PROMPT RECORD图像记录

SEEDANCE 2.5 — 动作捕捉 / 角色替换提示词 创建一个 11 秒、4:3 的动作捕捉视频，要求口型精准同步。 核心目标 仅将 @vid1 中位于画面中央的王座水獭替换为我在 @i...

中文说明

SEEDANCE 2.5 — 动作捕捉 / 角色替换提示词 创建一个 11 秒、4:3 的动作捕捉视频，要求口型精准同步。 核心目标 仅将 @vid1 中位于画面中央的王座水獭替换为我在 @img2 中提供的水獭，从 @img1 已建立的整体构图、位置和外观出发。这只是单角色替换。所有背景角色、所有环境元素、建筑、王座大厅、观众、控制台、窗户、舞台、灯光、道具、镜头行为、原始镜头推进方式，以及 @vid1 的原始音频都必须保持完整且可明显识别。 参考优先级 1. @img1 = 起始帧与构图锁定。 使用 @img1 作为第 1 帧的直接视觉起点。新视频应从 @img1 就像精确的开场帧或近乎相同的画面开始。保留中央坐着的水獭位置、王座位置、房间布局以及整体取景。 2. @img2 = 唯一权威的角色身份参考。...

原始 Prompt

SEEDANCE 2.5 — MOTION CAPTURE / CHARACTER REPLACEMENT PROMPT

Create an 11 second 4:3 motion-capture video with accurate lip sync.

CORE GOAL
Replace ONLY the centre-frame throne otter in @vid1 with my otter from @img2, starting from the composition, placement and look already established in @img1. This is a single-character swap only. All background characters, all environmental elements, the architecture, the throne room, the audience, the consoles, the windows, the stage, the lighting, the props, the camera behaviour, the original shot progression, and the source audio from @vid1 must remain intact and recognisably the same.

REFERENCE PRIORITY
1. @img1 = starting frame and composition lock.
   Use @img1 as the immediate visual starting point for frame 1. The new video should begin as though @img1 is the exact opening frame or its near-identical equivalent. Preserve the centre placement of the seated otter, the throne position, the room layout, and the overall framing.

2. @img2 = sole authoritative character identity reference.
   Use @img2 as the complete and final authority for the replacement otter’s identity. Preserve his species, anatomy, facial design, proportions, silhouette, costume, accessories, colours, rendering language, and every persistent visual detail.
   He is an anthropomorphic otter with:
   - warm brown fur
   - cream muzzle and lighter chest/front markings
   - bright expressive eyes
   - happy, friendly, clearly readable face
   - a clear bubble helmet
   - a mounted teal eyepiece / visor unit
   - brass/gold-toned headset hardware with orange-tipped antenna
   - white and navy upper suit
   - orange cargo trousers
   - backpack / harness system
   - chunky white boot cuffs
   - a single clear tail
   - clear, clean 2D hand-drawn styling
   - fur-covered anthropomorphic hands with exactly 5 digits per hand

3. @vid1 = motion, audio, timing, camera and background source only.
   Extract only the motion performance, lip-sync timing, shot timing, camera behaviour, edit timing, environmental continuity, and audio timing from @vid1.
   Do NOT import, imitate, reconstruct, or blend any identity traits from the source centre character other than the movement itself.
   Do NOT change or replace any other character in the video.
   Do NOT alter background extras.

ABSOLUTE CHARACTER SWAP RULE
This must be a true single-subject replacement.
Only the existing throne-seated centre performer is replaced.
Everyone else visible in @vid1 remains the same character they already are.
No other cast member may be redesigned, restyled, swapped, simplified, upgraded, corrected, or replaced.

SCENE LOCK
Preserve the science-fiction throne-room setting from @vid1:
- interior command-room / audience-chamber environment
- centre throne / pod seat
- seated audience in foreground and sides
- small background characters near the stage
- rear wall details
- side structures and large blue dome-like form
- control desks and glowing interface shapes
- windows / openings and exterior sky tones
- the overall stage perspective and room symmetry

The environment should remain stable and familiar to the source video.
No new props.
No missing props.
No crowd changes.

SHOT / EDIT / CAMERA LOCK
Match @vid1’s shot structure and duration as closely as possible.
Preserve the same overall pacing and edit logic:
- opening wider establishing framing
- transition to the closer centred throne view at the same moment and in the same way as the source
- same camera angle family
- same lens feeling
- same framing emphasis
- same cut timing or push-in timing, whichever the source uses
- same duration from first frame to last frame

Do not invent extra shots.
Do not add extra camera moves.
Do not crop in unpredictably.
Do not drift off-centre.
Keep the replacement otter clearly readable from head to toe or as fully as the source framing permits throughout the whole video.

