skills/ai-film-pipeline-master/references/phase-07-image-prompt-engineering/music-video.md

# Music Video and Performance

14 cards. These decide a frame: what is in it, how it is lit, and what it is made of.

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.

---

### Lip-Sync Framing

**Also called:** sync coverage, mouth framing  
**What it is:** Framing designed so the mouth and the eyes read clearly: a medium close-up or closer, with the face lit from a side that defines the lips.  
**Effect on the audience:** It proves this is the owner of the voice, and the closeness lets the viewer read the emotion in the eyes alongside the word.  
**Used for and where it works best:** Reserve the tightest size for the most important line in the song. Put a small catchlight in the eye so the face does not die. In vertical music video keep the mouth in the upper third, since the lower band is taken by text and interface.  
**Best in:** formats: Music Video, Lyric Video, Vertical Short-Form (Reels, Shorts, TikTok), Concert / Live Event, Kids Content, Dance Film | genres: Musical, Romance, Drama, Melodrama, Family, Comedy  
**Avoid when:** A wide shot on the key line wastes the most important moment in the song. A hard shadow across the mouth makes the sync look wrong.  
**Prompt:** `medium close-up on the singer, mouth and eyes clearly lit, catchlight in the eye, soft key from three quarters`  
**Yields to:** `mv.hidden_face` — The face must not be readable.  
`mv.lipsync_framing`

### Crowd-Energy Wide

**Also called:** group wide, crowd shot  
**What it is:** A wide shot placing the singer inside a large group moving with them: an audience, dancers, the neighbourhood.  
**Effect on the audience:** It makes the song feel bigger than one person, and the collective energy transfers straight to the viewer.  
**Used for and where it works best:** Reserve it for the final chorus so it becomes the peak. Keep the group action simple and synchronised: a jump, raised hands, one step. In vertical shoot it from above or from low so it fills the height of the frame.  
**Best in:** formats: Music Video, Concert / Live Event, Dance Film, Vertical Short-Form (Reels, Shorts, TikTok), Sports Broadcast / Sports Film, Kids Content | genres: Musical, Family, Sports Drama, Comedy, Slice of Life, Romance  
**Avoid when:** An unsynchronised crowd reads as chaos rather than energy. Do not spend it on the first chorus; you leave yourself no escalation for the end.  
**Prompt:** `wide low angle of the singer surrounded by a crowd raising their hands together on the downbeat of the final chorus`  
**Yields to:** `mv.handheld_party` — Crowd cannot be synchronised; let the chaos be the point.  
`mv.crowd_energy_wide`

### Strobe On The Drop

**Also called:** strobe lighting, flicker light  
**What it is:** A strobe light inside the scene itself, so movement breaks into a series of frozen images instead of continuous motion. How fast it flashes is not this card's to set - see `Avoid when:`.  
**Effect on the audience:** It gives club energy and a sense of explosion, and makes movement look stronger because each pose is caught on its own.  
**Used for and where it works best:** Reserve it for the drop or the final chorus so it registers as an event. Match the flash frequency to the shutter so you do not catch half frames. In dance let it catch the biggest moves.  
**Best in:** formats: Music Video, Concert / Live Event, Dance Film, Vertical Short-Form (Reels, Shorts, TikTok), Trailer / Teaser, Fashion Film | genres: Musical, Action, Cyberpunk, Horror, Thriller / Suspense, Sports Drama  
**Avoid when:** A long strobe tires the eye, and on a calm song it is simply wrong. Whether it can injure a photosensitive viewer is not a length this card gets to choose — see `safety.photosensitive_flash`, which owns the rule and where it stops.  
**Prompt:** `strobe lighting during the drop catching the dancer in frozen poses, shutter matched to the flash rate`  
*(the rate is deliberately not specified here. This field used to open with "high frequency", which is a rate — the exact thing the Avoid when above says this card does not get to choose. It goes to `safety.photosensitive_flash`.)*  
**Yields to:** `safety.photosensitive_flash` — AUTHORITY: rate and stopping point are not this card's.  
`mv.strobe_drop`

