sora-higgsfield-prompting

Original source / Raw SKILL.md

---
name: sora-higgsfield-prompting
description: AI video generation prompting guide for Sora 2 and Higgsfield.ai
version: 1.0
category: content
activation:
  mode: AUTO
  triggers:
    - "AI video generation"
    - "Sora prompt"
    - "Sora 2"
    - "Higgsfield"
    - "text to video"
    - "video from text"
    - "AI video script"
    - "cinematic shot"
    - "camera movement"
    - "video prompt template"
    - "character consistency AI video"
    - "multi-shot sequence"
related_skills:
  - video-production-guidelines
  - marketing-content-guidelines
references:
  - reference/PROGRESSIVE_EXAMPLES.md
  - reference/TECHNICAL_REFERENCE.md
  - reference/ADVANCED_WORKFLOWS.md
---

# Sora & Higgsfield AI Video Generation

## When This Skill Activates

This skill auto-activates when you're working with:
- **Sora 2** or **Higgsfield.ai** AI video generation
- Writing prompts for text-to-video AI models
- Creating multi-shot video sequences with AI
- Troubleshooting AI video generation issues
- Requesting cinematic shot types or camera movements

## What This Skill Provides

- **5 Core Prompting Principles** - Foundational guidelines for effective prompts
- **Quick Start Templates** - Copy-paste structures for common use cases
- **Common Mistakes & Fixes** - Troubleshooting guide
- **Quality Checklist** - Pre-generation validation

**For Advanced Topics:**
- [Progressive Examples](reference/PROGRESSIVE_EXAMPLES.md) - Beginner → Advanced
- [Technical Reference](reference/TECHNICAL_REFERENCE.md) - API specs, camera controls
- [Advanced Workflows](reference/ADVANCED_WORKFLOWS.md) - Multi-shot production

---

# Quick Start Guide

## What You're Working With

**OpenAI Sora 2** creates videos from your text descriptions or images. Think of it like hiring a filmmaker—the clearer your instructions, the better your video turns out.

**Higgsfield.ai** brings together multiple AI video tools (Sora 2, Google Veo, WAN, Kling, and more) in one place. It gives you easy camera controls and editing tools to make professional-looking videos.

## Your First Prompt

Start with this simple structure:

```
[Style]. [Subject] [action]. [Camera movement]. [Environment]. [Lighting].
```

### Example 1: Your First Video

**Prompt:**
```
Cinematic style. A golden retriever runs across a sunny meadow.
Camera pans left to right. Bright afternoon sunlight, green grass, blue sky.
```

**Why this works:**
- **Style** ("Cinematic") sets the visual aesthetic immediately
- **Subject + action** is clear and simple (one dog, one movement)
- **Camera movement** is specified and straightforward
- **Environment** uses concrete details (not "beautiful field" but "green grass, blue sky")
- **Lighting** is explicit ("bright afternoon sunlight")

**How to adapt it:**
- Change style: "Archival documentary" or "Film noir"
- Change subject: "A skateboarder" or "A red sports car"
- Change environment: "Urban alley" or "Snow-covered forest"

---

### Example 2: Adding More Control

**Prompt:**
```
1970s film aesthetic. A chef in white apron dices onions in three quick cuts,
scrapes them into bowl. Medium shot from 45-degree angle. Restaurant kitchen
with copper pots on back wall. Warm overhead pendant lights.
Colors: brass, cream, sage green.
```

**Why this works:**
- **Counting actions** ("three quick cuts, scrapes") tells the AI exactly how long things take
- **Camera angle** (45-degree) shows exactly where you want the camera
- **Background details** (copper pots, back wall) help keep the scene consistent
- **Specific colors** (brass, cream, sage green) keep the colors from changing randomly

**How to adapt it:**
- Count actions: "five hammer swings" or "walks four steps"
- Specify angles: "eye level" or "low angle looking up"
- Add 3-5 colors that match your brand or mood

---

### Example 3: Simple Dialogue

**Prompt:**
```
Archival documentary, 16mm grain. Woman in red coat stands at rain-soaked
window, looking out. Medium close-up, eye level. Soft window light from left.
Colors: crimson, navy, amber, steel gray.

Dialogue:
- Woman: "It's still raining..."
```

**Why this works:**
- **Separate dialogue section** tells the AI this needs lip-sync
- **Short line** (4 words) fits perfectly in a 4-8 second video
- **Light direction** (from left) creates depth and mood
- **Style choice** (archival documentary, 16mm grain) sets the overall look

**How to adapt it:**
- Keep dialogue under 10-15 words for short clips
- Always use the "Dialogue:" block format
- Match line count to clip duration (4s = 1-2 short lines)

---

## The Golden Rule

> **"Clarity wins."**
>
> Replace vague words with specific visual details.
>
> Instead of "beautiful" → say "wet asphalt with neon reflections"
>
> Instead of "moves quickly" → say "pedals three times, brakes, stops"

| ❌ Weak | ✅ Strong |
|---------|-----------|
| "A beautiful street at night" | "Wet asphalt, zebra crosswalk, neon signs reflecting in puddles" |
| "Person moves quickly" | "Cyclist pedals three times, brakes, stops at crosswalk" |
| "Cinematic look" | "Anamorphic 2.0x lens, shallow depth of field, volumetric light" |
| "Nice lighting" | "Soft window light with warm lamp fill, cool rim from hallway" |

---

# Core Prompting Principles

These five principles form the foundation of effective video generation prompting.

## Principle 1: Choose Your Style First

**Style is your most powerful tool.** Start with the look you want—it guides all the other visual choices the AI makes.

### Style Examples

- **"1970s film"** → Grain, warm color cast, specific aspect ratio conventions
- **"16mm black-and-white"** → High contrast, documentary feel, film texture
- **"IMAX-scale scene"** → Epic scope, cinematic framing, pristine quality
- **"90s home video"** → VHS artifacts, saturated colors, handheld shake
- **"Film noir"** → Dramatic shadows, high contrast, moody lighting

### Camera Direction Examples

- "Wide establishing shot, eye level"
- "Medium close-up shot, slight angle from behind"
- "Aerial wide shot, slight downward angle"
- "Over-the-shoulder, looking toward subject"
- "Tracking left to right with subject"

**Why this works:** The AI has seen thousands of 1970s films, so when you say "1970s film," it knows to add grain, warm colors, and that era's camera style.

**Quick tip:** Always start your prompt with the style or look you want.

---

## Principle 2: Keep Movement Simple

**Movement is hard for AI—keep it simple.**

Each shot should have:
- **ONE camera movement** (or none)
- **ONE main action** from your subject

**Count your actions** to give the AI a sense of timing.

### Examples

| ❌ Weak | ✅ Strong |
|---------|-----------|
| "Actor walks across room" | "Actor takes four steps to window, pauses, pulls curtain in final second" |
| "Car drives fast" | "Car accelerates in three seconds, reaches 60mph, tires screech" |
| "Bird flies" | "Hawk dives downward for two seconds, spreads wings, glides" |

**Why this works:** "Actor walks" doesn't tell the AI how long to walk or when to stop. "Takes four steps, pauses two seconds" gives clear timing.

