Guides users through professional filmmaking workflows in Higgsfield Cinema Studio, including creating multi-shot sequences, configuring optical stacks, applying color grading, managing Soul Cast AI actors, and structuring per-scene prompts with Director Panel camera movements. Use when the user mentions Cinema Studio, Cinema Studio 2.5, Cinema Studio 3.0, Soul Cast, color grading, multi-shot video, shot sequences, storyboard workflow, Hero Frame, optical stack, keyframe interpolation, Elements system (@Characters/@Locations/@Props), Speed Ramp, Director Panel, Higgsfield Popcorn, Single Shot / Multi-Shot Auto / Multi-Shot Manual modes, Reference Anchor, Smart shot control, or any professional filmmaking workflow inside Higgsfield.
Standardmäßig ist der Prompt ausgewählt, der zuerst die Quelle prüft. Sie können zu einem direkten Befehl wechseln oder eine lokale Kopie herunterladen.
Quelldateien prüfen
Lesen Sie SKILL.md und alle von SkillsMP angezeigten Begleitdateien, bevor Sie sich für eine Installation entscheiden.
Mit Codex oder Claude installieren Kopieren Sie diesen Prompt, fügen Sie ihn in Codex, Claude oder einen anderen Assistant ein und lassen Sie die Skill-Seite prüfen und installieren.
Ein direkter Befehl überspringt den Prüf-Prompt. Prüfen Sie die Quelle, bevor Sie ihn ausführen.
Guides users through professional filmmaking workflows in Higgsfield Cinema Studio, including creating multi-shot sequences, configuring optical stacks, applying color grading, managing Soul Cast AI actors, and structuring per-scene prompts with Director Panel camera movements. Use when the user mentions Cinema Studio, Cinema Studio 2.5, Cinema Studio 3.0, Soul Cast, color grading, multi-shot video, shot sequences, storyboard workflow, Hero Frame, optical stack, keyframe interpolation, Elements system (@Characters/@Locations/@Props), Speed Ramp, Director Panel, Higgsfield Popcorn, Single Shot / Multi-Shot Auto / Multi-Shot Manual modes, Reference Anchor, Smart shot control, or any professional filmmaking workflow inside Higgsfield.
Generated-checked block (build_index.py verifies anchors). Read the linked sections for full context — these lines are routing aids, not the rules themselves.
Three Cinema Studio versions coexist (2.5 / 3.0 / 3.5) — user-selected, no auto-routing; always detect the version first →
Hard 512-character prompt cap; 2.5 @ Element chips eat ~80–100 hidden chars each →
Elements System: define @Characters/@Locations/@Props once, call everywhere
3.5 main UI = three pills (Genre / Style / Camera), each Auto by default — override only with a creative reason →
7 confirmed 3.5 genres: General · Action · Horror · Comedy · Noir · Drama · Epic — never invent genre names →
Style presets: 8 Color Palettes · 6 Lighting · 9 Camera Moveset Styles, or free-form Manual Style →
Camera axes: 3 bodies · 5 lenses · focal 8/14/35/50/75mm · aperture f/1.4–f/4–f/11; two camera vocabularies coexist — vocabulary follows the selected model →
Lead every delivered shot with the one-line settings strip (UI presets ≠ prompt text; never restate strip values in the body) →
The UI shot counter caps internal cuts — "strictly N shots" must keep N within it (observed cap 4, TODO confirm) →
Manual Style = saved ≤2,000-char block of project LAWS (grade, lighting law, texture, performance register); it replaces the preset axes →
480p drafts validate the prompt, NOT the take — no seed param; transfer a look via Hero Frame + start/end-frame pinning →
Resolution matrix has two axes: shot physics × delivery context; 4K-finish pipelines master at model max res in std mode →
Seedance fast mode cannot output 1080p — drafting in fast then "switching up" silently changes mode AND res →
Cinema Studio is Higgsfield's professional filmmaking environment — a full production
workflow for multi-shot, character-consistent cinematic content. It's fundamentally
different from single-clip generation: you're building sequences, not individual videos.
