner.  
`lyric.music_clearance_constraint`

---

### The Spotting Session

**Also called:** spotting, the music spot, spotting notes, in and out points, hit markers, the cue-placement meeting
**What it is:** The meeting held after picture lock where the director, composer, editor and music supervisor watch the cut and decide, to the frame, where every cue starts, where it stops and what it hits. Everyone leaves with a timecoded spreadsheet — the spotting notes — of in points, out points, hit markers and descriptive markers, and that spreadsheet is the score's brief.
**Effect on the audience:** A cue's in-point is the loudest editorial statement a film makes without words. Music that starts *on* a line tells the audience the line mattered. Music that starts a beat *after* it tells them the silence mattered. Music that was already running tells them the outcome was never in doubt. The audience cannot name any of this and reacts to all of it.
**Used for and where it works best:** The writer is almost never in the room and usually wrote the scene a year earlier, so the work is done upstream, in the script. The craft rule: **wherever you want a cue to enter, put a physical event there.** A spotting session lands music on something visible — a door, a name spoken, a hand leaving a table, a chair scraping — because a cut gives a composer a frame and an intention does not. Write two or three candidate events into a scene and you have written the in-point without presuming to place it. Equally, mark where music must *not* go: the beat of dialogue that has to arrive dry is a decision made at the spotting session by default unless the writing makes the silence load-bearing. Use it on anything with a score, and hardest on wordless sequences, where the in and out points are the only structure the audience gets — see `seq.sequence_spine` for the sequence underneath and `emo.score_does_the_work` for the failure the session exists to prevent.
**Best in:** formats: Feature Film, Documentary, Series, Trailer, Animation Feature | genres: Drama, Heritage, Historical, Thriller, Biography
**Avoid when:** The piece is scoreless or diegetic-only, or the music is being written first and cut to — a music video, a lyric video, a musical number. There the picture serves the track and the arrow points the other way. The honest failure mode on the writing side is the stage direction "MUSIC SWELLS," which is an instruction with nothing under it; the session will land the cue wherever the picture offers a landing, and that is frequently four seconds earlier than the writer meant.
**Example:**
> *Before — nothing for a cue to land on:*
> She reads the letter. MUSIC SWELLS.
>
> *After — three candidate in-points written as events:*
> She reads the letter.
> She folds it back along its old creases. `[A]`
> She puts it in the drawer with the others. `[B]`
> Then she pushes the chair back and goes to the window. `[C]`
>
> The scene now tells the spotting session what the music is for. A cue at A is about the letter. At B it is about the other letters. At C it is about her giving up, which is the scene. The writer never named a timecode and has placed the cue anyway.

