tenglonghaiwen-hub/zoraagent/seedance-lighting-presets

Seedance Lighting Presets

Use this skill to choose, compress, and insert a lighting setup into a Seedance 2.0 prompt. It is designed for portrait, short-drama, product/person showcase, ancient-style, urban-night, stage, suspense, and emotional close-up shots.

Workflow

1. Identify the scene intention: tenderness, daily realism, romance, mystery, danger, ritual, isolation, commercial cleanness, conflict, or dreamlike calm.
2. Choose one primary lighting preset from references/lighting-presets.md. Do not combine more than two presets unless the user asks for a stylized conflict look.
3. Convert the preset into a compact lighting module. Preserve physical details: light source, direction, subject areas hit by light, fill strength, eye catchlight, background treatment, and haze/reflection if useful.
4. Merge the module with the full Seedance prompt: subject, action, camera move, lighting, sound, and constraints.
5. If references are present, assign each reference one primary role: identity, first frame, last frame, product, environment, motion, camera, timing, audio, or style. State what must not transfer.

Prompt Module Rules

Prefer concrete light behavior over generic mood words.

Keep:

• source: window, sky, cloud cover, sunset, foggy morning sun, moon, candle, lantern, fire, neon sign, streetlight, headlights, stage follow spot, hard spotlight, backlight, softbox, side light, reflected surface, water reflection, split color lamps
• direction: side front, overhead, side overhead, side back, back, low back, lower front, side lower, front upper
• visible effect: cheek, eyes, nose bridge, lips, hands, collar, hair edge, shoulder rim, clothing folds, wet ground reflection, catchlight
• control: weak fill, no dead-black shadows, no overexposure, preserve facial structure and skin texture

Remove:

• repeated "cinematic", "high quality", "beautiful", "realistic" if the physical light already expresses it
• long emotional explanations when the scene action already carries the mood
• secondary light effects that compete with identity, action, or camera

Native Chinese Output

When the user asks in Chinese or wants a Chinese prompt, output Chinese. Use compact prompt wording:

灯光：以[光源]为主光，从[方向]照向人物，[受光部位]形成[效果]；正面仅弱补光，保留五官结构、皮肤纹理和眼神光；背景[虚化/压暗/反光/雾气]。

For Japanese or Korean delivery, keep the reference tags unchanged and translate only the natural-language lighting module.

Seedance Integration Pattern

Standalone T2V:

[Subject] [visible action and endpoint] in [scene]. Camera: [one move]. Lighting: [compressed preset module]. Sound: [ambience/dialogue/SFX]. Constraint: [identity/IP/continuity].

I2V:

[Image1] preserves identity, wardrobe, and face structure; only [motion/camera/light] changes. Lighting: [compressed preset module]. Constraint: do not alter face, clothing, or background layout unless requested.

Reference-to-video:

[ReferenceTag] controls [one role] only. Do not transfer [identity/logo/background/style] unless assigned. Lighting: [compressed preset module].

Delivery Checks

Before finalizing:

• One lighting preset is dominant.
• The light source is physically motivated and visible or inferable.
• Camera, performance, and light serve one intention.
• Shadows are controlled: no dead-black face unless requested.
• Eye catchlight is specified for portrait/emotion shots.
• Prompt remains compact enough that lighting does not crowd out subject, action, camera, and endpoint.
• Chinese/Japanese/Korean output preserves exact reference tags.

Reference

Load references/lighting-presets.md when choosing a preset, comparing styles, or building a reusable lighting module.