Cinema Studio vs Standard Generation
Standard Generation
Cinema Studio 2.5
Cinema Studio 3.0 (Business/Team Plan)
Cinema Studio 3.5
Output
Single clip
Multi-shot sequence
Multi-shot sequence
Multi-shot sequence
Character consistency
Manual / Soul ID only
Reference Anchor system
@ reference system (up to 9 images)
@ reference system (extends 3.0)
AI actor generation
Not available
Soul Cast — generate actors from parameters (no photos)
Soul Cast — General (2K) / Character (4K) / Location (4K) modes, 0.125 credits
Same Soul Cast surface — see 3.0
Camera control
Named presets
Director Panel (18 movements)
Director Panel + Smart (auto camera planning)
Camera Settings 4-axis panel (Camera Body / Lens / Focal Length / Aperture) — see Cinema Studio 3.5 section
Optical physics
Not available
Full camera body + lens stack
Not available
Available — restored via Camera Settings panel (different vocabulary from 2.5 — see Cinema Studio 3.5 section)
Color grading
Not available
Built-in suite (temp, contrast, grain, bloom, etc.)
Not available
Available — Color Palette axis in Style Settings panel (8 named palettes)
Shot structure
One prompt = one clip
Up to 6 scenes, 12s total max, per-scene config
Smart (auto) + Custom multi-shot (up to 6 scenes, 15s)
Multi-shot supported — Duration 4s–15s
3D exploration
Not available
Gaussian splatting — move inside any generated image
Not available
Same as 3.0 (not available)
Batch generation
Not available
Grid generation — up to 16 variations per credit
Not available
Configurable Batch Size (exploration multiplier)
Storyboard
Not available
Higgsfield Popcorn integration
Not available
Same as 3.0
Speed control
Not available
Speed Ramp (6 modes)
Speed Ramp (7 modes + Bullet Time, Hero Moment)
Not separately tested in this release — see UI for 3.5-specific surface
Up to 4K (Character/Location) · Up to 2K (General)
Not separately tested in this release
Max video duration
Model-dependent
12s
15s
15s
Aspect ratios
Model-dependent
6 options
7 options (+ 21:9 ultrawide)
7 options (Auto, 16:9, 9:16, 4:3, 3:4, 1:1, 21:9)
Plan requirement
All plans
All plans
Business/Team plan only
Plan availability not separately verified in this release — see Higgsfield plan documentation
Use Cinema Studio when:
You need 2+ shots that must feel like the same film
Character geometry must be locked across cuts
You want professional optical physics (lens flare, depth of field, sensor grain) — 2.5 only
You're building a short film, branded content, or any sequence longer than one clip
Stick with standard generation when:
Single clip is sufficient
Speed matters more than consistency
Exploring ideas before committing to a full sequence
⚠ Version Detection — Ask First
Before generating any Cinema Studio output, always ask the user:
Are you working in Cinema Studio 2.5 or Cinema Studio 3.0?
If the user has already stated their version (e.g., "I'm using 3.0" or "Cinema Studio 3.0"), remember it and don't ask again. But never assume — 2.5 and 3.0 have fundamentally different feature sets.
Why this matters:
2.5 has optical physics (camera body + lens stack), color grading, 3D Mode, grid generation
3.0 has none of those — outputting them wastes the user's time and causes confusion
3.0 has features 2.5 doesn't: native audio, Smart shot control, 21:9 ultrawide, 15s duration
Speed Ramp options differ between versions
Genre lists differ between versions
Once the version is known, use ONLY that version's output format and feature set. Never mix features from one version into output for the other.
The 10-Step Cinema Studio 2.5 Workflow
Cinema Studio 2.5 extends the pipeline in both directions: pre-production (Soul Cast +
location prompt) before generation, and post-production (color grading) after.