**Yields to:** `lyric.needle_drop_lyric` — Diegetic source song rather than score.  
`lyric.spotting_session`

---

### The Needle-Drop's Lyric Problem

**Also called:** the second voice, lyrics under dialogue, source-song interference, the sung line in a spoken scene
**What it is:** A recorded song brought into a scene arrives with words, and those words are a second voice speaking at the same time as the dialogue, on the same channel, in the same language, competing for the same listener. Nothing else in a film's soundtrack does this.
**Effect on the audience:** Two speaking voices at once produce not emphasis but attrition — the audience loses some of both and knows only that the scene felt crowded. Handled the other way, the interference becomes the point: a sung line answering a line of dialogue is the cheapest irony in cinema and one of the most reliable, because the audience does the joining themselves and enjoys having done it.
**Used for and where it works best:** Decide which voice is speaking in every second of the scene and write accordingly. Three settings work and nothing between them does. **One: give the lyric air** — 30 to 90 seconds of scene with no dialogue at all, where the song is the text and the actors play listening. **Two: land on the instrumental** — write the dialogue to sit over a verse's instrumental bars or an intro, and write the *silences* in the dialogue where the vocal enters. **Three: let one sung line answer one spoken line**, deliberately, once, and never twice in the same film. The craft rule that governs all three: **the lyric is dialogue, so block it like dialogue** — who speaks, in what order, and who is interrupted. Write the song's vocal entry into the scene as a beat with a timing, the way you would write a character coming through a door. Pairs with `lyric.music_clearance_constraint`, which decides whether you get the record at all, and `lyric.on_screen_lyric_text` where the words are also shown.
**Best in:** formats: Feature Film, Series, Short Film, Music Video, Vertical Micro-Drama | genres: Drama, Coming of Age, Crime, Romance, Popular Music
**Avoid when:** The song is instrumental, or is in a language the audience does not read as language — then it behaves as score, and this card does not apply. Avoid it altogether in a scene carrying plot the audience must not miss: a needle-drop with a vocal is a scene where roughly one line in three will be half-heard, and that is an acceptable price for a mood and never for information. The commonest failure is a famous record dropped under a full dialogue scene, where the production has paid a featured-use fee for something the audience could not hear.
**Example:**
> *Before — both voices at once, and neither lands:*
> The chorus comes in. Big, wall-to-wall.
> MARWAN: You don't have to explain it to me. I was there.
> LEILA: You were in the car. That isn't the same as being there.
>
> *After — the lyric blocked as a third speaker:*
> MARWAN: You don't have to explain it to me.
> *The vocal enters — a man's voice, one line, the first words in the song: "I came back and the street had moved."*
> Neither of them says anything for eight seconds.
> LEILA: You were in the car.
> *Under the second verse's instrumental bars —*
> LEILA (CONT'D): That isn't the same as being there.
>
> Same dialogue, one word cut, and the song now answers Marwan instead of shouting over him.

**Source:** assumption — the package's working figure, not a published standard; set it by measuring the scene in the cut.  
**Yields to:** `lyric.spotting_session` — Instrumental or unread language; it behaves as score.  
`lyric.needle_drop_lyric`

---

### Temp Love

**Also called:** temp-track attachment, temp dependency, scoring the temp, the temp trap
**This is a named failure, not a recommendation.**
**What it is:** The cut is edited for months against borrowed music. Everyone involved stops hearing the film and starts hearing the temp, and the score, the scene lengths and sometimes the writing are then judged against a specific recording that cannot be cleared and must not be reproduced.
**Effect on the audience:** They never hear the temp, so they meet only the consequence: cues written to imitate something absent, scenes cut to a rhythm the final music cannot supply, and a general second-hand quality nobody in the room can locate any more because everybody in the room has heard the other version two hundred times.
**Used for and where it works best:** — There is no version of this that is a technique. Temp music itself is legitimate and unavoidable; the failure is the attachment, and it hurts most where post-production is long and the temp is famous — features, studio trailers, and heritage documentary cut against archive recordings that will never clear.
**Best in:** formats: — | genres: —
**Avoid when:** Always, and the defence is procedural rather than artistic. Temp a scene with several unlike pieces rather than one, so no single recording becomes the scene's identity. Hold the spotting session against the cut with the temp muted at least once (`lyric.spotting_session`), because a cue's in-point should be decided by the picture and not by where a borrowed record happened to change. And on the writing side, never adjust a scene's length to a temp cue — that is the point at which a rights problem becomes a structure problem. The diagnostic: if you can only describe what a scene needs by naming a record, the film has temp love and the note is unanswerable as written.
**Example:**
> *A note with temp love in it:*
> "It just needs to feel like the temp. Can it build the way the temp builds at 1:12?"
>
> *The same note, made answerable:*
> "At 1:12 she stops arguing and starts agreeing with him, and we need to hear that she has lost. Right now the cue is still on her side."
>
> The first note asks a composer to copy something they are legally forbidden to copy. The second names a value turn, which is a brief.

