Audio prompts
Tempo, instrumentation and the structure a generated cue needs to be cuttable.
Take audio prompts with youOne dense page to pin up, or a deck with one section per page. Both built from this page's own content.
Cheatsheet · 1 pageDeck · 18 pagesThe order
Working down this list is also the recommended order inside the prompt.
- Duration
- Role
- Mood
- Key and tonality
- Tempo
- Instrumentation
- Arrangement and dynamics
- Vocals
- Ending
Audio tags are a separate control surface
Some audio models take instructions inside the text in square brackets: [whispers] to change a delivery, [Chorus] to mark a section. ElevenLabs documents a vocabulary for this; Suno's structure tags are convention rather than published specification. Which one you are using changes the advice on this page about emotion cues.
Duration
Length of the piece.
A generator will otherwise give you a loop of arbitrary length that has to be cut, which is where the musical ending gets lost.
- 15 seconds
- 30 seconds
- 60 seconds
- 90 seconds
- 120 seconds
Phrases that work
- 90-second bed.
- 30 seconds with a clean ending.
- 15-second sting.
Sonilo Music and Mirelo Text to Audio both require an explicit --duration; Seed Audio does not.
Role
What the audio is for.
A bed under narration and a standalone track need opposite decisions about melody and midrange, so state the job before the style.
Phrases that work
- Instrumental music bed to sit under narration.
- Standalone track, no voice over it.
- Short sting to close a scene.
This is the single richest audio prompt guidance in the repo. One sound per prompt is the hard rule.
Mood
The feeling.
Pairs with tempo and key. Together those three settle most of the composition.
- warm
- playful
- cheerful
- buoyant
- wholesome
- ambient
- uplifting
- driving
- triumphant
- minimal
- dreamy
- suspenseful
Phrases that work
- Cheerful, wholesome and playful.
- Calm and spacious.
- Suspenseful and restrained.
The visual style contract is carried into audio prompts by translation, not by pasting the formula.
Key and tonality
Major or minor, and the key if it matters.
The cheapest emotional lever in audio. Major reads as safe and bright, minor as unresolved, and models respond to both.
- major key
- minor key
- modal
- pentatonic
Phrases that work
- Bright major key.
- Minor key, unresolved.
- Warm, clean major-key resolution in the final eight seconds.
Tempo
Speed, ideally as a number.
A number is obeyed far more reliably than a word like upbeat.
- 70 BPM
- 90 BPM
- 100 BPM
- 116 BPM
- 128 BPM
Phrases that work
- Buoyant mid-tempo around 116 BPM.
- Slow, roughly 70 BPM.
- Driving, 128 BPM, steady four on the floor.
Instrumentation
Which instruments play.
Listing instruments works better than naming a genre, which models interpret loosely.
- percussion
- ukulele
- piano
- bass
- handclaps
- marimba
- bells
- strings
- guitar
- xylophone
- drums
- harp
- synthesizer
- pads
- 808
- synth
- flute
- choir
- synthesized
Phrases that work
- Marimba, ukulele, xylophone, light percussion and soft handclaps.
- Solo piano with a distant string pad.
- Deep electronic ambience with soft mechanical clicks.
Tempo is named as a required aspect but the repo gives no BPM values or key/scale vocabulary anywhere, this is genuinely thin.
Arrangement and dynamics
How the piece develops.
Without this you get four bars looped flat. Naming a build and an ending is what makes it usable against picture.
- airy
- midrange
- headroom
- no abrupt hits
- narration-friendly
- smooth transitions
- gentle dynamics
Phrases that work
- Airy and narration-friendly, gentle dynamics, no abrupt hits.
- Build subtle energy throughout, then a celebratory lift at the end.
- Clear midrange space left for a voice.
Post-generation mixing, not prompt text, but it is the only mix guidance in the corpus.
Vocals
Whether anything sings or speaks.
Generators add humming and wordless vocals unprompted.
Phrases that work
- No vocals, no lyrics, no spoken words.
- Wordless humming only, no intelligible words.
- No prominent lead melody competing with the narration.
Ending
How it stops.
Ask for a resolution.
Phrases that work
- End on a clean major-key resolution, no fade.
- Ring out and decay naturally.
- Hard stop on the final beat.
Voice and spoken line
The prompt IS the line to be spoken; the voice itself is selected by id, never described in prose. Adjust --speech_rate modestly rather than rewriting a take.
Treat it as practice worth copying, not as a count.
- one voice per speaking entity
- exact line
- performance direction
Phrases that work
- The gate is open. Move!
- same voice, calmer delivery
- Voice prompts contain the exact line; performance direction belongs in the model-supported prompt, not undocumented flags.
- Voice: lock one voice per speaking entity and reuse it. Keep lines short enough not to block gameplay.
Narration writing (spoken script)
Narration is the one thing written in the user's chosen language; every image and video prompt stays English.
Treat it as practice worth copying, not as a count.
- 20-24 words
- about 8-9 seconds
- under roughly 9.5 seconds
- plain spoken text only
- no timecodes, emotion cues, parentheticals, or stage directions
- Spell numbers out
- hook through payoff
- ~150 words per minute
Phrases that work
- Write one plain line per clip, sized for about 8-9 seconds and normally 20-24 words. Keep every take under roughly 9.5 seconds.
- Use no timecodes, emotion cues, parentheticals, or stage directions.
- Spell numbers out.
- Set tone through word choice and concrete detail.
- Never say "in this video."
- For four and a half thousand years, the pyramids of Egypt have stood against the desert, silent and immense.
- Write scripts for speech, not text: short sentences, natural pauses, ~150 words per minute target. A 60-second video = ~150 words.
- Don't pad scripts to fit duration. The model paces itself; over-stuffed scripts get rushed delivery.
Prompt structure
The shapes these prompts actually take, rather than a description of one. For a complete worked set built on one of them, see hand-drawn VFX fusion.
Thin and single-sentence. One sound per prompt, described as material + action + recording character + an isolation clause, e.g. "short isolated sword impact, dry studio recording, no music".
Music prompts name mood, tempo, instrumentation and then `instrumental/no vocals`. Voice prompts contain the exact line to be spoken and nothing else, with the voice chosen by id rather than described.
Narration for assembled video is written as one plain labelled line per block (`Block 1` then the line), 20-24 words, about 8-9 seconds, with no timecodes, emotion cues, parentheticals or stage directions and numbers spelled out.
Negatives
Grouped by the failure they address. Keep the list short: on a model with no separate negative field, every one of these is just another word competing for attention.
Text and branding
- spoken words
- words
- other words
- harsh textures
- ad-libs that obscure the words
Content and safety
- scary sounds
Audio
- vocals
- lyrics
- sound effects
- speech
- music
- spoken narration
- voices
- melody
- singing
- ad-libs
- background music
- vocal-like sounds
- prominent lead melody
Other
- medical sounds
- abrupt peaks
- dark
- dramatic tension
- heavy drums
- long instrumental passages
- sung parts
- dialogue
- chanting
- tense passages
- harsh timbres
- abrupt stops
Take it with you
Take audio prompts with youThe same reference, as a printable sheet or a presentation.
Cheatsheet · 1 pageDeck · 18 pages