**Quick tip:** Count things. Use words like "quick," "slow," "gradual," or "in final second."

---

## Principle 3: Describe Your Lighting and Colors

**Lighting sets the mood.** If you don't specify lights and colors, the AI will pick random ones that might change throughout your video.

### Tell the AI About Your Lights

Describe:
- **Where light comes from** ("window light with desk lamp")
- **Which direction** ("from left" or "from right")
- **What kind** ("harsh," "soft," "bright," "dim")

**Example:**
```
Soft window light from left with warm desk lamp fill from right,
cool backlight creating rim on shoulders
```

### Pick 3-5 Specific Colors

Choose **3-5 exact colors** to keep your video consistent:

- "Teal, sand, rust"
- "Deep navy, cream, burnt orange"
- "Forest green, burgundy, gold, slate gray"
- "Crimson, charcoal, ivory"

**Why this works:** Naming exact colors keeps the AI from randomly changing colors between frames.

**Quick tip:** Say "burnt sienna" or "slate blue" instead of "earthy tones" or "cool colors."

---

## Principle 4: Put Dialogue in Its Own Section

**Important:** Keep dialogue separate from your scene description.

### The Format

```
[Visual description of shot and environment]

Dialogue:
- Character A: "Short, natural line here"
- Character B: "Brief response"
```

### Best Practices

✅ **DO:**
- Place dialogue in separate block below prose
- Keep lines concise and natural (under 15 words for short clips)
- Limit exchanges to match clip length
- Label speakers consistently
- Include tone descriptors: "calm," "excited," "whispered"

❌ **DON'T:**
- Embed dialogue in scene description
- Write long speeches for 4-second clips
- Mix dialogue with camera direction

### Duration Guidelines

- **4-second clips:** 1-2 short exchanges maximum
- **8-second clips:** 2-3 brief lines
- **12+ second clips:** Short conversation possible

**Why this works:** When you mix dialogue with scene description ("She says 'hello' while waving"), the AI might not know this needs lip-sync. A separate section makes it clear this is spoken dialogue.

---

## Principle 5: Use Reference Images for Better Control

**Starting with an image** helps keep your videos consistent.

### When to Use Image Input

- Lock in composition
- Define character design
- Establish aesthetic/style
- Ensure setting consistency across shots

### Technical Requirements

- Image must match target video resolution
- Supported formats: JPEG, PNG, WebP
- Model uses image as first-frame anchor
- Text prompt defines what happens next

**Example workflow:**
1. Upload reference image of character in specific outfit/location
2. Write prompt: "Same character walks forward three steps, turns to camera"
3. Generate video with visual consistency

**Why this works:** Converting words to images can vary a lot. Starting with an actual image gives the AI a clear starting point.

**Quick tip:** Make reference images of your characters or locations when creating multiple related videos.

---

# Beginner Examples

## Example 1: Simple Product Shot

**Prompt:**
```
Cinematic ad. iPhone 15 Pro rotating slowly on marble pedestal.
Dolly zoom in. Minimalist studio, soft shadows.
Lighting: rim light with cool backlight. Colors: titanium, deep blue, white.
```

**Breakdown:**
- **Style:** "Cinematic ad" (commercial aesthetic)
- **Subject + action:** "iPhone 15 Pro rotating slowly" (one clear action)
- **Camera:** "Dolly zoom in" (single movement from Higgsfield presets)
- **Environment:** "Marble pedestal, minimalist studio" (simple, clear)
- **Lighting:** "Rim light with cool backlight" (specific sources)
- **Colors:** Three anchors (titanium, deep blue, white)

**Why it works:** Simple, focused, one subject doing one thing. All elements clearly specified.

**How to adapt:**
- Change product: "Sneaker" or "Coffee mug"
- Change surface: "Wood table" or "Glass shelf"
- Change movement: "Emerges from shadow" or "Unfolds"

---

## Example 2: Nature Scene

**Prompt:**
```
Documentary style. A deer walks through morning mist in forest clearing.
Camera static, eye level. Dappled golden hour sunlight through trees.
Colors: forest green, gold, earth brown, soft white mist.
```

**Breakdown:**
- **Style:** "Documentary style" (naturalistic)
- **Subject + action:** "Deer walks through mist" (gentle, clear action)
- **Camera:** "Static, eye level" (no movement, simple framing)
- **Environment:** "Forest clearing" with "trees" (specific location)
- **Lighting:** "Dappled golden hour sunlight" (time of day + quality)
- **Colors:** Four nature-appropriate anchors

**Why it works:** No camera movement makes it easier. One animal doing one thing. Natural scenes work well for AI.

**How to adapt:**
- Change animal: "Fox" or "Owl landing on branch"
- Change time: "Twilight blue hour" or "Noon sunlight"
- Change weather: "Light rain falling" or "Snow drifting down"

---

## Example 3: Simple Character Action

**Prompt:**
```
Handheld camera, authentic lighting. Young woman in yellow raincoat
walks forward four steps, stops, looks up at sky.
Residential street, autumn leaves on sidewalk. Overcast daylight.
Colors: yellow, burgundy leaves, gray pavement, muted blue sky.
```

**Breakdown:**
- **Style:** "Handheld camera, authentic lighting" (documentary feel)
- **Subject + action:** Clear description + counted steps ("four steps, stops, looks up")
- **Camera:** "Handheld" implies natural shake, following subject
- **Environment:** "Residential street, autumn leaves" (specific season/setting)
- **Lighting:** "Overcast daylight" (soft, even)
- **Colors:** Four seasonal anchors

**Why it works:** Counting the steps gives clear timing. Simple actions in order (walk, stop, look). Everyday setting the AI knows well.

**How to adapt:**
- Count different actions: "Opens umbrella in three seconds"
- Change environment: "Park path" or "Beach boardwalk"
- Change season: "Spring flowers blooming" or "Snow-covered street"

---

# Prompt Templates

Copy these templates and fill in the brackets with your specific details.

## Template 1: Simple Product Showcase

**Use for:** Product videos, commercial content, social media ads

```
[Style preset: Cinematic/Archival/Modern]. [Product name] [key action:
rotate/emerge/unfold]. [Higgsfield camera preset: Dolly In/Crash Zoom/etc.].
[Minimal environment: 2-3 elements]. [Lighting: type + direction].
[Palette: brand colors + 2 supporting colors].
```

**Example:**
```
Cinematic commercial. Wireless headphones rotating slowly on wooden surface.
Dolly zoom in. Minimalist studio with soft shadows, blurred plant background.
Soft overhead lighting with cool rim light from left. Colors: matte black,
walnut wood, sage green, silver accents.
```

---

## Template 2: Social Media / UGC Content

**Use for:** Instagram Reels, TikTok, authentic creator content

```
[Handheld/authentic camera style]. [Person description] speaks directly
to camera, [natural gesture/movement]. [Location: home/outdoor/casual setting].
Natural/window light. [Mood: energetic/calm/conversational].

Dialogue:
- [Name]: "[Enthusiastic, brief statement - under 15 words]"
```

**Example:**
```
Handheld camera, natural lighting. Young creator in graphic tee sits on
bedroom floor, gestures excitedly while speaking to camera. Bedroom with
string lights and posters in soft focus background. Window light from right.
Energetic, authentic mood.

Dialogue:
- Creator: "You won't believe what happened today!"
```

---

## Template 3: Atmospheric / Establishing

**Use for:** Setting scenes, mood creation, B-roll

```
[Wide shot or aerial preset]. [Location with distinctive features].
[Time of day + weather]. [Camera: slow drift/static/orbit].
[Lighting: environmental source]. [Palette: mood-appropriate 3-5 colors].
[Audio: ambient soundscape].
```

**Example:**
```
Aerial wide shot. Coastal lighthouse on rocky cliff with waves crashing below.
Foggy dawn, mist rolling in from ocean. Camera: slow crane up revealing more
coastline. Soft pre-sunrise light, diffused by fog. Colors: slate gray rocks,
white lighthouse, navy ocean, soft pink dawn, white foam.

Audio: Waves crashing, distant foghorn, seabirds calling, wind.
```

---

# Common Mistakes & Solutions

## Mistake 1: Vague or Incomplete Prompts

### ❌ The Problem
```
"A beautiful street scene at night with nice lighting"
```

**Why it fails:** The AI has to guess everything—time, place, weather, lights, camera position, what happens, and colors. Results will be random.

