Trailer music is the most formally constrained writing in the business. Ninety seconds, a fixed dramatic shape, and an editor who will cut your cue to pieces if it does not land where the picture needs it. The constraint is the reason it can be taught.

A film score follows picture. Trailer music is written before there is any picture to follow. It goes into a library, and an editor finds it later and cuts a trailer to it. That inversion is the single most important fact about the form, and almost everything below follows from it.

It means your cue has to contain the dramatic shape that the editor needs, at the timings they expect, with clean points to cut at. A beautiful piece of music with no structural landmarks is unusable. A fairly ordinary piece with the landmarks in the right places gets placed.

The shape

Nearly every trailer cue is built in four movements plus a tail. The names vary by library; the shape does not.

Section	Rough position	Job
Intro	0:00 – 0:15	Atmosphere. Establish key and world. Often almost no rhythm.
Act 1	0:15 – 0:45	The pulse arrives. Setup. Something is happening.
Act 2	0:45 – 1:15	The build. Stakes rise. Density and register climb.
Act 3	1:15 – 1:45	Full statement. The loudest, widest, most complete version.
Button	1:45 – 2:00	The hit after the silence. Title card.

Those timings are for a two-minute cue. The three common lengths are 60, 90 and 120 seconds, and you compress or expand the middle. Intro and Button stay roughly the same length regardless, because they are tied to the logo and the title card rather than to the story.

The editor is not looking for your best sixteen bars. They are looking for a structure they can hang a trailer on.

Why the sections must be separable

An editor will loop your Act 2 for forty extra seconds because the dialogue runs long. They will drop Act 1 entirely. They will take the Button and paste it somewhere else.

Write so that this is possible. In practice:

• Start sections on a downbeat of a phrase, in four- or eight-bar blocks. Sections that begin mid-phrase cannot be moved.
• End sections cleanly. A section whose last bar spills a reverb tail into the next one is fine; one whose musical idea only completes two bars later is not.
• Avoid unique events inside a section that make it unloopable. A one-off melodic turn in bar 6 of 8 means the editor cannot repeat those eight bars.
• Keep the tempo constant. Ritardandi and tempo changes are lovely and make a cue very hard to cut against. If you want the feeling of slowing down, use half-time instead: same grid, half the rhythmic density.

The vocabulary

A handful of sound-design devices do specific structural work. They are worth naming because using them without knowing the job produces cues that sound busy but do not build.

The braam

A low, distorted, sustained brass-and-synth hit, usually two to four seconds long, usually on a structural downbeat. Its job is punctuation. It ends a thought. One per section is plenty. Three in a row is a parody of the form, and the style has been ubiquitous long enough that libraries now reject cues built around them.

The riser

Any sound that ascends in pitch or intensity across two to eight bars and terminates on a downbeat. Its job is to make the listener feel a landing coming. The rule that matters: a riser must resolve. A riser that fades out instead of hitting something is the most common structural failure in amateur trailer writing, because it promises an arrival and then does not deliver one.

The reverse

A reversed cymbal, piano or impact placed so that its swell ends exactly on the downbeat of a new section. Subtler than a riser, and more useful, because you can put one at every section change without it becoming a gimmick.

The whoosh

Movement between moments. Carries the ear across a transition. Mostly a mixing decision rather than a compositional one, and easy to overdo. If you can count the whooshes on a first listen, there are too many.

The drop

Not a sound but an absence: everything stops except one element. Usually placed at 1:10 to 1:20, immediately before the final act. This is the most powerful device in the form and it costs nothing.

Silence is structural

The single change that improves most amateur trailer cues is to take something out at around the two-thirds mark.

Loudness is relative. After seventy seconds of accumulation, the listener's ear has adjusted, and Act 3 arriving at full density sounds only slightly larger than Act 2, because it is. Insert two bars where almost everything drops out, and the same Act 3 lands like a different piece of music. Nothing about it changed except what came immediately before.

The false ending

A refinement of the same idea. Act 3 finishes, everything stops, and there is genuine silence for one to two seconds, long enough that the cue appears to be over. Then a single hit, or a short tag.

