Templates C, E and F

Three fill-in prompts for speaking and singing shots, written for this guide. Fill every slot, delete the blocks a shot doesn't need, then validate (see Video with Seedance → Validate, then pre-flight). The block order is the one taught in Video with Seedance → The prompt grammar.

Worked prompts:

• Template C: Video with Seedance → 8. Dialogue close-up and Video with Seedance → 9. Presenter line;

• Template E: Video with Seedance → 12. Arabic audio-referenced close-up and Video with Seedance → 13. English audio-referenced line;

• Template F: Video with Seedance → 14. Egyptian pop chorus.

Template C: the sound-off speaking plate

Settings:

• omni-reference, audio off;

• duration = onset + spoken span + tail, rounded up to whole seconds;

• images: the character sheet, the identity close-up, the location plate;

• no start still;

• draft, then the 1080p finish, then lip sync on the finish.

The beat times come from the approved take's word timestamps.

SCENE CONTEXT: {who speaks to whom, where and when; the partner off screen or not speaking}.
ACTIVE REFERENCES: @image1 is {ALIAS}: {the anchor line, with nothing about the voice}. This sheet gives identity and wardrobe and nothing else: the face and hair come from the close-up panel alone, the wardrobe and body from the full-body panels alone; the front panel's cropped head is never repeated; leave out its backdrop, layout and flat light. Match it exactly. @image2 is {ALIAS}'S FACE: the identity close-up, used for the face alone. @image3 is {PLACE}: its architecture, materials and light, nothing else; it does not decide where the camera stands.
LOCATION MAP: {layers from the lens, each at its distance}. Live motion: {as the plan counts it: 2–4 outdoors or wherever the place has them, 0–1 in a quiet room; kept clear of the face; each with its cause, direction and rate}.
FIRST FRAME AND BLOCKING (in motion from frame one): Medium close-up {or closer}. {ALIAS} is {position, facing}, lips together, already {a physical task}; hands {out of frame | resting still on …}. EYELINE — {ALIAS}: looks {frame-left | frame-right} of the lens, just past it, at {the partner's near eye}, level, and stays on it through the line.
FORMAT MODE: A single unbroken shot with no cuts of its own. Real time.
OPTICS: {°} diagonal field of view; the mouth clear; the lens holds steady.
CAMERA: {rig} at {height} m, {distance} m from {ALIAS}. While the words run, {one very slow push of {cm} over {n} s, about a sixth of the distance, seen as the face growing by about a fifth | a handheld frame that sways in place with the operator's breath}, from {start composition} to {landing composition}, {its job}. {Any bigger move: before the first word or after the last.} The face keeps within about 45° of the lens, and nothing reframes fast while a word is spoken.
ACTION / TIMELINE: 0–{onset} s lips together, {the beat before the line}; {a}–{b} s {ALIAS} speaks, {manner}; {b}–{c} s a pause, lips together, {behaviour}; {c}–{d} s {ALIAS} speaks again, {manner}; {d}–{duration} s lips together, {the hold after the line}.
{Only where the plan names the option for a hero line that plays flat: DELIVERY MAP (direction only, never spoken): {phrase by phrase at the word timestamps, one physical cause each, with its time}.}
PERFORMANCE — {ALIAS}: {the task, keyed to what the line means, written as behaviour; the line itself is never quoted}. {A blink in a pause, where the thought turns}. {One involuntary detail}. {Emotion}, never {its overshoot}. {ALIAS} talks in a natural, {manner} conversational voice; the mouth moves naturally through every stretch of speech, at ordinary speaking size, with no exaggerated shapes; the lips come to rest, closed, in every pause, before the first word and after the last.
TEXTURE: {2–4 materials that show wear}; the frame sits off-centre.
LIGHTING: {one setup, in Kelvin}; the mouth lit by {the key | in low key, a motivated source in frame}.
STYLE: {the film's look line, word for word}.
POSITIVE LOCKS: Exactly one person speaks. {Anyone else}: speaks no words in this beat. No subtitles and no on-screen text.