### ✅ The Solution
```
"Film noir style. Wet cobblestone street with zebra crosswalk, neon bar
signs reflecting in puddles. Static shot, eye level. Midnight, light rain.
Single streetlamp creating pool of warm light, neon signs providing cool
blue-pink accent. Colors: deep blue shadows, warm amber streetlight, neon
pink-blue, wet black pavement."
```

**What changed:** Style defined, specific environmental elements, camera position, time/weather, light sources with direction, color palette.

---

## Mistake 2: Complex or Unclear Movement

### ❌ The Problem
```
"Person walks across room quickly while waving and talking on phone"
```

**Why it fails:** Three things happening at once (walking, waving, talking) confuses the AI. No timing information.

### ✅ The Solution
```
"Woman takes six quick steps across living room, phone to ear with right
hand. Camera pans right following movement, eye level. Mid-sentence, raises
left hand in brief wave gesture toward someone off-camera, then drops hand.
One action beats per 2 seconds."
```

**What changed:** Actions broken into steps with counts. Clear timing. Gestures happen one at a time (not all together). Camera separate from action.

**Better option:** Split into two shots—one for walking/talking, one for waving.

---

## Mistake 3: Missing Lighting Details

### ❌ The Problem
```
"Interior office scene with manager at desk"
```

**Why it fails:** Without lighting info, the AI picks randomly—might be bright office lights, dim mood lighting, or mixed inconsistent sources.

### ✅ The Solution
```
"Interior office scene. Manager at wooden desk with laptop and papers.
Overhead fluorescent tubes providing cool even light, warm desk lamp
adding fill from left side, window behind creating slight backlight.
Late afternoon."
```

**What changed:** Three light sources specified (overhead, desk lamp, window) with quality (cool, warm) and direction.

---

## Mistake 4: No Color Anchors

### ❌ The Problem
```
"Cinematic coffee shop scene with warm tones"
```

**Why it fails:** "Warm tones" is vague—could be orange, red, yellow, brown, or any combination. Colors may drift frame-to-frame.

### ✅ The Solution
```
"Cinematic coffee shop scene. Colors: rich espresso brown, cream ceramic
mugs, brass espresso machine, warm Edison bulb amber light, dark walnut
furniture."
```

**What changed:** Five specific color anchors stabilize palette across generation.

---

## Mistake 5: Dialogue Embedded in Prose

### ❌ The Problem
```
"Woman in cafe says 'I can't believe this happened' while looking worried
and holding coffee cup nervously"
```

**Why it fails:** When dialogue is mixed with actions, the AI might treat it as part of the scene instead of spoken words that need lip-sync.

### ✅ The Solution
```
"Woman in cafe holds coffee cup, shifts nervously in seat, speaks with
worried expression. Medium close-up, slight angle from across table.
Cafe background with blurred patrons, warm pendant lights overhead.

Dialogue:
- Woman: "I can't believe this happened."
```

**What changed:** Dialogue isolated in separate block. Visual description focuses on actions and camera.

---

# Quick Reference Card

## Prompt Structure
```
[Style]. [Subject] [action in beats]. [Camera]. [Environment].
[Lighting + direction]. [Colors: 3-5 anchors].

Dialogue:
- Character: "Brief line"
```

## Common Style Presets
- **Cinematic** - Film-like visuals, professional polish
- **Archival** - Vintage footage, film grain, period aesthetic
- **Film Noir** - Black & white, dramatic shadows, high contrast
- **Pixar-like 3D** - Animated 3D Pixar aesthetic
- **Hand-Drawn 2D** - Traditional animation look
- **Whimsical Stop Motion** - Claymation/handmade feel

## Essential Camera Moves
- **Static** - Locked camera (no movement)
- **Pan Left/Right** - Horizontal rotation
- **Dolly In/Out** - Moving on track toward/away
- **Crane Up/Down** - Vertical lift movement
- **FPV Drone** - Fast, agile weaving (action shots)
- **Handheld** - Documentary natural shake
- **Dolly Zoom** - Vertigo effect (zoom + dolly)
- **360 Orbit** - Complete circle around subject

## Shot Types
- **Wide** - Full scene with context
- **Medium** - Waist up, balanced detail
- **Close-Up** - Head and shoulders, emotions
- **Over-the-Shoulder (OTS)** - Conversation framing
- **Two-Shot** - Two subjects together
- **POV** - See through character's eyes
- **High/Low Angle** - Power dynamics

## Duration Tips
- **4s:** Simple single action
- **8s:** 2-3 action beats
- **12s+:** Multi-beat sequence or dialogue

## Credit Optimization
- **Square format** = 50% fewer credits
- **Shorter clips** = higher reliability
- **Iterate at lower resolution** first
- **Final render** at target quality

---

## Reference Files

### Need More Examples?
See [Progressive Examples](reference/PROGRESSIVE_EXAMPLES.md) for:
- 3 Beginner examples with full breakdowns
- 3 Intermediate examples with lighting/dialogue
- 3 Advanced examples with multi-shot sequences

### Need Technical Specs?
See [Technical Reference](reference/TECHNICAL_REFERENCE.md) for:
- API parameters and resolution options
- 7 built-in style presets with descriptions
- 50+ Higgsfield camera movements catalog
- 12 cinematic shot types reference

### Creating Complex Videos?
See [Advanced Workflows](reference/ADVANCED_WORKFLOWS.md) for:
- Multi-shot sequence production techniques
- Character consistency across shots
- Remix iteration strategy
- Recovery protocol for failing prompts
- Production optimization checklist

---

*Last updated: January 2025*
*Based on Sora 2 and Higgsfield.ai current capabilities*


---

## Referenced Files

> The following files are referenced in this skill and included for context.