### Lyrics On Screen

**Also called:** on screen lyrics, sung text  
**What it is:** The song's words are printed over the image in sync with the singing, word by word or line by line.  
**Effect on the audience:** They make the song understandable even to someone who does not know the language or is watching muted, and add a design layer to the frame.  
**Used for and where it works best:** Essential in vertical and for songs in another language. Let the typeface serve the mood of the song and keep its position fixed. In Arabic keep it to one centred line. For emphasis, enlarge only the key word.  
**Best in:** formats: Music Video, Lyric Video, Vertical Short-Form (Reels, Shorts, TikTok), Kids Content, Concert / Live Event, Commercial / TVC | genres: Musical, Romance, Family, Comedy, Drama, Faith / Biblical Epic  
**Avoid when:** Small type over a busy image is not read. Heavy animation on every word pulls attention off the performance.  
**Prompt:** `burned lyrics appearing line by line in sync with the vocal, fixed position in the safe band, typeface matching the mood`  
`mv.lyrics_on_screen`

### Stage Lighting Cues

**Also called:** lighting changes, programmed light  
**What it is:** The lighting inside the scene changes with the structure of the song: a colour enters on the chorus, a backlight ignites on the drop, darkness in the bridge.  
**Effect on the audience:** It makes the song visible in the light itself and gives the video the feel of a live show even when it was shot in a studio.  
**Used for and where it works best:** Programme the changes before the shoot against the song's timecode and shoot every take to the same programme so the edit stays coherent. Use two contrasting colours and no more. Ideal for night videos and concert work.  
**Best in:** formats: Music Video, Concert / Live Event, Dance Film, Vertical Short-Form (Reels, Shorts, TikTok), Title Sequence / Opening Credits, Trailer / Teaser | genres: Musical, Action, Cyberpunk, Romance, Sports Drama, Drama  
**Avoid when:** Random lighting changes between takes make the edit impossible. Three or more colours cheapen the image.  
**Prompt:** `programmed stage lighting, cool key in the verse, a red backlight igniting on the chorus, blackout on the bridge`  
**Yields to:** `mv.practical_lights` — No programmable rig; take colour from visible sources.  
`mv.stage_light_cues`

### Mirror / Kaleidoscope

**Also called:** mirrored frame, symmetry effect  
**What it is:** The image is mirrored horizontally or vertically or repeated in symmetrical arrays, turning the performer into a geometric pattern. It differs from the Mirror Shot, which films a real reflection in a mirror inside the scene: here the mirroring is applied to the picture itself in the edit.  
**Effect on the audience:** It turns the real into abstraction and produces a hypnotic, repeating feel that suits electronic and dreamlike sections.  
**Used for and where it works best:** Keep it for the bridge or an instrumental section rather than the whole song. Build it from a shot whose movement is already symmetrical, such as raised arms. In heritage music video it works well with repeating ornament and carved pattern.  
**Best in:** formats: Music Video, Experimental / Art Film, Lyric Video, Vertical Short-Form (Reels, Shorts, TikTok), Title Sequence / Opening Credits, AI-Generated Clip | genres: Musical, Psychological / Arthouse, Surreal / Absurdist, Fantasy / Epic, Cyberpunk, Heritage (civilisation-focused)  
**Avoid when:** Using it through the whole video loses the performer and bores the viewer. If the movement is not symmetrical the mirror comes out distorted and unpleasant.  
**Prompt:** `vertically mirrored frame of the performer raising both arms, kaleidoscope repetition building through the bridge`  
**Example:** Queen, Bohemian Rhapsody (1975), the mirrored multiplied faces of the four members.  
**Yields to:** `frame.mirror` — You want a real reflection in the location, not applied symmetry.  
`mv.mirror_kaleidoscope`

### Hidden-Face Performance

**Also called:** silhouette performance, back of head sync  
**What it is:** The performer is shot in silhouette, from behind, or against a strong backlight, so the shape reads but the features do not.  
**Effect on the audience:** It makes the voice the protagonist and creates mystery, letting the viewer occupy the figure rather than watch a specific person.  
**Used for and where it works best:** The practical answer when you do not want a face on screen or have no performer: a back walking through window light, a silhouette against a lit doorway. Essential in heritage content that avoids depicting faces. In vertical it pairs well with a mystery hook.  
**Best in:** formats: Music Video, Lyric Video, Heritage / Historical Documentary, Vertical Short-Form (Reels, Shorts, TikTok), Experimental / Art Film, Short Film, AI-Generated Clip | genres: Musical, Heritage (civilisation-focused), Drama, Mystery / Whodunit, Psychological / Arthouse, Faith / Biblical Epic, Romance  
**Avoid when:** A whole video with a hidden face leaves the viewer with nobody to attach to. If the song is about a specific person, hiding them defeats the purpose.  
**Prompt:** `performer in full silhouette against a bright doorway, features unreadable, slow walk toward the light`  
**Yields to:** `mv.lipsync_framing` — The song is about a specific person; hiding them defeats it.  
`mv.hidden_face`