**Yields to:** `lyric.spotting_session` — Spot the cut once with the temp muted.  
`lyric.temp_love`

---

## Placement Inside a Musical

### The Eleven O'Clock Number

**Also called:** the eleven o'clock song, the late big number, the second-act showstopper, the bookend to the "I want" song
**What it is:** A large, character-defining song placed late in the second act, shortly before the finale but never as the finale, in which the protagonist arrives at a decision, a revelation or a change of heart. The name is literal: when curtains went up at 8:30, this song landed around eleven.
**Effect on the audience:** It is the last thing they will be given room to feel before the plot closes, so it is the one they leave humming and the one they quote. It also converts a plot resolution into a personal one — without it an audience watches a situation being sorted out; with it they watch a person deciding.
**Used for and where it works best:** Write it as the answer to the "I want" song, because in the Lehman Engel model that is exactly what it is — same character, same want, and the whole show between them as evidence. The craft rules are positional and strict. **It is not the finale**; something has to happen after it, or the show ends on a solo and the company has nowhere to go. **The change happens inside the song, not before it** — a character who has already decided is singing a report, and a report cannot stop a show. **It is the character's largest vocal and emotional demand**, which is a casting fact as much as a writing one, so it must be written to a voice (`lyric.written_for_the_voice`). And it is nearly always sung by the character the audience has waited longest for. Pairs with `lyric.i_want_song` as its bookend and with `lyric.reprise`, which is what this number becomes if you write it as a restatement rather than an arrival.
**Best in:** formats: Musical Theatre, Stage Play, Opera, Feature Film, Animation Feature | genres: Musical, Drama, Comedy, Redemption, Coming of Age
**Avoid when:** The show is one act, or through-composed with no dynamic room left to make a peak, or the protagonist's arc is deliberately one of *not* changing — a satire or a tragedy of stasis, where a big late aria of self-knowledge would break the argument. Avoid it also for a character who has not carried the evening; giving a showstopper to a supporting role at eleven o'clock reorganises the show around them, which occasionally is the intention and usually is an accident. The honest failure: writing a number that is large but not a turn. Size is not a function. An audience will applaud it and the second act will not have moved.
**Example:**
> *Act One "I want," the character at the top of the show:*
> "Somebody's going to ask me what I do — / and I'm going to have an answer for them."
>
> *Eleven o'clock, same character, same want, everything in between as evidence:*
> "Nobody asked me. Not in twenty years. / I asked. I stood in my own hallway and I asked, / and the house said nothing, / and I heard it, / and I am going to be the one who answers."
>
> The want has not changed. What changed is who she is waiting on, and the song is where she finds that out — not where she reports it.

`lyric.eleven_oclock_number`

---

### The Act-One Finale

**Also called:** the pre-interval number, the first-act curtain, the act one closer, the bar-and-back number
**What it is:** The last number before the interval. Structurally it is the one song in a musical with a commercial job: raise everything, resolve nothing, and send an audience out of the room in a state that brings them back into it twenty minutes later.
**Effect on the audience:** It manufactures the conversation they have in the bar. A working act-one finale leaves them with a question they want to argue about and enough energy to argue; a weak one leaves them checking how long act two is.
**Used for and where it works best:** Treat it as the show's second-largest number and write it for the whole company. The rules are unusually mechanical for musical theatre and they hold. **It must be bigger than anything before it** — musically, dramatically, or in the number of people on stage — because the audience reads scale as importance. **It must turn, not conclude**: the protagonist commits to something irreversible, or learns the thing that makes act two necessary, and the curtain falls with the consequence unplayed. **The last sung line should be answerable, not answered** — a question, a promise, or a claim the audience already doubts. And it should carry the whole company, because a solo curtain hands the interval to one person and a company curtain hands it to the story. Cross-refs: `seq.act_rhythm` for the act shape underneath it and `tens.stage_blackout` for what the blackout itself does.
**Best in:** formats: Musical Theatre, Stage Play, Opera, Live Event, Series | genres: Musical, Drama, Comedy, Ensemble, Epic
**Avoid when:** There is no interval — a one-act, a ninety-minute film musical, a streaming episode. Writing a false curtain into a continuous piece produces a dead stop at the two-thirds mark that the audience reads as the ending, and everything after it plays as epilogue. Avoid the literal cliffhanger too, where possible: a shock at the curtain is cheaper than a commitment, and an audience that has been shocked comes back curious, where an audience that has watched someone commit comes back invested. The real failure mode is the act-one finale that resolves its own question, which is the commonest structural fault in new musicals and unfixable in act two.
**Example:**
> *Resolved — the interval has nothing to do:*
> COMPANY: "And so we signed it, and the land was ours, / and everything we lost came home tonight."
>
> *Turned — the same beat, left open:*
> FATHER: "I signed it."
> DAUGHTER: "You signed *what*?"
> COMPANY: "He signed it, he signed it, the ink is still wet on the table —"
> FATHER (alone, over them): "I will explain it to you in the morning."
> COMPANY: "— and the morning is coming, and the morning is coming, and the morning is *coming* —"
> *Blackout.*
>
> Nothing is concluded, the company is on stage, the last line is a promise the audience does not believe, and there is an argument to have in the bar.