### reference/PROGRESSIVE_EXAMPLES.md

```markdown
# Progressive Examples - Sora & Higgsfield Prompting

**← Back to [Main Skill](../SKILL.md)**

Learn through increasingly complex examples, each building on previous skills.

---

## Intermediate Level

**Goal:** Control lighting, color, camera movement, and add dialogue

### Example 1: Controlled Product Demo

**Prompt:**
```
Archival film aesthetic, 16mm grain. Hand wearing silver watch
pours coffee from French press in slow motion. Three-second pour,
steam rising. Medium shot, 45-degree angle. Kitchen counter with
morning newspaper, ceramic mug. Warm window light from left side,
soft shadows. Colors: brass press, cream ceramic, walnut counter,
amber coffee, silver watch.
```

**Breakdown:**
- **Style:** "Archival film, 16mm grain" (vintage texture)
- **Subject + action:** "Hand pours coffee" with duration ("three-second pour")
- **Detail anchor:** "Silver watch" (distinctive element for visual interest)
- **Camera:** "Medium shot, 45-degree angle" (specific framing)
- **Environment:** Multiple props (newspaper, mug, counter) create lived-in feel
- **Lighting:** "Warm window light from left, soft shadows" (direction + quality)
- **Colors:** Five specific anchors creating cohesive palette

**Why it works:** Timing the pour ("three seconds") controls the speed. Exact camera angle. Enough details to feel real without overwhelming the AI. Light direction creates depth.

**How to adapt:**
- Change beverage: "Tea from teapot" or "Wine into glass"
- Change props: "Notebook and pen" or "Phone and keys"
- Change lighting: "Overhead pendant" or "Candlelight"

---

### Example 2: Character with Dialogue

**Prompt:**
```
Cinematic interview style. Woman in leather jacket sits in armchair,
shifts position, speaks directly to camera. Medium close-up, slight
low angle. Living room with bookshelves and plants in soft focus
background. Natural window light from right creating soft shadows on
left side of face. Colors: black leather, sage green plants, warm wood
shelves, cream walls.

Dialogue:
- Woman: "I never expected it to turn out this way."
```

**Breakdown:**
- **Style:** "Cinematic interview style" (professional talking-head aesthetic)
- **Subject + action:** "Sits, shifts position, speaks" (natural movements)
- **Camera:** "Medium close-up, slight low angle" (flattering interview framing)
- **Environment:** Specific background elements ("bookshelves, plants") in soft focus
- **Lighting:** "Natural window light from right, shadows on left" (dimensional)
- **Colors:** Four anchors matching environment
- **Dialogue:** Single reflective line, appropriate length for 4-6 second clip

**Why it works:** Dialogue is separate for lip-sync. Short line fits the video length. Lighting adds depth. Background has details but stays out of focus.

**How to adapt:**
- Change setting: "Coffee shop" or "Office with window"
- Change lighting: "Golden hour backlight" or "Soft overhead"
- Change dialogue tone: Make line excited, calm, or mysterious

---

### Example 3: Dynamic Action Sequence

**Prompt:**
```
FPV drone style. Skateboarder accelerates down hill, performs kickflip,
lands cleanly. Camera follows behind and above, slight downward angle.
Urban residential street with parked cars, tree-lined sidewalk. Late
afternoon sunlight, long shadows. Colors: concrete gray, red skateboard,
green trees, golden light, blue jeans.

Audio: Board wheels on pavement, skateboard snap on flip, landing impact.
```

**Breakdown:**
- **Style:** "FPV drone" (Higgsfield preset for fast, agile movement)
- **Subject + action:** Sequence broken into beats (accelerate, kickflip, land)
- **Camera:** "Follows behind and above, downward angle" (dynamic tracking)
- **Environment:** Urban with specific elements (parked cars, trees)
- **Lighting:** "Late afternoon, long shadows" (time creates mood)
- **Colors:** Five anchors including action elements (skateboard, jeans)
- **Audio:** Foley cues specified (wheels, snap, impact)

**Why it works:** The FPV drone preset handles tricky camera work for you. Actions happen in clear order. Sound effects improve the result. Realistic setting for this type of action.

**How to adapt:**
- Change sport: "Cyclist weaving through pylons" or "Roller skater spinning"
- Change camera: "Steadicam tracking shot" or "Static low angle"
- Change environment: "Skate park" or "Beach boardwalk"

---

## Advanced Level

**Goal:** Multi-shot sequences, character consistency, complex cinematography

### Example 4: Multi-Shot Character Sequence

**Prompt:**
```
Shot 1:
Narrative drama, anamorphic 2.0x lens. Wide establishing shot, eye level.
Young hiker (blue backpack with yellow straps, tan jacket, dark beanie)
approaches forest trail entrance, takes first three steps onto path.
Morning golden hour, dappled sunlight through pine trees. Colors: forest
green, gold light, earth brown path, blue backpack, tan jacket.

Shot 2:
Same hiker (blue backpack with yellow straps, tan jacket, dark beanie,
maintain exact facial features). Medium shot from behind, continues walking,
pushes low pine branch aside with right hand. Same forest trail environment
(pine trees, dappled light). Golden hour lighting now with more shadows as
deeper into forest.

Shot 3:
Same hiker (blue backpack with yellow straps visible, tan jacket, dark beanie,
same facial features and proportions). Close-up, hiker stops, turns face toward
camera, expression of awe as looks up at towering trees. Backlit by sun breaking
through canopy, rim light on shoulders and edge of beanie. Same forest environment.
```

**Breakdown:**
- **Multi-shot structure:** Three connected shots with clear separations
- **Character consistency:** Exact description repeated each shot (blue backpack with yellow straps, tan jacket, dark beanie)
- **Explicit reference:** "Maintain exact facial features," "same facial features and proportions"
- **Environmental continuity:** "Same forest trail environment" with progressive lighting changes
- **Progression:** Wide → Medium → Close-up (classic coverage)
- **Action continuity:** Approaches → walks → stops and looks up
- **Lighting evolution:** Golden hour shifts as hiker moves deeper into forest

**Why it works:** Repeating unique details (yellow backpack straps) keeps the character looking the same. Clear instructions to keep features consistent. Same forest location throughout. Natural camera progression from wide to close.

**How to adapt:**
- Use for any multi-shot scene requiring character consistency
- Always repeat exact distinctive features (clothing, accessories, physical traits)
- Progress camera positions logically (wide → medium → close)
- Maintain environmental anchors across shots

---

### Example 5: Complex Dialogue Scene

**Prompt:**
```
Narrative thriller, 85mm portrait lens, f/2.8. Over-the-shoulder shot
from behind Detective Chen (gray suit, silver watch on left wrist,
short black hair). Detective Chen faces Suspect Martinez (leather jacket,
gold chain visible, stubble) across metal interrogation table. Detective
leans forward slightly, Suspect shifts back. Concrete walls, fluorescent
tube overhead creating harsh top light, deep shadows under eyes. Single
frosted window right side with dim daylight. Colors: teal shadows, warm
sodium highlights from overhead, desaturated mids, gray concrete, silver
watch, gold chain.

Dialogue:
- Chen: "Where were you on Thursday night?"
- Martinez: "I already told you—home."