1. SCRIPT → Write or paste your scene description / shot list
2. SOUL CAST → (New in 2.5) Generate AI actors from parameters or use saved Elements
3. REFERENCE → Upload character photo → create Reference Anchor (or use Soul Cast actor)
4. ELEMENTS → (Optional) Define @Characters, @Locations, @Props if needed
5. OPTICAL STACK → Select camera body + lens + focal length + aperture (image mode)
6. HERO FRAME → Generate a key image that defines the visual tone
7. COLOR GRADE → (New in 2.5) Apply color grading to keyframes before video generation
8. CAMERA CONFIG → Set Director Panel movement + Speed Ramp + Duration in UI
9. SHOT MODE → Choose Single Shot / Multi-Shot Auto / Multi-Shot Manual
10. GENERATE → EXPORT → Chain into timeline or export to editing
Elements System — Define Once, Call Everywhere
Elements are Cinema Studio's reusable asset library. Create a Character, Location, or Prop
once and reference it with @ in any subsequent prompt.
Save → now available across all shots in this project
Per-Character Emotion: In Multi-Shot mode, each character can have an emotion setting
per scene. The available emotions are:
Emotion
Effect
Joy
Smiling, warm expression, positive energy
Fear
Wide eyes, tense posture, defensive body language
Surprise
Raised brows, open mouth, alert stance
Sadness
Downcast eyes, slumped posture, muted energy
Set emotion per character per scene in the UI — this keeps expression changes out of the
prompt field and lets the model handle the subtlety of facial animation.
@ tag rule — use exactly what the user provides, nothing more
Only use @ tags for Elements the user has explicitly given you in this conversation.
If they give you @Marcus but no location or prop tags, write the location and props
as plain description. Never invent or assume @ tags for anything the user hasn't named.
Each @ tag the user gives you = they have that Element set up. Silence = they don't.
No Elements provided — write everything as description:
A woman with dark hair and a red coat walks through a narrow downtown alley at night.
She carries a worn leather briefcase. She stops under a streetlight, turns to camera.
User provides @Sarah only — use it, describe the rest:
@Sarah walks through a narrow downtown alley at night.
She carries a worn leather briefcase. She stops under a streetlight, turns to camera.
User provides @Sarah, @DowntownAlley, @LeatherBriefcase — use all three:
@Sarah walks through @DowntownAlley, carrying @LeatherBriefcase.
She stops under a streetlight, turns to camera.
Key rule: Match exactly what the user gives you. No more, no less.
Tags they give you = Element exists. Anything else = write it as description.
Element Library Surface — Source Tabs and Categories
The @ symbol opens a unified asset picker. The picker is organized as a two-dimensional structure: source tabs across the top (where the asset comes from), and element categories within the Elements tab (how project assets are grouped).
Source tabs (5):
Tab
What's in it
Uploads
Assets the user brings in directly — original photos, plates, prop shots
Image Generations
Prior generated stills available for reference (useful for self-referencing continuity across shots — pull a specific look or pose that emerged from a previous generation back into a new one)
Video Generations
Prior generated clips available for reference (useful for camera-memory or motion reference — feed a previous clip's framing or movement back into a new generation)
Elements
Curated/saved assets in the project library — the primary continuity surface
Liked
Favorited assets across the picker — fast recall for go-to references
Element categories within the Elements tab (6):
Category
What it groups
All
Everything in the project
Pinned
Surfaced for quick access in this project — the user's working set for the current sequence
Shared
Accessible across collaborators or contexts (exact sharing scope is platform-managed)
Build the element library before generation. For a project with recurring characters, locations, or props — a short film, a series of branded videos, a multi-shot sequence — establish each recurring asset as an Element first, then reference it via @ in every shot. This is the platform-native implementation of the Reference-Based Prompt Mode discipline already documented elsewhere: references with assigned roles, not references as inspiration. See ../higgsfield-seedance/SKILL.md for the Reference-Based Prompt Mode discipline, and the existing "@ tag rule — use exactly what the user provides" guidance earlier in this section for how to use @ tags inside the prompt itself.
Cross-shot continuity tip. For continuation work, the Image Generations and Video Generations source tabs let you reference prior outputs from the same project — useful when a Character Element doesn't capture a specific look or pose that emerged from a particular generation. Pull the exact frame back in via the Generations tab rather than re-prompting from scratch.