### Instrument Close-Up On Its Note

**Also called:** instrument insert, played note shot  
**What it is:** A close shot of the instrument at the exact instant its sound is heard: fingers on the strings, the stick on the drum, breath into the reed.  
**Effect on the audience:** It ties a sound to its source so the viewer believes the music is being played in front of them, and it communicates craft.  
**Used for and where it works best:** Distribute them according to the arrangement: the drum shot on the rhythm entry, the string shot on the solo. Shoot very close with side light so the texture reads. It matters most in heritage music video, where it introduces the traditional instruments.  
**Best in:** formats: Music Video, Concert / Live Event, Heritage / Historical Documentary, Vertical Short-Form (Reels, Shorts, TikTok), Lyric Video, Kids Content | genres: Musical, Heritage (civilisation-focused), Historical / Biopic, Family, Drama, Faith / Biblical Epic  
**Avoid when:** A shot of an instrument that is not heard in that section reads as fake. Fingers playing something else are spotted instantly by any musician.  
**Prompt:** `extreme close-up of fingers striking the strings timed exactly to the note heard, hard side light, shallow focus`  
`mv.instrument_closeup`

### The Bridge As A Different World

**Also called:** bridge break, world shift  
**What it is:** The bridge section is shot in a different place or a completely different style: another palette, another lens, or animation instead of live action.  
**Effect on the audience:** It gives a visual rest and makes the return to the final chorus feel like a homecoming, which strengthens the peak.  
**Used for and where it works best:** Plan it as a miniature world: one place, one colour, one idea. Ideal for concept shots, dreams and memory. In heritage music video make the bridge the past: the same place in another era.  
**Best in:** formats: Music Video, Lyric Video, Vertical Short-Form (Reels, Shorts, TikTok), Experimental / Art Film, Short Film, AI-Generated Clip | genres: Musical, Psychological / Arthouse, Surreal / Absurdist, Romance, Heritage (civilisation-focused), Fantasy / Epic, Drama  
**Avoid when:** A bridge that looks like the rest of the video wastes the best chance for variation. An over-complicated world there distracts and breaks the flow.  
**Prompt:** `the bridge staged in a single different world, one dominant colour and one idea, returning to the main world on the last chorus`  
**Yields to:** `mv.mirror_kaleidoscope` — No second world available; vary the bridge in the edit.  
`mv.bridge_other_world`

### Projection On The Performer

**Also called:** projected visuals, light projection  
**What it is:** Images, video or patterns are thrown by a projector onto the performer's body and the wall behind, becoming part of the lighting.  
**Effect on the audience:** It merges two visual layers in one shot with no compositing and makes the person feel inside the image rather than in front of it.  
**Used for and where it works best:** Ideal for heritage music video: project a relief, a map or ancient script onto the face and the clothing. Shoot in a dark room with the performer close to the wall. Keep the projected image slow so the face stays readable.  
**Best in:** formats: Music Video, Concert / Live Event, Lyric Video, Experimental / Art Film, Heritage / Historical Documentary, Vertical Short-Form (Reels, Shorts, TikTok), Title Sequence / Opening Credits | genres: Musical, Heritage (civilisation-focused), Historical / Biopic, Psychological / Arthouse, Cyberpunk, Faith / Biblical Epic, Drama  
**Avoid when:** A fast or over-bright projection erases the features completely. If the room is not dark the projection reads washed out and does nothing. Keeping it slow is a legibility rule and this card owns that one; where a bright projection in a dark room becomes a fast alternation instead of a moving image, how fast it may go and when it must stop are `safety.photosensitive_flash`.  
**Prompt:** `carved relief patterns projected across the performer's face and the wall behind, dark room, slow moving projection`  
**Yields to:** `mv.stage_light_cues` — Room cannot be darkened; colour the performer with light.  
`mv.projection`