Audio: Metal chair scrape, fluorescent buzz, distant door echo.
```

**Breakdown:**
- **Technical specs:** "85mm portrait lens, f/2.8" (cinematic detail)
- **Shot type:** "Over-the-shoulder" (classic dialogue framing)
- **Two characters:** Each with distinctive anchors (silver watch vs. gold chain)
- **Blocking:** "Leans forward... shifts back" (physical tension)
- **Environment:** Specific setting (interrogation room) with architectural details
- **Lighting:** Complex setup (harsh overhead, window fill, shadows specified)
- **Color grade:** Sophisticated palette (teal shadows, sodium highlights)
- **Dialogue:** Two-line exchange, interrogation tension
- **Audio:** Environmental sounds (chair, fluorescent, echo) enhance realism

**Why it works:** Unique details for each character (watch vs. chain) help tell them apart. Body language shows the power dynamic. Lighting sets the tense mood. Camera specs add polish. Sound effects bring the scene to life.

**How to adapt:**
- Use for any two-character dialogue scene
- Give each character unique visual marker (jewelry, clothing detail, physical trait)
- Describe blocking that shows relationship (leaning in = aggressive, pulling back = defensive)
- Match lighting to emotional tone (harsh = interrogation, soft = intimate)

---

### Example 6: High-Production Commercial

**Prompt:**
```
High-end commercial, IMAX-quality cinematography. Anamorphic 1.85:1 aspect,
35mm lens, shallow depth of field (f/1.4). Slow-motion product reveal:
Luxury perfume bottle (frosted glass, gold cap, minimalist label) emerges
from water surface, droplets cascading down sides in slow motion over
5 seconds. Camera: dolly zoom in synchronized with emergence. Black gradient
background, single spotlight from above creating bright rim on bottle top,
soft blue-green underlighting from water surface creates reflection on glass.
Foreground: water ripples in sharp focus. Midground: bottle emerging.
Background: soft black gradient. Grade: desaturated with punchy highlights on
gold cap, teal water reflection, pure white rim light. Colors: frosted glass
transparency, 24k gold cap, teal water, deep black, pure white highlights.

Audio: Gentle water movement, droplets falling, subtle orchestral swell.

Camera notes: Dolly zoom synchronized with 5-second emergence creates impossible
perspective shift, product remains same size in frame while background appears
to expand.
```

**Breakdown:**
- **Production value:** "IMAX-quality," "High-end commercial" (premium aesthetic)
- **Technical cinematography:** Anamorphic aspect, 35mm lens, f-stop specified
- **Slow-motion timing:** "5 seconds" with action beat (emergence)
- **Complex camera move:** "Dolly zoom synchronized with emergence" (advanced technique)
- **Layered environment:** Foreground/midground/background explicitly described
- **Sophisticated lighting:** Multiple sources (spotlight, underlighting) with effects (rim, reflection)
- **Color grade:** Technical terms (desaturated, punchy highlights) with specific palette
- **Audio layers:** Water foley + music cue
- **Camera notes:** Explains dolly zoom effect and intention

**Why it works:** Lots of detail creates a premium look. Professional camera language. Three layers of depth (front, middle, back). Precise lighting setup. Sound design adds quality. Exact timing for slow-motion effect.

**How to adapt:**
- Use for premium product shots requiring high production value
- Specify lens, f-stop, aspect ratio for cinematic control
- Layer environment (foreground, midground, background)
- Describe lighting with multiple sources and specific effects
- Add camera notes to explain complex techniques

---

## Key Learnings

### Intermediate Level Takeaways
1. **Precise timing** - Count seconds and beats
2. **Camera angles** - Specify exact degrees and heights
3. **Dialogue separation** - Always use separate dialogue blocks
4. **Audio cues** - Sound effects enhance realism
5. **Lighting direction** - Specify where lights come from

### Advanced Level Takeaways
1. **Character consistency** - Repeat exact distinctive details across shots
2. **Multi-shot planning** - Wide → Medium → Close progression
3. **Technical specs** - Lens, f-stop, aspect ratio add cinematic polish
4. **Blocking** - Physical movements show emotional dynamics
5. **Layered environments** - Foreground, midground, background create depth
6. **Color grading** - Professional color language (desaturated, punchy, etc.)

---

**← Back to [Main Skill](../SKILL.md)** | **See also:** [Technical Reference](TECHNICAL_REFERENCE.md) | [Advanced Workflows](ADVANCED_WORKFLOWS.md)