Template E: the audio-referenced shot

For Route 1's audio-driven plate (also the first-plates check's fallback), the Franco quote, and English principal lines. For Arabic, use it only for a close-up hero line you chose for Route 1, or for shots you moved to it after the first-plates check failed twice; the Arabic default is Template C.

Settings:

• omni-reference, audio on;

• images in this order: the character sheet, the identity close-up, the location plate;

• then the audio reference: the approved take, voice only, at least 2 s (a shorter line uses the copy padded with room tone);

• duration = start + speech span + tail, in whole seconds, at least 4 s;

• draft, then the 1080p finish, then lip sync on the finish where the route says so.

If Draft Mode refuses the audio reference, draft at 720p.

SCENE CONTEXT: {who speaks to whom, where and when; the partner off screen or not speaking}.
ACTIVE REFERENCES: {@image1 to @image3 as in Template C}. @audio1 is {ALIAS}'S VOICE: the approved recording; it gives the voice, accent, pace and pauses, and nothing in the picture.
VOICE AND TIMING: {ALIAS}, voice: {the voice in a few words}, as in @audio1. The speech in @audio1 runs {span, to 0.1 s} s; {she | he} speaks from {start} s to {end} s.
FIRST FRAME AND BLOCKING (in motion from frame one): {size}. {ALIAS} is {position, facing}, lips together, already {a physical task}; hands {out of frame | resting still on …}. EYELINE — {ALIAS}: {side of the lens, target, height}.
FORMAT MODE: A single unbroken shot with no cuts of its own. Real time.
OPTICS: {°} diagonal field of view; the mouth clear; the lens holds steady.
CAMERA: {as in Template C: while the words run, a breathing frame or a very slow push, seen as the face growing by about a fifth; any bigger move only before or after the line}. The face keeps within about 45° of the lens, and nothing reframes fast while a word is spoken.
ACTION / TIMELINE: 0–{start} s lips together, {the beat before the line}; {start}–{end} s {ALIAS} speaks: {behaviour beats from the take's timestamps: the tactic, where the eyes go, one involuntary detail; nothing about the mouth}; {end}–{duration} s lips together, {the hold after the line}.
{Only where the plan names the option for a hero line that plays flat: DELIVERY MAP (direction only, never spoken): {phrase by phrase, one physical cause each, with its time}.}
PERFORMANCE — {ALIAS}: {the task, the eye life, the emotion clamped on both sides}. {ALIAS} says, {manner}: "{the line}"
Only {ALIAS} speaks; {the listener} speaks no words in this beat: {listening, with a single visible change as the key word arrives}. Nothing outside the quotation marks is spoken; the rest of the text is direction only.
TEXTURE: {2–4 materials that show wear}; the frame sits off-centre.
LIGHTING: {source, Kelvin}; the mouth lit by {the key | a motivated source in frame}.
AUDIO: {Her | His} voice, accent, pace and pauses come from @audio1; close, dry dialogue; {one or two real sounds of the shot}, kept low. No music, no subtitles and no on-screen text.
STYLE: {the film's look line, word for word}.
POSITIVE LOCKS: Exactly one person speaks. {Anyone else}: speaks no words in this beat.

• Seedance re-times an audio reference, so the timeline's windows are only an order: measure the draft's real onset and span against the take before you judge it.

The own-voice form (decided), for Seedance's own voice in English or another language it speaks well, never Arabic: Template E as written, with no audio reference and no @audio1. The quoted line is written exactly as it is to be spoken, in its own alphabet. Two blocks change:

VOICE AND TIMING: {ALIAS}, voice: {the character's fixed voice description, exactly as the client approved it from the sample, the same words in every shot: age, timbre, pitch, pace, accent, texture; no person's name}.
Emotion: {this line's emotion and delivery}.
{She | He} says the line once, exactly as written, from {start} s to {end} s, with about {0.5–1} s of quiet before and after; the lips stay still whenever {she | he} is not speaking.
AUDIO: {Her | His} voice as described; close {room} dialogue; {one or two real sounds}, kept low. Only the quoted words are spoken. No music, no subtitles and no on-screen text.