Soul Cast — AI Actor Generation
Soul Cast is Cinema Studio 2.5's character generation system — create AI actors from
parameters instead of uploading photos. This is fundamentally different from Soul ID.
Casual, Formal, High Fashion, Military, Sporty, Workwear, Vintage (7 styles)
Key Features
Add up to 3 Soul Cast characters per keyframe — choose from saved Elements or generate on the spot
"Save to elements" button to reuse a specific Soul Cast actor across projects
Every actor auto-generates a backstory + character sheet (personality traits, motivation, fear, flaw, strength)
Designed to eliminate the "plastic/waxy" AI look — excels at skin textures and emotions
Powered by the Nano Banana 2 model under the hood
Soul Cast Workflow
1. Open Cinema Studio → Navigate to Soul Cast panel
2. Set Genre + Era + Budget to establish the visual world
3. Select Archetype + Identity + Physical Appearance
4. Add Details (imperfections) + Outfit
5. Generate → Review backstory + character sheet
6. Save to Elements → Now available as @CharacterName across all shots
7. Repeat for additional characters (up to 3 per keyframe)
Built-in Color Grading Suite
Cinema Studio 2.5 adds a post-production color grading suite applied to keyframes
before video generation — enabling unified visual cohesion across all clips.
Controls
Control
Effect
Color temperature
Warm ↔ cool shift
Contrast
Shadow/highlight separation
Saturation
Color intensity
Sharpness + effects
Detail enhancement
Highlights
Bright area control
Film grain
Analog texture overlay
Exposure
Overall brightness
Bloom
Highlight glow/diffusion
Workflow
Generate your keyframe image (Hero Frame or grid selection)
Click the keyframe → "Colorgrade" button
Adjust settings to taste
Apply — grade is baked into the keyframe before video generation
Repeat for each keyframe to maintain visual cohesion across the sequence
Tip: Grade your Hero Frame first, then match subsequent keyframes to it.
This is the post-production equivalent of a DIT (Digital Imaging Technician) on set.
Optical Physics Engine
Cinema Studio's Image Mode gives you a full camera body + lens stack. These are the
exact names from the UI (previous versions had wrong names — these are corrected).
Claude's job when building a Cinema Studio image prompt: Always recommend a specific
optical stack — body + lens + focal length + aperture — with a one-line reason why.
Never leave the optical stack blank or say "choose based on preference." Make a call.
Optical Stack Recommendations by Intent
Use the user's creative intent to select the stack. The most important signal is
what the image needs to feel like — not just the genre.
Portrait / Character Focus
Feel
Body
Lens
Focal
Aperture
Why
Warm, intimate, skin-flattering
Full-Frame Cine Digital
Warm Cinema Prime
50mm
f/1.4
Flattest most natural face rendering, creamy background separation
UI dropdowns — Camera body, Lens, Focal length, Aperture, Genre, Director Panel, Speed Ramp
Prompt field — Scene description only. No camera/lens/aperture text goes here.
Claude must always output these as two clearly separated blocks:
Block 1 — UI Settings (select these in Higgsfield before pasting the prompt)
Camera: Full-Frame Cine Digital
Lens: Warm Cinema Prime
Focal: 50mm
Aperture: f/1.4
Genre: Intimate
↳ Why: Flattest, most natural face rendering. Creamy background separation
keeps all attention on the character without any lens distortion.
Block 2 — Prompt (paste this into the Cinema Studio prompt field)
[Scene description only — no camera, lens, aperture, or genre language here]
Rule: Never put camera body names, lens names, focal lengths, apertures, or genre
names inside the prompt text. They live in the UI, not the prompt field.
f/11 — Deep depth of field. Everything sharp front to back. Product, landscape, architecture
Director Panel — 18 Camera Movements
⚠ Use ONLY these exact movement names in Cinema Studio output. General cinematic terms like "Crane Down", "Whip Pan", "FPV Drone", "Crash Zoom" etc. (from vocab.md) are valid for standard video generation and image prompts outside Cinema Studio, but they are NOT Director Panel options. Inside Cinema Studio, map to the closest equivalent (e.g., Crane Down → Jib Down, Crane Up → Jib Up).