```

### reference/TECHNICAL_REFERENCE.md

```markdown
# Technical Reference - Sora & Higgsfield

**← Back to [Main Skill](../SKILL.md)**

Complete technical specifications for Sora 2 and Higgsfield.ai platform.

---

## API Parameters & Constraints

### Model Options

- **sora-2** (Standard tier)
- **sora-2-pro** (Advanced tier with extended capabilities)

### Resolution Options

**Both models:**
- 1280x720 (HD landscape)
- 720x1280 (HD portrait)

**Sora-2-pro only:**
- 1024x1792 (portrait)
- 1792x1024 (landscape)

**Square formats** (both models):
- 720x720
- 1080x1080

> **💰 Save Credits:** Square videos use ~50% fewer credits than widescreen!
>
> Use square format whenever your video allows it.

### Video Length Options

**Sora 2 Standard:**
- 4 seconds
- 8 seconds
- 12 seconds (maximum)
- Default: 4 seconds

**Sora 2 Pro:**
- 10 seconds
- 15 seconds
- 25 seconds (maximum)

> **💡 Pro Tip:** Shorter is better!
>
> Two 4-second clips are easier to make (and more reliable) than one 8-second clip.
> Make short videos and combine them later.

---

## Built-In Style Presets

Invoke these styles in your prompts for consistent aesthetics:

### 1. **Cinematic**
High-quality, film-like visuals with professional lighting and color grading.

**When to use:** Product showcases, narrative scenes, polished commercial content

**Example:** "Cinematic style. Sports car driving through desert highway at sunset."

---

### 2. **Archival**
Historical/vintage footage aesthetic with muted tones, film grain, sepia effects.

**When to use:** Period pieces, nostalgic content, documentary-style footage

**Example:** "Archival 1960s documentary. Family picnic in park, 8mm film grain."

---

### 3. **Film Noir**
Black-and-white, dramatic lighting with deep shadows, high contrast, moody atmosphere.

**When to use:** Mystery scenes, dramatic moments, stylized content

**Example:** "Film noir. Detective in fedora walks through rainy alley, streetlight shadows."

---

### 4. **Pixar-like 3D**
Animated 3D style with characteristic Pixar aesthetic.

**When to use:** Animated content, playful scenes, family-friendly videos

**Example:** "Pixar-like 3D. Cute robot character discovers flower in junkyard."

---

### 5. **Hand-Drawn 2D**
Traditional animation aesthetic with hand-drawn look.

**When to use:** Artistic content, stylized narratives, illustration-style videos

**Example:** "Hand-drawn 2D animation. Bird flying over watercolor landscape."

---

### 6. **Cardboard & Papercraft**
Tactile, handcrafted diorama look with paper textures.

**When to use:** Whimsical content, children's content, unique artistic style

**Example:** "Papercraft style. Tiny paper car driving through cardboard city."

---

### 7. **Whimsical Stop Motion**
Claymation/stop-motion aesthetic with tactile, handmade feel.

**When to use:** Playful content, artistic projects, nostalgic style

**Example:** "Whimsical stop motion. Clay character baking cookies in kitchen."

---

## Higgsfield Camera Controls

Higgsfield offers **50+ professional camera movements** without physical equipment:

### Zoom Controls

- **Crash Zoom In/Out** - Rapid, dramatic zoom
- **Dolly Zoom In/Out** - Vertigo effect (zoom + dolly)
- **Rapid Zoom In/Out** - Fast zoom for impact
- **YoYo Zoom** - Back and forth zoom movement

### Dolly Movements

- **Dolly In/Out/Left/Right** - Moving camera on track
- **Super Dolly In/Out** - Extended dolly movement
- **Double Dolly** - Complex dolly combination

### Crane & Jib Operations

- **Crane Up/Down** - Vertical camera lift
- **Crane Over The Head** - Sweeping overhead move
- **Jib Up/Down** - Similar to crane, smaller scale

### Rotation & Panning

- **Pan Left/Right** - Horizontal camera rotation
- **Whip Pan** - Fast, motion-blurred pan
- **Arc Left/Right** - Curved movement around subject
- **360 Orbit** - Complete circle around subject
- **3D Rotation** - Multi-axis rotation

### Specialized Techniques

- **FPV Drone** - Fast, agile drone weaving through spaces (perfect for action)
- **Snorricam** - Character-mounted perspective (face stays centered, world moves)
- **Head Tracking** - Auto subject-following as they move
- **Dutch Angle** - Tilted horizon for tension/unease
- **Bullet Time** - Slow-motion with dynamic rotating camera (Matrix effect)
- **Handheld** - Documentary-style natural shake
- **Steadicam** - Smooth continuous motion, floating feel

### Temporal Effects

- **Hyperlapse** - Accelerated time with camera movement
- **Timelapse (Landscape)** - Clouds, sunset, nature time passage
- **Timelapse (Human)** - People moving in accelerated time
- **Timelapse (Glam)** - Beauty/transformation time-lapse

### Transition Techniques

- **Flying Cam Transition** - Dynamic movement between scenes
- **Through Object In/Out** - Camera passes through object for transition
- **Object POV** - See from object's perspective

### Vehicle-Specific

- **Car Grip** - Racing cinematography mounted on vehicle
- **Car Chasing** - Following vehicle in pursuit
- **Road Rush** - Fast road movement (driver POV feel)
- **Buckle Up** - Interior vehicle perspective

### Additional Techniques

- **Static** - Locked camera position (no movement)
- **Low Shutter** - Motion blur effect for dreamy/ethereal feel
- **Focus Change** - Depth of field shifting between subjects
- **Lazy Susan** - Slow 360-degree rotation
- **Wiggle** - Subtle vibration or shake

---

## Cinematic Shot Types

### Close-Ups & Details

**Close-Up (CU)**
- Frame: Head and shoulders, captures facial expressions
- Use for: Emotional moments, reactions, dialogue emphasis
- Example: "Close-up of woman's face as she receives news"

**Extreme Close-Up (ECU)**
- Frame: Isolates specific details (eyes, hands, objects)
- Use for: Critical details, building tension, product features
- Example: "Extreme close-up of eye widening in fear"

**Medium Close-Up (MCU)**
- Frame: Face and shoulders, slight more context than CU
- Use for: Interviews, vlogs, personal connection
- Example: "Medium close-up of chef explaining recipe to camera"

### Establishing Shots

**Wide Shot**
- Frame: Full scene with subject in environment
- Use for: Establishing location, showing scale, context
- Example: "Wide shot of lone figure on beach at sunset"

**Extreme Wide Shot**
- Frame: Vast landscapes with subject small or distant
- Use for: Epic scale, isolation, environmental grandeur
- Example: "Extreme wide shot of hiker on mountain ridge"

**Medium Shot**
- Frame: Waist up, balances detail and context
- Use for: Conversations, actions, general coverage
- Example: "Medium shot of barista making coffee"

### Conversation & Dialogue

**Over-the-Shoulder (OTS)**
- Frame: From behind one person looking at another
- Use for: Conversations, establishing spatial relationship
- Example: "Over-the-shoulder shot of interviewer facing subject"

**Two-Shot**
- Frame: Two subjects in frame together
- Use for: Relationships, conversations, connections
- Example: "Two-shot of couple sitting on bench, talking"

### Perspective & Mood

**High Angle**
- Frame: Camera above subject looking down
- Effect: Subject appears vulnerable, small, weak
- Example: "High angle of child lost in crowd"

**Low Angle**
- Frame: Camera below subject looking up
- Effect: Subject appears powerful, imposing, dominant
- Example: "Low angle of CEO standing at window"

**Dutch Angle**
- Frame: Camera tilted, horizon diagonal
- Effect: Creates tension, unease, disorientation
- Example: "Dutch angle of person running through maze"

**Bird's Eye View**
- Frame: Directly overhead, looking straight down
- Effect: Omniscient perspective, pattern emphasis
- Example: "Bird's eye view of traffic intersection"

**POV (Point of View)**
- Frame: See through character's eyes
- Effect: Immersive, subjective experience
- Example: "POV shot of driver's view on highway"

---

## Camera Language Quick Reference

### Basic Framing
| Term | Meaning |
|------|---------|
| **ECU** | Extreme Close-Up (eyes, details) |
| **CU** | Close-Up (head/shoulders) |
| **MCU** | Medium Close-Up (face/chest) |
| **MS** | Medium Shot (waist up) |
| **WS** | Wide Shot (full scene) |
| **EWS** | Extreme Wide Shot (landscape) |

### Camera Angles
| Term | Meaning |
|------|---------|
| **Eye level** | Camera at subject's eye height |
| **High angle** | Looking down (vulnerable) |
| **Low angle** | Looking up (powerful) |
| **Dutch/Canted** | Tilted horizon (unease) |
| **Bird's eye** | Directly overhead |
| **Worm's eye** | Ground level looking up |

### Camera Movement
| Term | Meaning |
|------|---------|
| **Pan** | Horizontal rotation (left/right) |
| **Tilt** | Vertical rotation (up/down) |
| **Dolly** | Camera moves on track (in/out/left/right) |
| **Crane** | Vertical lift movement |
| **Zoom** | Lens focal length change (no camera movement) |
| **Tracking** | Camera follows subject motion |
| **Orbit** | Circle around subject |
| **Handheld** | Natural camera shake |
| **Steadicam** | Smooth floating motion |

### Lens Specifications
| Term | Meaning |
|------|---------|
| **16mm, 24mm, 35mm** | Wide angle lenses (more in frame) |
| **50mm** | Standard/normal lens (natural perspective) |
| **85mm, 105mm** | Portrait lenses (flattering for faces) |
| **200mm+** | Telephoto lenses (compression, isolation) |
| **f/1.4, f/2.8** | Aperture (lower = shallow depth, blurry background) |
| **f/8, f/11** | Aperture (higher = deep focus, everything sharp) |

### Cinematic Effects
| Term | Meaning |
|------|---------|
| **Shallow DOF** | Blurry background, subject in focus |
| **Deep focus** | Everything in frame is sharp |
| **Rack focus** | Shift focus from foreground to background (or reverse) |
| **Bokeh** | Aesthetic quality of out-of-focus areas |
| **Anamorphic** | Widescreen aspect with oval lens flares |
| **Film grain** | Texture from film stock (vintage feel) |
| **Vignette** | Darkened corners/edges |

---

## Duration & Pacing Guidelines

### 4-Second Clips
**Best for:**
- Single action or movement
- Simple product rotation
- Brief establishing shot
- One dialogue line (5-8 words)

**Example:** "Product rotates 180 degrees on pedestal"

### 8-Second Clips
**Best for:**
- 2-3 action beats
- Short conversation exchange
- Simple narrative moment
- Action with reaction

**Example:** "Person walks to door (4s), turns back to camera (2s), waves goodbye (2s)"

### 12+ Second Clips
**Best for:**
- Multi-beat sequences
- Longer dialogue exchanges
- Complex camera movements
- Establishing shots with movement

**Example:** "Wide shot of city skyline at dawn, camera cranes up as sun rises, revealing harbor below"

---

## Credit Optimization Strategies

### Resolution Strategy
1. **Iterate at 720x720** (square, lowest credit cost)
2. **Test prompt variations** without spending many credits
3. **Final render** at target resolution (1920x1080, etc.)

### Duration Strategy
1. **Start with 4-second clips** (most reliable, fastest generation)
2. **Combine clips in post** using video editing software
3. **Only use longer durations** when single continuous action required

### Format Strategy
- **Square (1:1)** = 50% fewer credits than widescreen
- **Portrait (9:16)** = Standard credit cost
- **Landscape (16:9)** = Standard credit cost

### Model Strategy
- **Sora 2 Standard** for testing and iteration
- **Sora 2 Pro** only for final renders requiring extended length or resolution

---

**← Back to [Main Skill](../SKILL.md)** | **See also:** [Progressive Examples](PROGRESSIVE_EXAMPLES.md) | [Advanced Workflows](ADVANCED_WORKFLOWS.md)