MOTION CAPTURE / PERFORMANCE TRANSFER
Transfer the source centre performer’s motion from @vid1 to my otter from @img2 precisely and rhythmically, while intelligently adapting the motion to his anthropomorphic otter anatomy and seated throne pose.

Performance requirements:
- replicate the source performer’s exact timing, rhythm, gesture emphasis, and speaking cadence
- preserve the same arm gestures and upper-body motion pattern
- preserve the same head movement pattern
- preserve the same torso bounce / body emphasis
- preserve the same seated performance energy
- preserve the same moments of emphasis toward the camera
- preserve the same conversational / presenter-like delivery to camera

The otter should feel like he is talking directly to the viewer from the throne.
He should maintain strong camera-facing readability and direct engagement.

LIP SYNC
Use the source audio from @vid1 and sync the otter’s mouth animation convincingly to it.
The mouth shapes should feel intentional and readable.
Match the syllabic timing closely.
The speech must feel clearly spoken to camera.
Include natural blinks and subtle facial animation, but do not let them interfere with the clarity of the lip sync.

ANATOMY ADAPTATION
Adapt the source motion cleanly into the otter’s established body plan.
This is an anthropomorphic otter, not a human in costume.
His movement must feel natural for his design:
- compact torso
- soft rounded muzzle
- seated lower body
- visible tail
- fur-covered anthropomorphic hands
- exactly 5 digits on each hand
- exactly 2 arms, 2 legs, 2 feet, 1 tail
- no duplicate limbs
- no broken wrists
- no extra fingers
- no missing fingers
- no warped paws
- no collapsing suit parts

COSTUME / ACCESSORY LOCK
Preserve the full space-suit identity from @img2:
- clear bubble helmet remains on throughout
- teal eyepiece / visor remains consistent in design and placement
- headset and orange-tipped antenna remain attached and stable
- backpack / harness remains consistent
- white, navy and orange costume colours remain consistent
- trouser shape remains consistent
- boots / ankle cuffs remain consistent
- insignia and suit details should stay recognisable
- no weapon
- no extra equipment beyond what belongs to this design

Helmet reflections and visor details should remain clean, simple and stable.
Do not let the helmet distort the face excessively.
The face must stay readable through the helmet.

STYLE / RENDERING
Fluid 2D hand-drawn animation.
Pure 2D look.
No 3D appearance.
No photorealism.
No puppety cutout look.
No messy interpolation look.

Visual style requirements:
- clean polished linework
- soft cel shading
- smooth colour application
- readable silhouette
- consistent proportions
- stable design from frame to frame
- gentle, appealing expression work
- clear mouth animation
- very clean final image
- cinematic but simple readability
- centre subject always easy to understand

MOVEMENT QUALITY
Movement should be fluid and confident, not jittery.
Keep the gesture clarity of the source performance.
Do not over-exaggerate beyond the source.
Do not underanimate the speech gestures.
Preserve subtle secondary motion where appropriate:
- slight head bob
- natural suit movement
- mild tail response if visible
- subtle ear and whisker response if helpful
- gentle body follow-through
But keep it controlled and readable.

BACKGROUND LOCK — VERY IMPORTANT
All non-centre characters from @vid1 must remain the same.
Do not swap them.
Do not redraw them as my otter.
Do not convert the room into a new style.
Do not redesign the crowd.
Do not alter their costumes, colours, species, poses, or placements except for whatever minimal motion already exists in the source video.

NEGATIVE / EXCLUSION RULES
- only 1 character swap
- do not replace any background character
- do not blend my otter with the source centre character’s identity
- no human face remnants
- no foreign costume elements from the source centre character
- no extra characters
- no duplicate otters
- no extra tails
- no missing tail
- no extra fingers
- no missing fingers
- no mutated anatomy
- no style drift
- no frame-to-frame off-model inconsistency
- no major helmet redesign
- no weapon
- no camera invention
- no new edits
- no scene redesign
- no background replacement
- no extra props
- no text overlays
- no subtitles unless already present in source
- no breaking the 4:3 frame

FINAL INTENT
The final video should feel like the exact original @vid1 scene and performance, except the throne-seated centre character has been cleanly and convincingly replaced with my own space-suited otter from @img2, beginning from the look and positioning established in @img1. He talks directly to the camera with accurate lip sync, preserves the source timing and gestures, and remains visually clear, charming, stable, and fully on-model for the entire 11 second clip.