Template F: the singing plate

For a sung line on screen: a music video, a jingle, a sung voiceover played to camera (see Speaking shots and lip sync → Other speaking routes).

Settings, as Template C:

• omni-reference, audio off;

• images: the character sheet, the smiling close-up if one was made for this character (otherwise the identity close-up), the location plate;

• no start still;

• duration = head + the vocal segment + tail, in whole seconds;

• draft, then the 1080p finish, then lip sync on the finish.

SCENE CONTEXT: {who sings, where, and to whom or to the lens; the song's section (verse, chorus, hook) and its energy}.
ACTIVE REFERENCES: @image1 is {ALIAS}: {the anchor line, with nothing about the voice}. {The sheet's role line, as in Template C.} @image2 is {ALIAS}'S FACE: {the smiling close-up, used for the face, teeth and jaw alone | the identity close-up, used for the face alone}. @image3 is {PLACE}: its architecture, materials and light, nothing else; it does not decide where the camera stands.
LOCATION MAP: {layers from the lens}. Live motion: {background life that moves on the beat: lights, a crowd, fabric, water}.
FIRST FRAME AND BLOCKING (in motion from frame one): {size}. {ALIAS} is {position, facing}, lips together, already moving to the tempo: {a sway, a step, a nod on the beat}.
FORMAT MODE: A single unbroken shot with no cuts of its own. Real time.
OPTICS: {°} diagonal field of view; the mouth clear; the lens holds steady.
CAMERA: {one dosed move timed to the bars, such as an arc of {°} across the {n} bars of the phrase}, from {start composition} to {landing composition}, {its job}. While {she | he} sings, the face keeps within about 45° of the lens; any bigger move waits for the instrumental gap.
ACTION / TIMELINE: 0–{onset} s lips together, the breath before the phrase; {a}–{b} s {ALIAS} sings {phrase 1, as behaviour}, the mouth wide on the held open vowels at {t1} s and {t2} s; {b}–{c} s a breath, lips apart; {c}–{d} s {phrase 2}; {d}–{duration} s {the hold}.
{Only where the plan names the option for a hero line that plays flat: DELIVERY MAP (direction only, never sung): {phrase by phrase, one physical cause each, such as the chin rising into the long note, or the eyes closing on the held vowel}.}
PERFORMANCE — {ALIAS}: {what the song means to the character now, as a task aimed at someone}. The body moves at {BPM} beats a minute, its weight dropping on {the beat}. {Eye life}. {She | He} sings in full voice; the mouth moves naturally through every sung stretch, open wide on the held vowels, at real singing size, with no exaggerated shapes; between phrases the lips rest closed or just apart.
TEXTURE: {2–4 materials that show wear}; the frame sits off-centre; {the background falling into dark beyond {n} m}.
LIGHTING: {source, Kelvin}; the mouth lit by {the key | a motivated source in frame}.
STYLE: {the film's look line, word for word}.
POSITIVE LOCKS: Exactly one person sings. {Anyone else}: neither sings nor mouths words. No subtitles and no on-screen text.

• Held notes and phrase lengths come from the stem, never from the lyric sheet, because singers stretch words. Write each held vowel's time, the mouth open and holding its shape for that length, and measure the draft's onset against the stem again before you judge it.

• A chorus repeats, so each repeat is its own plate, from a different angle and size, and the cut can move on the bar lines.

Sources

• Higgsfield's prompt builder for Seedance 2.5 (the block order); JOEY, @joeysnoisygroup (project) (listeners who say nothing); Welcome to the Team (@chip_frames; project) (the delivery map; only quoted words are spoken); Soulless by Yoav Yariv and The Elevator by @curious_media_2026 (one fixed voice description with a separate emotion line).