```

### reference/ADVANCED_WORKFLOWS.md

```markdown
# Advanced Workflows - Sora & Higgsfield

**← Back to [Main Skill](../SKILL.md)**

Professional techniques for multi-shot sequences, character consistency, and production optimization.

---

## Multi-Shot Sequence Production

### The Challenge
Keeping your character looking the same across multiple videos (AI doesn't remember previous videos).

### The Solution: Write Down Your Character Details

Create a reference guide for your character before you start:

```
CHARACTER: Detective Sarah Chen

DISTINCTIVE FEATURES (use in EVERY prompt):
- Gray tailored blazer with silver buttons
- White collared shirt
- Silver watch on left wrist (always visible)
- Black straight hair in low bun
- Square silver earrings
- Black slim trousers
- Black leather ankle boots

FACIAL FEATURES ANCHOR:
"Sharp jawline, defined cheekbones, almond-shaped eyes, straight nose"

REUSABLE PROMPT BLOCK:
"Detective Chen (gray blazer with silver buttons, white shirt, silver watch
visible on left wrist, black hair in low bun, square silver earrings, sharp
jawline, defined cheekbones, almond eyes, maintain exact facial features)"
```

### Shot-by-Shot Workflow

**Shot 1: Establishing**
```
Wide shot. Detective Chen (gray blazer with silver buttons, white shirt,
silver watch visible on left wrist, black hair in low bun, square silver
earrings) enters precinct, walks five steps to desk. Same character details
throughout sequence.
```

**Shot 2: Action**
```
Same Detective Chen (gray blazer with silver buttons, white shirt, silver
watch on left wrist, black hair in low bun, square silver earrings, sharp
jawline, defined cheekbones, maintain exact features from previous shot).
Medium shot, sets briefcase on desk, opens it.
```

**Shot 3: Close-up**
```
Same Detective Chen (maintain all previous features: gray blazer, silver
watch, low bun, square earrings, same facial structure). Close-up, pulls
file from briefcase, expression focused.
```

### Key Principles

1. **Repeat exact phrasing** for distinctive elements
2. **Include "maintain exact/same features"** instruction
3. **Reference previous shot** explicitly
4. **Use distinctive accessories** (watch, jewelry, specific clothing items)
5. **Anchor facial features** with consistent description

---

## Remix Iteration Strategy

### Change One Thing at a Time

Only change **ONE thing** at a time so you know what makes it better or worse.

### Example Progression

**Base prompt:**
```
Cinematic. Woman walks through park. Camera follows.
```

**Iteration 1 - Add style detail:**
```
1970s film aesthetic. Woman walks through park. Camera follows.
```

**Iteration 2 - Improve action (keep previous change):**
```
1970s film aesthetic. Woman takes six steps down park path, pauses at bench.
Camera follows.
```

**Iteration 3 - Add lighting (keep all previous):**
```
1970s film aesthetic. Woman takes six steps down park path, pauses at bench.
Camera follows at eye level. Golden hour sunlight through trees, dappled shadows.
```

**Iteration 4 - Add color palette (keep all previous):**
```
1970s film aesthetic. Woman takes six steps down park path, pauses at bench.
Camera follows at eye level. Golden hour sunlight through trees, dappled shadows.
Colors: forest green, gold light, burgundy jacket, earth brown path.
```

### Recovery Protocol

**If a shot fails repeatedly:**

1. **Freeze the camera** - Change to static shot
2. **Simplify action** - One simple movement only
3. **Clear the background** - Minimal environment
4. **Test** - Does it work now?
5. **Layer complexity** - Add back elements one at a time

### Example Recovery

**Failing prompt:**
```
FPV drone weaving through crowded market, following running child through
vendors, shoppers, stalls, under canopies, around corners
```

**Step 1 - Freeze camera:**
```
Static wide shot. Child runs straight down market aisle between two vendor stalls.
```

**Step 2 - If that works, add camera back:**
```
Camera pans right following child running straight down market aisle between
two vendor stalls.
```

**Step 3 - If that works, add environment:**
```
Camera pans right following child running straight down market aisle between
two vendor stalls selling fruit, colorful awnings overhead.
```

**Step 4 - If that works, increase camera complexity:**
```
Camera tracking shot following child running down market aisle between vendor
stalls selling fruit, colorful awnings overhead.
```

Add things back slowly until you find what makes it fail.

---

## Higgsfield Multi-Model Workflow

### Choosing the Right Model

**Use Sora 2 when:**
- Photorealistic output required
- High production value commercial content
- Natural human motion and lip sync
- Complex lighting and cinematography

**Use WAN 2.5 when:**
- Camera control is paramount
- Need synchronized audio generation
- Character animation with expressive movements
- 10-second clips with complex camera choreography

**Use Kling when:**
- Dialogue-heavy scenes with multiple speakers
- Character interaction primary focus
- Specific lip-sync requirements

**Use Popcorn when:**
- Need to replace character in existing video
- Maintaining motion while changing appearance

### Integrated Pipeline Example

**Project: Product launch video with spokesperson**

1. **Generate base scene** (Sora 2 Max)
   - High-quality studio environment
   - Professional lighting setup
   - Product on pedestal

2. **Add camera movement** (WAN Camera Controls)
   - Dolly zoom on product
   - Adjust motion timing
   - Refine framing

3. **Generate spokesperson** (Kling)
   - Dialogue-focused generation
   - Speaking to camera
   - Lip sync with script

4. **Composite in Higgsfield Canvas**
   - Combine product shot with spokesperson
   - Match lighting
   - Create cohesive final video

---

## Character Consistency Techniques

### Technique 1: Distinctive Accessories

**Why it works:** Unique accessories are easier for AI to remember than faces.

**Examples:**
- Specific watch (silver Rolex on left wrist)
- Unique jewelry (gold chain with pendant)
- Distinctive clothing (red backpack with yellow straps)
- Props (always carrying blue notebook)

**Implementation:**
```
Shot 1: "Woman (red backpack with yellow straps, blue denim jacket,
silver hoop earrings)"