All movements available in Cinema Studio's Director Panel:
Movement
Description
Best for
Static
Locked off, no movement
Dialogue, tension, composition
Handheld
Organic shake, documentary feel
Urgency, realism, action
Zoom Out
Focal length pulls back
Revelation, isolation
Zoom In
Focal length pushes in
Intensity, focus on subject
Camera Follows
Tracks subject movement
Chase, pursuit, accompaniment
Pan Left
Horizontal sweep left
Reveal, environment scan
Pan Right
Horizontal sweep right
Reveal, environment scan
Tilt Up
Vertical sweep up
Scale, aspiration, reveal
Tilt Down
Vertical sweep down
Weight, consequence, ground
Orbit Around
360° around subject
Isolation, drama, examination
Dolly In
Physical push toward subject
Emotional intimacy
Dolly Out
Physical pull from subject
Distance, context reveal
Jib Up
Camera rises on arm
Scale, context, god's eye
Jib Down
Camera descends on arm
Grounding, arrival
Drone Shot
Aerial movement
Landscape, scale, geography
Dolly Left
Lateral push left
Parallel tracking, reveal
Dolly Right
Lateral push right
Parallel tracking, reveal
360 Roll
Camera rolls on axis
Disorientation, stylized
Auto
Model selects best movement
When unsure
Speed Ramp
Controls temporal feel of the shot. Set per-scene in Cinema Studio.
⚠ Speed Ramp options differ between versions. Use ONLY the table for the version the user specified. Never output a 2.5 ramp name in 3.0 output or vice versa.
Cinema Studio 2.5 Speed Ramps (6 modes)
Mode
Effect
Best for
Linear
Consistent speed throughout
Standard, natural
Slow Mo
Reduced playback speed
Impact moments, emotion
Speed Up
Accelerated playback
Montage, energy, passage of time
Impact
Fast → sudden slow at key moment
Action, hits, reveals
Auto
Model selects based on content
When unsure
Custom
Draw your own speed curve
Precise creative control
Custom curve: Blue line with draggable nodes. Pull up = slow down, pull down = speed up.
Left = beginning of clip, right = end.
2.5-only values — NEVER use in 3.0 output: Linear, Slow Mo, Speed Up, Impact, Custom
Cinema Studio 3.0 Speed Ramps (7 modes)
Mode
Effect
Best for
Auto
Model selects based on content
When unsure
Slow-mo
Reduced playback speed
Impact moments, emotion
Ramp Up
Gradual acceleration
Building energy, montage
Flash In
Fast start → ease to normal
Dramatic entrances, openings
Flash Out
Normal → sudden snap acceleration
Launches, exits, explosive action
Bullet Time
Ultra-slow at key moment, normal around it
Action hits, reveals, hero beats
Hero Moment
Slow build → dramatic pause → release
Character reveals, power moves
3.0-only values — NEVER use in 2.5 output: Ramp Up, Flash In, Flash Out, Bullet Time, Hero Moment
Genre Selection
⚠ Genre lists differ between versions. Use ONLY the genres for the version the user specified.
Cinema Studio 2.5 Genres (8)
General · Action · Horror · Comedy · Western · Suspense · Intimate · Spectacle
Genre affects lighting defaults, color grading baseline, and motion style suggestions.
You can always override with explicit prompt language — genre is a starting point.
Genre
Lighting default
Color tendency
Motion style
General
Neutral
Balanced
Varies
Action
Hard, high contrast
Desaturated, cool
Dynamic
Horror
Low key, harsh shadows
Desaturated, teal/green tint
Slow or sudden
Comedy
Bright, even
Warm, saturated
Light, energetic
Western
Golden hour / dusty
Warm amber
Steady, wide
Suspense
Low key, motivated
Cold, muted
Slow build
Intimate
Soft, warm
Warm, skin-flattering
Gentle
Spectacle
Dramatic, high contrast
Bold, saturated
Epic, sweeping
2.5-only genres — NEVER use in 3.0 output: Western, Suspense, Intimate, Spectacle
Cinema Studio 3.0 Genres (7)
General · Action · Horror · Comedy · Noir · Drama · Epic
3.0-only genres — NEVER use in 2.5 output: Noir, Drama, Epic
Shot Modes
Single Shot
Standard generation — one prompt, one clip. Same as regular Higgsfield but with the
optical stack and Director Panel applied.