`lyric.act_one_finale`

---

## The Album as a Structure

### Album Sequencing as a Craft Object

**Also called:** the running order, tracklisting, the sequence, side-one side-two planning
**What it is:** The order of the songs, treated as a structure with a shape rather than as a list of finished things. A running order decides what the record's argument is, in the same way that scene order decides a film's.
**Effect on the audience:** On a first listen, almost everything. The same eleven songs in two orders are two different records: one has a thesis and a turn, the other has a good first half. Listeners rarely notice sequencing and universally feel it, which is why the complaint that a record "sags" is nearly always a sequencing complaint rather than a songwriting one.
**Used for and where it works best:** Sequence the way you would outline: decide what the record claims, then order the songs so the claim is stated, tested and answered. Working rules that survive contact with real records — **track one is a statement of identity, not the single**; the song a listener should meet second is the most immediately likeable, because that is where a stranger decides to stay; **the strongest song is rarely last** — put it around two-thirds, where a record either lifts or dies; the slow middle is where records lose people, so no two consecutive songs may share tempo, key *and* subject; and **the last track is the one the record is about**, which is a different question from which one is best. Sequence by key and tempo relationships as well as by meaning, because a listener feels an awkward key change between tracks without being able to name it. Under streaming, add one more: **the first thirty seconds of every track is now a decision point**, because that is where a stream counts and a skip registers, so a record that opens three tracks in a row with long instrumental intros will be measured as three skips. Related: `seq.sequence_spine`, borrowed from narrative and written for it; `lyric.second_verse_repeat`, which happens at album scale as well as verse scale.
**Best in:** formats: Song Lyric, Music Video, Podcast, Audio Drama, Live Event | genres: Popular Music, Heritage, Faith, Testimony, Coming of Age
**Shelf life:** dated — review 2028-06. The craft of sequencing is durable; the streaming pressures — the thirty-second stream threshold, skip-rate reporting, playlist-first release patterns — are 2026 conditions and are the part most likely to move.
**Avoid when:** The release is a single, an EP of three, or a playlist-first campaign where the tracks will be encountered separately and in an order you do not control. Sequencing a record nobody will hear in order is effort spent on an experience that will not occur — though note that the people who care most about you are precisely the ones who will hear it in order, which is why the practice survives. The honest failure: sequencing to protect weak songs by burying them, which produces a record with no back half and is always more obvious than cutting them would have been.
**Example:**
> *A running order written as a list — best songs first:*
> 1. the single · 2. the other single · 3. the big one · 4–9. everything else · 10. the slow one
>
> *The same ten songs sequenced as an argument:*
> 1. **"Name the Street"** — not the single; it says who is talking and where from
> 2. **"Forty Minutes North"** — the single; the stranger stays here or not at all
> 3. **"Cheap Rooms"** — first contradiction: he likes it there
> 4. **"Cheap Rooms (What My Mother Said)"** — 50 seconds, her voice only, no chorus
> 5. **"Work"** · 6. **"Work Again"** — deliberately similar; the record is being boring on purpose and says so
> 7. **"The Long Bridge"** — the strongest song, at two-thirds, where the lift has to come
> 8. **"Nobody Came"** · 9. **"The Letter I Didn't Send"** — the cost
> 10. **"Name the Street (Reprise)"** — same title, same melody, one word different in the last line
>
> Track 4 is an interlude (`lyric.interlude_track`), tracks 5 and 6 are a deliberate flat stretch, and track 10 answers track 1. That is a structure, not a running order.

