Contenu du pack

Un fichier SKILL.md (le guide pratique) + un modèle de départ. Aperçu :

Title Scene Takeover

When to use

• You need a scene change that feels like an event, not a cut — a trailer beat, a chapter change, a before/after reveal
• One word can carry the transition: a place name, a date, a chapter title, a single dramatic verb
• The video's pace is fast enough that a 1.5–2.5 s full-screen title beat will energize rather than stall it
• You have a strong outgoing shot and a strong incoming shot, and the title can bridge them instead of sitting on either one
• You want a repeatable transition language: the same takeover move at every chapter change becomes the video's signature

Workflow

1. Pick the takeover word and the two shots (input: edit timeline → output: word + outgoing/incoming shots)
   Choose one word, 1–12 characters, that names what comes next — a city, a year, a chapter, a verdict. Long phrases cannot grow to fill a frame without turning into an unreadable blur, so keep it short. Pick an outgoing shot that ends on a stable frame (no fast camera whip mid-takeover) and an incoming shot that opens with movement or a reveal — the takeover works because the exit is calm and the entrance is alive. Mark the cut point, then back the takeover up so it starts 0.6–0.8 s before the cut: the word begins growing while the old scene is still visible, so the audience feels the takeover coming.
2. Grow the title to full frame (input: word + outgoing shot → output: title filling the frame)
   Place the word centered, in a heavy weight (800–900), starting at a size that reads comfortably over the outgoing shot — roughly 15–25% of frame height. Over 0.6–0.8 s, scale it with an ease-in curve from 100% to 900–1200% until a single letterform fills the frame edge to edge. Ease-in matters: the growth should accelerate, starting gentle and then rushing, like something taking over. As the letters grow past the frame edges, let them crop — do not shrink to keep the word visible. At full scale the frame should be one flat field of letterform and background: for maximum punch, the letters' fill color should be the outgoing shot's darkest tone or a solid brand color, so the frame reads as a deliberate title card for 2–4 frames, not as an accident. Hold that full-frame moment for 0.15–0.25 s — a beat, not a pause.
3. Cut inside the takeover (input: full-frame title → output: hidden cut)
   Make the actual scene cut while the title is at full scale, at the moment the frame is pure letterform. Because the frame is a flat field of color at that instant, the cut is invisible: the audience perceives one continuous takeover, not two shots. On the incoming side, start with the same full-frame title over the new shot — identical size, position, and color — and hold it for 2–4 frames so the eye registers continuity. This is the hinge of the whole effect: if the title jumps position, size, or color across the cut, the illusion breaks and the viewer sees a sloppy edit instead of a takeover.
4. Release into the new scene (input: full-frame title over new shot → output: title shrinking away to reveal the scene)
   Reverse the move: over 0.5–0.7 s, scale the title back down with an ease-out curve from full frame to its resting size (or to zero — shrinking to nothing and letting the scene take over completely is the cleaner exit for trailers). As the title shrinks, the new scene is revealed from behind it, like a curtain pulling back. Time the reveal so the new scene's key subject is uncovered first — if the incoming shot is a person, the title should shrink toward a point away from the face, or shrink to zero before the face matters. Add a subtle 0.2 s cross-dissolve on the title layer only (not the video) at the very end of the shrink, so the last trace of the word melts rather than pops off.
5. Add sound and momentum (input: finished visual takeover → output: takeover with audio punch)
   The takeover lives or dies on its sound. Place a rising whoosh or swell under the growth phase (0.6–0.8 s, rising pitch or volume), a short impact hit — a thud, a clap, a synth stab — exactly on the full-frame hold, and a breath of room tone or the new scene's natural sound as the title releases. If the incoming scene has music, duck the music 3–4 dB during the growth and bring it back on the release so the takeover feels like it owns the beat. Keep the whole takeover between 1.5 and 2.5 s end to end. Shorter feels like a glitch; longer feels like an interruption. Use the identical takeover timing at every chapter change in the same video — consistency is what turns a trick into a language.

Aperçu du modèle

Title Scene Takeover — Build Template

One word, one takeover, one hidden cut. Fill this in before animating.

Takeover word: {WORD — 1–12 characters}

Why this word: {names what comes next: place / date / chapter / verdict} Font: {heavy weight 800–900} | Title color: {solid — identical on both sides of the cut}

The two shots

Takeover starts: {0.6–0.8 s before the cut}

Growth phase (outgoing shot)

The hidden cut

Cut placed: {inside the full-frame hold, frame is pure letterform} Continuity check: [ ] identical size, position, color on both sides for 2–4 frames

Release phase (incoming shot)

Le pack de compétence lui-même (SKILL.md + modèle) est rédigé en anglais.

Acheter le pack complet — $4.99 Pro — les 511 compétences