### Practical Lights In Frame

**Also called:** in shot practicals, visible sources  
**What it is:** The light sources themselves are visible in frame: neon, hanging bulbs, candles, car lights, screens.  
**Effect on the audience:** They provide depth, colour and a real sense of place, and act as eye catchers that stop the frame feeling empty.  
**Used for and where it works best:** The backbone of low budget night videos: place bulbs or neon behind the performer out of focus so they bloom into circles. Use two contrasting colours. In heritage music video use candles and torches as the only source.  
**Best in:** formats: Music Video, Vertical Short-Form (Reels, Shorts, TikTok), Concert / Live Event, Lyric Video, Wedding / Event Film, Fashion Film, Heritage / Historical Documentary | genres: Musical, Romance, Cyberpunk, Noir / Neo-Noir, Heritage (civilisation-focused), Drama, Slice of Life  
**Avoid when:** Many sources in mixed colours make the frame busy and cheap. If a source is brighter than the face the eye goes to it instead of the performer.  
**Prompt:** `hanging bulbs and neon visible behind the singer thrown out of focus into glowing circles, two colour palette, night`  
**Yields to:** `mv.stage_light_cues` — Sources overpower the face; control colour off-camera.  
`mv.practical_lights`

### Performance To Narrative Match

**Also called:** world bridge cut, match between layers  
**What it is:** A cut between the performance world and the story world on two shots that match in movement or composition: the same head turn, the same hand, the same direction.  
**Effect on the audience:** It bonds the two worlds and makes the viewer feel the singer is the character in the story, or is watching it.  
**Used for and where it works best:** This is the cleanest way to move between performance and story instead of cutting arbitrarily. Design both shots together at the shoot: same angle, same size. Place the match at the end of the verse before the chorus.  
**Best in:** formats: Music Video, Lyric Video, Short Film, Vertical Short-Form (Reels, Shorts, TikTok), Trailer / Teaser | genres: Musical, Romance, Drama, Melodrama, Tragedy, Crime / Procedural  
**Avoid when:** A forced match between two unrelated shots reads as a gimmick. Jumping between the worlds too often means neither one gets built.  
**Prompt:** `the singer turns his head left and the cut lands on the character turning left in the story, matched size and angle`  
**Yields to:** `mv.whip_pan_beat` — No matching shots; join the worlds on a whip.  
`mv.perf_narrative_match`

### Literal Lyric Image

**Also called:** lyric visual pun, word to image  
**What it is:** A shot that executes the literal meaning of a word at the instant it is sung: the word fire with a flame, the word road with a road.  
**Effect on the audience:** It gives an instant matching pleasure and helps the words stick, and it is the backbone of the lyric video form.  
**Used for and where it works best:** Pick three or four key words in the song and build images for them, not every word. Best for simple songs, children's songs and lyric video. In Arabic choose the word with a concrete image rather than a metaphor.  
**Best in:** formats: Lyric Video, Music Video, Kids Content, Vertical Short-Form (Reels, Shorts, TikTok), Explainer, AI-Generated Clip | genres: Musical, Family, Preschool / Early Learning, Comedy, Romance, Faith / Biblical Epic  
**Avoid when:** Executing every word literally makes the video naive and tedious. In serious songs, literalism kills the poetry and shrinks the meaning.  
**Prompt:** `the image cuts to a flame exactly on the sung word for fire, one literal image per key word only`  
**Yields to:** `mv.concept_shot` — Serious song; literalism kills the poetry.  
`mv.lyric_visual_pun`

### Flash Frames

**Also called:** white frames, single frame flash  
**What it is:** One or two white or single colour frames placed on the beat between shots, at the shortest possible duration.  
**Effect on the audience:** It gives a shock of light and sharpens the cut, and it is felt more than seen because it is shorter than conscious perception.  
**Used for and where it works best:** Use it at the drop and in the fast montage: one white frame immediately before the new shot. It differs from a strobe in that it is an edit device rather than set lighting. Keep it for the big accents only - and read that as a rule about attention, not as the answer to the photosensitivity question, which the `Avoid when:` line below hands to the card that owns it.  
**Best in:** formats: Music Video, Trailer / Teaser, Vertical Short-Form (Reels, Shorts, TikTok), Dance Film, Meme / Comedy Edit, Title Sequence / Opening Credits | genres: Musical, Action, Thriller / Suspense, Horror, Cyberpunk, Sports Drama  
**Avoid when:** It has no place in calm or children's content. Repeated flash frames are a photosensitivity question rather than a taste one, and *keep it rare* is not the answer to it — `safety.photosensitive_flash` carries the rule and says who decides.  
**Prompt:** `single white frame inserted immediately before the cut on an accent of the drop`  
*(how many accents get one is deliberately not specified here. This field used to say "each cut on the four big accents" — a repetition, which is the photosensitivity question the Avoid when above hands to `safety.photosensitive_flash`, not a taste one.)*  
**Yields to:** `safety.photosensitive_flash` — AUTHORITY: repetition is a photosensitivity question.  
`mv.flash_frames`