`lyric.album_sequencing`

---

### The Segue

**Also called:** the crossfade, the gapless join, the run-on, the attacca, "tracks one and two are one track"
**What it is:** A written decision to join two tracks so that the second begins before the first has finished being heard — by crossfade, by a shared note or chord, by a held sound, by a line of speech carried across the join, or by simply removing the silence.
**Effect on the audience:** It removes the listener's exit. A gap between tracks is a place to decide whether to keep listening; a segue takes that decision away and, done twice or three times, converts a collection into a single object the listener is inside.
**Used for and where it works best:** Use it where two songs are one argument and the gap between them would let the listener treat them as two opinions. The craft rule: **a segue must carry something across the join** — a word, a pitch, an image, a rhythm — because a crossfade with nothing in common is a mixing effect and a listener hears it as sloppiness. The strongest joins are lyrical rather than sonic: end track four on a question and open track five on its answer, and the segue is written into the words before anyone touches a fader. Use it hardest where the pair is a before-and-after — the argument and the apology, the leaving and the arriving. Note the practical consequence: segued tracks are hard to playlist and are frequently separated anyway, so a segue is a choice to serve the album listener at the singles listener's expense, and should be made on purpose. Cross-refs: `lyric.outro_tag` for how the first track ends, `lyric.intro_line_cold_open` for how the second begins.
**Best in:** formats: Song Lyric, Audio Drama, Podcast, Live Event, Musical Theatre | genres: Popular Music, Epic, Testimony, Heritage, Faith
**Avoid when:** The record is singles-led, or the two tracks belong to different worlds, or the join is being used to hide the fact that neither track has an ending. Avoid segueing into the strongest song on the record: a listener who wants to go straight to it will be sent to the end of the previous track every time, and will eventually stop. The honest failure is the gapless album where every join is a crossfade, which does not read as continuity — it reads as a record with no endings, and by track six the listener has lost the ability to tell where anything began.
**Example:**
> *Two tracks, gap between them, no relationship written:*
> Track 4 ends: "...and I never said a word about it." `[fade to silence, 2s gap]`
> Track 5 opens: "Monday morning, half past six —"
>
> *The same two tracks, joined in the writing:*
> Track 4 ends: "...and I never said a word about it."
> *The last word is held. Under it, a kettle starts.*
> Track 5 opens, over the kettle, same pitch as the held word: "So here is the word. / Monday morning, half past six, / and here is the word I didn't say."
>
> The join carries three things — a pitch, a domestic sound, and the sentence one track left unfinished. None of them is a fader move.