Shot 2: "Same woman (red backpack with yellow straps visible, blue denim
jacket, silver hoop earrings, maintain exact facial features from previous shot)"
```

---

### Technique 2: Color Anchors

**Why it works:** Specific colors help maintain consistency.

**Bad approach:**
```
Shot 1: "Man in dark suit"
Shot 2: "Same man in suit"
```
Result: Suit color changes

**Good approach:**
```
Shot 1: "Man in charcoal gray suit with burgundy tie"
Shot 2: "Same man (charcoal gray suit with burgundy tie, maintain colors)"
```

---

### Technique 3: Explicit Facial Feature Description

**Why it works:** Anchoring specific facial features helps AI maintain consistency.

**Template:**
```
"[Character name] ([distinctive features], [jawline description],
[eye shape], [nose description], maintain exact facial features and proportions)"
```

**Example:**
```
"Marcus (square jaw, wide-set brown eyes, aquiline nose, prominent cheekbones,
maintain exact facial structure)"
```

---

### Technique 4: Reference Image Workflow

**Best approach for character consistency:**

1. Generate initial character shot with detailed prompt
2. Save best result as reference image
3. Use image as input for subsequent shots
4. Add text prompt: "Same character [action]"

**Benefits:**
- Visual consistency guaranteed
- Less prompt repetition needed
- Facial features locked in
- Clothing/accessories consistent

---

## Production Optimization Checklist

### Pre-Generation

- [ ] **Define shot purpose** - What is this shot achieving?
- [ ] **Select model** - Sora/WAN/Kling based on requirements
- [ ] **Choose resolution** - Square format if aspect allows (50% credit savings)
- [ ] **Set duration** - Start with 4-5s, extend only if needed
- [ ] **Prepare references** - Character images, location photos, style examples
- [ ] **Draft dialogue** - If speech required, write brief, natural lines

### Prompt Construction

- [ ] **Establish style** - Lead with aesthetic/preset
- [ ] **Specify camera** - Shot type, angle, movement
- [ ] **One clear action** - Subject doing one main thing
- [ ] **Count beats** - "Three steps," "five-second pour"
- [ ] **Light sources** - Type, direction, quality
- [ ] **3-5 color anchors** - Specific hue names
- [ ] **Distinctive details** - Character markers for consistency
- [ ] **Audio cues** - Foley and ambient sounds
- [ ] **Dialogue block** - Separate from visual description if applicable

### Post-Generation

- [ ] **Review for issues** - Physics violations, inconsistencies
- [ ] **Iterate one variable** - If changes needed, modify one element
- [ ] **Use camera controls** - Higgsfield presets for motion refinement
- [ ] **Check continuity** - If multi-shot, verify consistency
- [ ] **Stitch sequences** - Combine shots in editing software
- [ ] **Add effects** - Transitions, color grading, sound design
- [ ] **Export settings** - Final resolution and format

---

## Common Production Scenarios

### Scenario 1: Product Demo Video

**Goal:** Create professional 30-second product demo

**Workflow:**
1. Create 6-8 short clips (4-5s each)
2. Mix angles: Wide → Medium → Close-up → Detail
3. Use consistent lighting across all shots
4. Add camera presets: Dolly In, Orbit, Crash Zoom
5. Stitch in editing software
6. Add music and voiceover

**Credit optimization:** Use square format for testing, final render at 16:9

---

### Scenario 2: Character Story Sequence

**Goal:** Multi-shot narrative with same character

**Workflow:**
1. Create character design document
2. Generate initial hero shot (save as reference image)
3. Use reference image + prompts for subsequent shots
4. Repeat distinctive details in every prompt
5. Progress camera: Wide → Medium → Close-up
6. Maintain environmental continuity

**Key challenge:** Character consistency
**Solution:** Reference image workflow + explicit feature descriptions

---

### Scenario 3: Social Media Content

**Goal:** Quick, engaging 8-second clip

**Workflow:**
1. Start with square format (1:1) for Instagram
2. Simple prompt: One action, handheld camera, authentic lighting
3. Add dialogue if needed (5-8 words max)
4. Test variations quickly (square = 50% fewer credits)
5. Export best result

**Credit optimization:** Square format, 4-8s duration, minimal iterations

---

## Troubleshooting Decision Tree

### Problem: Character looks different across shots

**Try:**
1. Use reference image from first shot as input
2. Repeat exact distinctive features (accessories, clothing)
3. Add "maintain exact facial features from previous shot"
4. Simplify: Focus on 2-3 unique markers

---

### Problem: Video quality is inconsistent

**Try:**
1. Use same style preset across all shots
2. Specify consistent lighting (direction, color temperature)
3. Match camera angles (all eye-level, or all low-angle)
4. Use same color palette

---

### Problem: Generation fails repeatedly

**Try:**
1. Freeze camera (static shot)
2. Simplify action (one movement only)
3. Reduce environment details
4. Remove complex lighting
5. Use beginner-level prompt structure

---

### Problem: Timing feels off

**Try:**
1. Count actions explicitly ("three steps," "five-second pour")
2. Use shorter duration (4s instead of 8s)
3. Simplify to single action per shot
4. Add temporal markers ("in first second," "at end")

---

### Problem: Colors keep changing

**Try:**
1. Specify 3-5 exact colors (not "warm tones")
2. Use color names in environment ("walnut counter," "brass lamp")
3. Add color palette line at end of prompt
4. Test with simpler prompt first

---

## Advanced Prompt Patterns

### Pattern 1: Layered Environment

**Structure:**
```
Foreground: [sharp focus elements]
Midground: [subject and main action]
Background: [soft focus environment]
```

**Example:**
```
Foreground: Water ripples in sharp focus. Midground: Perfume bottle emerging.
Background: Soft black gradient.
```

---

### Pattern 2: Progressive Lighting

**Structure:**
```
Shot 1: [Initial lighting state]
Shot 2: [Same environment, lighting evolved]
Shot 3: [Final lighting state]
```

**Example:**
```
Shot 1: Golden hour, dappled sunlight through trees
Shot 2: Same forest, golden hour with more shadows (deeper into forest)
Shot 3: Same forest, backlit by sun breaking through canopy
```

---

### Pattern 3: Emotional Blocking

**Structure:**
```
[Character A action showing emotion] [Character B reaction showing opposite]
```

**Example:**
```
Detective leans forward (aggressive), Suspect shifts back (defensive)
```

---

## Learning Path

### Week 1: Foundations
- Master 5 core principles
- Generate 10+ single-shot videos
- Test different style presets
- Experiment with camera movements

### Week 2: Sequences
- Create first 3-shot sequence
- Practice character consistency
- Add dialogue to clips
- Use remix iteration strategy

### Week 3: Professional
- Produce 30-60 second story (10+ shots)
- Use reference images
- Combine multiple models
- Optimize credit usage

### Week 4: Mastery
- Develop personal style
- Build prompt template library
- Create character design documents
- Plan productions with shot lists

---

**← Back to [Main Skill](../SKILL.md)** | **See also:** [Progressive Examples](PROGRESSIVE_EXAMPLES.md) | [Technical Reference](TECHNICAL_REFERENCE.md)

```