Editors love this because the title card goes in the silence and the hit lands on the logo. If you write nothing else structural into a cue, write this.

Rhythm and the doubling trick

Trailer cues almost never change tempo, and they almost always feel like they accelerate. The mechanism is rhythmic density, and it works in a specific order.

1. Intro: no pulse at all, or very slow. Half notes, or free time.
2. Act 1: quarters. Weight, no urgency.
3. Act 2: eighths. Twice the motion, same tempo.
4. Act 3: sixteenths on top, quarters still underneath. Maximum apparent speed.
5. Button: back to nothing. One hit.

The crucial detail: the slow layer never leaves. Act 3 has sixteenths and the original quarters. If you replace the slow pulse rather than adding to it, the music gets busier but loses its floor, and it will sound thinner at the loudest point in the cue, which is the opposite of what you want.

The first fifteen seconds

Library editors audition in bulk. They will skip through dozens of cues in an afternoon, and most of what they hear gets ten or fifteen seconds before they move on. Whatever else the opening does, it has to establish within that window that the cue knows what it is.

That does not mean starting loud. It means starting with something specific: a texture with a character, a rhythmic idea, an interval that is going to matter later. What kills a cue at the audition stage is a generic pad while the composer works out what the piece is about.

Two openings that reliably work. The first is a single sound with a slow swell: one note, one texture, growing across four bars into the first pulse. The second is the cold open: the pulse starts immediately, bar 1, with nothing in front of it. Both tell the listener in two seconds what kind of cue this is. A vague atmospheric wash tells them nothing, and it is what most rejected cues open with.

Tempo

Most trailer music lives between 80 and 140 BPM, and the choice interacts with the section lengths in ways that are easy to get wrong.

Range	Feel	Note
70–85	Weight, dread, ceremony	Bars are long. An eight-bar section runs nearly half a minute, so plan fewer sections.
90–110	The default	A bar is roughly 2.2–2.7 seconds, so eight bars is about twenty. Section lengths land naturally on trailer timings.
120–140	Drive, action, chase	Bars fly past. Needs sixteen-bar sections or the cue feels chopped.
150+	Rarely useful	Usually better written at half the tempo with doubled subdivision: same energy, easier to cut against.

The half-time device is worth knowing properly. Write at 100 BPM and have the percussion play as though it were 50: backbeat on beat 3 instead of 2 and 4, kick on beat 1 only. The grid does not change, so everything stays cuttable, but the music sits back and feels enormous. Switching from half-time to straight time at the start of Act 3 is one of the largest structural gear changes available, and it costs nothing in the arrangement.

Key, mode and the harmonic decisions

Minor is the default, and it is the default for a reason: unresolved tension is the emotional register the form is selling. But the specific flavour of minor does a lot of work.

Choice	Character
Natural minor (Aeolian)	Neutral, heroic, the standard. Everything works.
Dorian (raised sixth)	Less tragic, more determined. Good for action rather than loss.
Phrygian (flat second)	Menace, the exotic, the antagonist. Use sparingly.
Harmonic minor	Gothic, occasionally cheap. Easy to overdo.
Minor with a major IV	The lift. One borrowed chord in an otherwise minor cue is the most reliable emotional hit in the form.

Two more decisions are worth making deliberately. First, harmonic rhythm: slower chord changes read as larger. Act 1 at one chord per bar, Act 3 at one chord per two bars, and the music has grown without a single new instrument. Second, the pedal point: hold the tonic in the bass through the whole of Act 2 while the upper harmony moves against it. It is the cheapest way to make a build feel like it is straining against something.

There is more on all of this in the guide to chord progressions for epic music.

Orchestration: spend it slowly

The commonest structural mistake is running out of orchestra. If Act 1 has full strings, brass, choir and percussion, Act 3 has nowhere to go.

Budget it. A workable allocation for a two-minute cue:

• Intro: one or two elements. A texture, a held note, a distant pulse. No brass.
• Act 1: low strings and a pulse. Maybe a solo instrument carrying a fragment of the theme. Still no brass.
• Act 2: full strings, horns entering quietly, percussion underneath. Choir may appear, low and wordless.
• Act 3: everything. Trumpets and high choir held back until here, because they are the two things that unmistakably say largest.
• Button: one hit, everything.