**Yields to:** `lyric.outro_tag` — The join would hide a track with no ending.  
`lyric.segue`

---

### The Interlude

**Also called:** the skit, the short track, the connector, the breather, the spoken-word track
**What it is:** A deliberately short, deliberately incomplete track — thirty seconds to a minute and a half — that is not trying to be a song. It carries a voice, a field recording, a fragment of melody, a piece of dialogue or a change of room, and it exists to do something to the tracks either side of it.
**Effect on the audience:** It resets the ear. After three songs in the same voice at the same scale, a listener's attention has flattened without their noticing, and a minute of something that is not a song restores the contrast the next song needs. It also tells them, unmistakably, that someone is arranging this — which is the moment a collection becomes a record.
**Used for and where it works best:** Use an interlude to do one of four jobs and never two at once: **change the room** (a different space, a different fidelity, so the next song sounds like a new place); **hand the record to another voice** (a parent, an archive tape, a child, a crowd — the one place on a record where the artist need not be the speaker); **state the thesis plainly**, in speech, once, so the songs never have to; or **absorb a tonal collision** between two tracks that cannot sit next to each other. The craft rules: it must be shorter than the listener expects, it must not have a chorus, and it must be *about* something the surrounding songs are about, or it is a sketch that got onto a record. Do not put an interlude before the strongest song — the listener will skip it every time and land in the wrong place. Best placed at the quarter and three-quarter points, where records sag. Related: `lyric.album_sequencing`, `lyric.side_break`.
**Best in:** formats: Song Lyric, Podcast, Audio Drama, Audiobook, Live Event | genres: Popular Music, Heritage, Testimony, Faith, Migration
**Avoid when:** The record is short, singles-led, or pop-facing — there an interlude reads as a track that failed to become a song, and the skip data will say so. Avoid it on a first record by an artist nobody knows yet: an interlude asks for a listener's patience on credit. And avoid the comedy skit unless the comedy is the record's subject; nothing dates a record faster, because a joke has one good listen in it and an album is asking for fifty.
**Example:**
> *Track 4 — "Cheap Rooms (What My Mother Said)" — 0:51, no music until the last eight seconds:*
>
> > *(room tone, a kettle, a chair)*
> > "You want me to say it in English? Fine. I will say it in English.
> > He had a room. It had one window. The window faced a wall.
> > He wrote to me every month for four years and in every letter he said the room was fine.
> > *(pause)*
> > The room was not fine."
> > *(the piano figure from track 3 enters underneath, one octave lower, and does not resolve)*
>
> It is under a minute, has no chorus, hands the record to the one voice that can contradict the singer, and leaves an unresolved chord for track 5 to answer.

`lyric.interlude_track`

---

### The Side Break

**Also called:** the two-side structure, the flip, side one / side two, the built-in pause, the act break of a record
**What it is:** The pause designed into a record at its midpoint, inherited from vinyl and cassette, where the listener physically had to get up. It gives a record two halves, a first-half ending, and a second-half opening — three structural positions that a continuous playlist does not have.
**Effect on the audience:** A break the listener performs themselves is the strongest punctuation available to a record. The last track of side one is heard as an ending, the first track of side two as a fresh start, and the listener's own body enforces both. On a streaming service none of that is compelled, and the effect survives only in weakened form — as a shape the attentive listener notices and the shuffling listener never encounters.
**Used for and where it works best:** Write the two positions even when the format will not enforce them, because they are where a long record earns its length. **The side-one closer is a false ending**: it must be complete enough to satisfy and unfinished enough to need a second half — the album's act-one finale, and the same rules apply as in `lyric.act_one_finale`. **The side-two opener is a second first impression**: a listener returning from a break has partially reset, so it can change tempo, voice or register harder than any other position on the record, and it is the best place on an album to do something the first half did not license. Sequence to it deliberately: put the record's sharpest turn at track 6 of 11 and a record with no vinyl release still has a spine. On streaming, mark the break with the strongest available substitute — a long track, an interlude, a segue that stops rather than carries — because the platform gives you no other way to say "half".
**Best in:** formats: Song Lyric, Podcast, Audio Drama, Audiobook, Live Event | genres: Popular Music, Epic, Heritage, Testimony, Faith
**Shelf life:** dated — review 2028-06. The two-half structure is durable and predates vinyl; what expires is the claim about how much of it survives a given listening platform. Vinyl's commercial revival has pushed real side lengths back into sequencing conversations, and that could deepen or reverse.
**Avoid when:** The record is under about thirty minutes, where a midpoint break halves it into two things too small to have shape, or where the release is explicitly playlist-first. And do not let the side break become an excuse: "side two is where it gets weird" is how records end up with a good first half and a demo reel after it. The honest failure mode is a side-one closer written as a full stop — a genuine ending, beautifully done — after which a meaningful number of listeners simply do not return.
**Example:**
> *Side one closer, track 5 — complete and unfinished at once. Last verse only:*
> > "So we put the whole house in a van, and the van went south,
> > and I stood in the road until I couldn't see it,
> > and then I stood in the road.
> > *(long instrumental, ends on the unresolved chord from track 4)*"
>
> *Side two opener, track 6 — a second first impression. New tempo, new speaker, no preamble:*
> > "Forty years later, a girl in Michigan
> > who has never seen that road
> > is being told about it wrong
> > by me."
>
> Track 5 ends the story and refuses to close it. Track 6 changes decade, country, and narrator inside four lines — a move that would be a jolt at track 4 and reads as a new act at track 6.

**Yields to:** `lyric.interlude_track` — Streaming enforces no break; mark the half instead.  
`lyric.side_break`

---

## Coverage

**10 cards** in this file, across 3 sections: Music as a Written Decision (4), Placement Inside a Musical (2), The Album as a Structure (4).

**What the research showed that the brief did not anticipate.**

*The clearance card should not have been split and nearly was.* The brief describes one card; the material is large enough for three (the two-licence structure, the four price variables, the craft consequence) and it is written as one anyway, because a writer never meets those three separately. The price variables are only interesting as the reason the alternates get written. Splitting it would have produced two cards nobody would open.

*The eleven o'clock number and the act-one finale are genuinely two cards, and the music kit was right to say so.* They are frequently discussed together and they are opposite problems. The eleven o'clock number is a character solving something in public; the act-one finale is a company refusing to solve something, on a schedule set by a bar. The one rule they share — bigger than what came before — is the least useful thing about either.

*Temp love is the card this section needed and the brief did not list.* It is named in the literature, it is the failure the spotting session exists to prevent, and it reaches the writer as an unanswerable note ("make it feel like the temp"). Without it, `lyric.spotting_session` describes a meeting with no account of how that meeting goes wrong.

*The side break is the most shelf-life-exposed card in the category* and was the hardest to write honestly. Half the practice literature treats it as dead and half treats it as revived by vinyl; what is durable is the two-half structure, and the card says so and dates the rest.

**Still missing, and recorded rather than hidden.** The music kits named one gap this file does not close: **the cutdown** — one piece of music built so that six, fifteen, thirty and sixty-second versions are each complete. It is written in `cat-rhythm-3.md` as `rhythm.cutdown_ladder`, because the craft is duration-writing rather than lyric-writing, and the lyric-specific part of it is already carried by `lyric.jingle_micro_hook`. Three further gaps surfaced while writing this file and are not covered anywhere: **the title track** as a decision (naming the record after a song changes the song); **the deluxe edition and the bonus track**, which is a real sequencing problem now that a record can be re-sequenced after release; and **the live rewrite** — the lines an artist changes when a song has been sung for ten years, which is `lyric.refrain_mutation` applied across a career rather than across a song.

Named failures in this file: `lyric.temp_love`.

Carrying a review date: `lyric.music_clearance_constraint` (review 2027-09), `lyric.album_sequencing` (review 2028-06), `lyric.side_break` (review 2028-06). The other seven are durable.

*Build notes from when this file was drafted — targets, running totals and what remained to be written — are kept verbatim in the [coverage history](../tools/history/coverage-history.md). They were true of the draft, not of the file, so they no longer sit here as if they were measurements.*

## Sources

Verified 17 September 2026.