High trumpets and high choir are the ceiling of the form. If they appear at 0:30, the cue has already peaked and has ninety seconds left to fill.

Mixing decisions that are really structural

Three things get cues rejected that have nothing to do with the writing.

The low end is not cleared. Trailer mixes are dense and loud. If the braam, the taiko and the basses all occupy 40–80 Hz, the mix turns to porridge at exactly the moment it should be biggest. Give each one a band and high-pass the others out of it.

There is no headroom for the sound effects. The trailer's own sound design goes on top of your cue. Leave room in the 2–5 kHz range where dialogue lives, and do not master so hard that the editor cannot fit anything above it.

Stems are not provided. Most libraries require them, and the split is standard: drums and percussion, orchestra, synth and design, and a full mix. An editor who can mute your percussion for the dialogue section will use your cue. One who cannot will find another.

A ninety-second map

At 100 BPM in 4/4, one bar is 2.4 seconds. Ninety seconds is roughly 37 bars, which is awkward, so write 36 and let the tail run. Eight-bar blocks, one 4-bar break:

Bars	Time	Content
1–4	0:00–0:10	Intro. Texture and one sustained note. No pulse.
5–12	0:10–0:29	Act 1. Pulse in quarters, low strings, ostinato figure enters at bar 9.
13–20	0:29–0:48	Act 2. Eighths. Horns, percussion. Octave double on the figure. Riser in bars 19–20.
21–24	0:48–0:58	Act 2b. Harmonic rhythm halves. Choir enters low. Pedal point in the bass.
25–26	0:58–1:02	The drop. Everything out except one element. Reverse swell across bar 26.
27–34	1:02–1:22	Act 3. Full. Sixteenths on top, quarters still underneath. Trumpets and high choir.
35	1:22–1:24	Hard stop. Silence.
36	1:24–1:30	Button. One hit, long tail.

That structure is not a formula to hide. It is the shape the form asks for, and every choice inside it is still yours. The material in bars 5–12 is what makes the cue good or forgettable. The map only guarantees it is usable.

What gets a cue rejected

• It takes too long to start. Twenty seconds of atmosphere before anything happens. Libraries want the first hook inside fifteen seconds.
• No clear peak. The cue is loud throughout, so there is no moment to cut the money shot to.
• Tempo changes. Unusable for anyone cutting to a grid.
• A melody too distinctive to sit under dialogue. Trailer music accompanies. A tune that demands attention competes with the voiceover.
• It sounds like one specific well-known cue. Libraries are cautious about anything that could be mistaken for an existing recording, and so should you be.

What a library expects to receive

If the cue is going to a production library rather than a specific project, the deliverable is not one file. The expectations are fairly standard across libraries, and a cue that arrives incomplete gets held up or returned.

• The full mix, usually 48 kHz / 24-bit WAV, mastered but with headroom left.
• Alternate lengths. A 60-second edit and a 30-second edit, cut from the same material rather than rewritten. Make these yourself. If you leave it to the library, the edit points will not be the ones you would have chosen.
• An underscore version. The cue with the melodic content removed, so it can sit under dialogue. Frequently the version that actually gets placed.
• A no-drums version. Same reason.
• Stems. Typically four to six: percussion, orchestra, synth and design, choir if present, and the full mix.
• A hit or sting lifted from the Button, as its own short file.

This is worth knowing before you write rather than after, because it constrains the arrangement. A cue where the melody is woven inseparably through every layer cannot produce an underscore version. One where the melodic material lives on its own tracks can, and it takes ten minutes rather than an afternoon.

Where to start

Write Act 2 first. It is the part with the most content, it sets the key, the tempo and the ostinato, and everything else is derived from it. Act 1 is Act 2 with things removed, Act 3 is Act 2 with things added, the Intro is the texture left over, and the Button is the last chord.

Composers who start at bar 1 usually spend their best ideas on the Intro and arrive at Act 3 with nothing left. Start in the middle and the ladder builds itself.