- [Sync Licensing Fees: Film, TV, Ads, Games (2026 Data) — Dynamoi](https://dynamoi.com/learn/music-distribution/sync-licensing-fee-benchmarks)
- [Sync Licensing Rate Card 2026: Real Fees by Medium — Chartlex](https://www.chartlex.com/blog/business/sync-licensing-rate-card-2026)
- [How Much Does It Cost to License a Song? The 2026 Sync Fee Guide — Music Oracle](https://musicoracle.app/blog/how-much-does-it-cost-to-license-a-song)
- [Music Clearance for Independent Filmmakers: What to Clear, How, and Why Scope Is Everything — Thoolie](https://thoolie.com/creator_resource/music-clearance-independent-filmmakers/)
- [What is Sync Licensing? The 2026 Guide — Melody](https://usemelody.co/sync-licensing/what-is-sync-licensing)
- [How to License Your Music for TV, Film and Advertising in 2026 — FWD Music](https://fwdmusic.com/en/news/how-to-license-music-for-tv-film-advertising-2026)
- [Spotting (filmmaking) — Wikipedia](https://en.wikipedia.org/wiki/Spotting_(filmmaking))
- [How to Spot a Film — Art of Composing](https://www.artofcomposing.com/how-to-spot-a-film)
- [Spotting Session in Film — Soundbridge](https://www.soundbridge.io/spotting-session-film)
- [Spotting Sessions Demystified: Collaborating with Directors to Find the Perfect Musical Moments — Adrian Walther](https://www.adrianwalther.com/post/spotting-sessions-demystified-collaborating-with-directors-to-find-the-perfect-musical-moments)
- [Film editors and music — spotting session — Film Scoring Tips](https://filmscoringtips.com/film-editors-spotting-session/)
- [Temp track — Wikipedia](https://en.wikipedia.org/wiki/Temp_track)
- [How to Avoid the Temp Love Trap: Music Supervisors Weigh In — Soundstripe](https://www.soundstripe.com/blogs/how-to-avoid-temp-love)
- [What Is Temp Music in Film? — MasterClass](https://www.masterclass.com/articles/what-is-temp-music-in-film)
- [A Case Study for Film Composers Working with Temp Tracks — West Virginia University Research Repository](https://researchrepository.wvu.edu/cgi/viewcontent.cgi?article=8297&context=etd)
- [11 o'clock number — Wikipedia](https://en.wikipedia.org/wiki/11_o%27clock_number)
- [What Is an 11 O'Clock Number? — Playbill](https://playbill.com/article/what-is-an-11-oclock-number)
- [What's an Eleven O'Clock Number? — Counter Arts](https://medium.com/counterarts/whats-an-eleven-o-clock-number-b7fa14361d2d)
- [The 11 O'Clock Number! — CPA Theatricals](https://cpatheatricals.org/2017/the-11-oclock-number/)
- [How to Write a Great Act One Finale — London Musical Theatre Orchestra](https://www.lmto.org/how-to-write-a-great-act-one-finale/)
- [Hadestown and Act One Finales — Sophie A. Katz](https://sophieakatz.substack.com/p/hadestown-and-act-one-finales)
- [Elements of a Musical: The Book — Musicals101](https://www.musicals101.com/book.htm)
- [Album Sequencing in the Age of Streaming — iZotope](https://www.izotope.com/en/learn/album-sequencing-in-the-age-of-streaming.html)
- [Album Sequencing: 6 Ways to Give Your Tracklist That Perfect Flow — LANDR](https://blog.landr.com/album-sequencing/)
- [How to Sequence an Album — Mat Leffler-Schulman Mastering](https://matlefflerschulman.com/mastering-articles/how-to-sequence-an-album)
- [Bang Out of Order: The Importance of Great Album Sequencing — Music Musings & Such](https://www.musicmusingsandsuch.com/musicmusingsandsuch/2021/3/31/feature-bang-out-of-order-the-importance-of-great-album-sequencing)
- [What is an Interlude in Music, and Why Do Albums Have Them? — By The Barricade](https://www.bythebarricade.com/why-do-some-albums-have-interludes/)
- [What is an Interlude in an Album? — VIPZONE Samples](https://www.vipzone-samples.com/en/what-is-an-interlude-in-an-album/)
- [Fading Into the Next Song — TV Tropes](https://tvtropes.org/pmwiki/pmwiki.php/Main/FadingIntoTheNextSong)
- [The Power of Needle Drops: How Strategic Song Choices Shape Storytelling — LWKS](https://lwks.com/blog/the-power-of-needle-drops-how-strategic-song-choices-shape-storytelling)
- [The Nuances of the Needle Drop — Splice Today](https://www.splicetoday.com/music/the-nuances-of-the-needle-drop)
- [The Art of the Needle Drop: How Music Can Transform a Movie Scene — Trill](https://www.trillmag.com/entertainment/music/the-art-of-the-needle-drop-how-music-can-transform-a-movie-scene/)