Multi-Shot Auto
Describe the full sequence in one prompt. Cinema Studio breaks it into shots automatically.
Good for: rough sequences, exploration, when you don't need per-shot control.
Example: "A woman walks into a bar, sits down, orders a drink. Bartender slides it over.
She takes a sip and stares at her reflection in the mirror behind the bottles."
→ Cinema Studio Auto generates: Establishing shot / Walk to seat / Order / Drink delivery / Reflection closeup
Multi-Shot Manual
Full per-scene control. Up to 6 scenes, each with its own:
Prompt
Camera movement (Director Panel)
Speed Ramp
Duration
⚠ 12-Second Total Runtime Cap: The combined duration of all scenes in a Multi-Shot
sequence cannot exceed 12 seconds total. Plan your per-scene durations to stay within
this limit (e.g., 6 scenes × 2s each, or 4 scenes × 3s, or 3 scenes × 4s).
Cost transparency: Multi-Shot Manual with 4 variations = 24 generations total (6 scenes × 4 variations). Plan credits accordingly.
Multi-Shot Manual workflow:
Scene 1: Establishing shot — wide, sets location
Scene 2: Character introduction — medium, shows protagonist
Scene 3: Action beat — closer, movement
Scene 4: Reaction — close-up, emotion
Scene 5: Consequence / turn — medium or wide
Scene 6: Resolution / button — varies
Morph-Cut and Smooth-Cut breathing room — 2-second still moments. Editors
land morph cuts and smooth cuts (transition styles in DaVinci Resolve, Premiere,
and similar editors that hide the cut by morphing similar frames) on stillness,
not on motion. If every shot starts and ends mid-action, every cut between them
has to happen mid-action too — and the model renders mid-action transitions
poorly. Build cut-friendly material by prompting an explicit 2-second
still-or-near-still moment at the start AND end of every shot. Examples
of buildable stillness: a held look, a settled pose, a held pause before motion
begins, a held breath after motion stops. The middle of the shot can be as
kinetic as the script demands; the boundaries earn the cuts. This rule is
adapted from the Mr. Core methodology and pairs with editor-side workflow in
../higgsfield-pipeline/SKILL.md § Stage 8 — Assembly.
Reference Anchor System
The Reference Anchor locks character geometry (facial structure, proportions, costume silhouette)
across all shots in a sequence — preventing the character drift that happens in standard generation.
Setup:
Upload a clear, well-lit reference photo of your character
Cinema Studio generates an anchor from it
Every subsequent shot in the project references this anchor
Combine with Soul ID for full face + geometry lock
Best practices:
Use a front-facing, neutral expression photo for the anchor
Avoid glasses, hats, or accessories in the anchor image if they won't always be present
One Reference Anchor per character — add multiple if you have a cast
The anchor persists across the whole Cinema Studio project
Location Reference Sheets
Locations deserve the same asset-first treatment as characters. If a room, alley,
vehicle interior, or exterior reappears across multiple shots, generate it once as
a standalone asset and reuse it — don't re-describe it inside each scene prompt.
Scene-by-scene re-description is the root cause of the "every shot reinterprets
the environment" failure: walls drift, furniture rearranges, scale shifts, and the
world stops feeling like a single place.
The Five-View Location Sheet
A location sheet captures the space from enough angles that the model has a
complete spatial model to draw from. Generate these five views as one asset set:
View
What it locks
Straight-on wide
The overall scale, focal point, and horizon line
Left-angle perspective
Side geometry, depth on the left
Right-angle perspective
Side geometry, depth on the right
Reverse / opposite view
What's behind the camera in the wide — finishes the 360° footprint
Close-up environmental details
Texture, wear, signage, small props that anchor identity
Use Grid Generation (2×2 or 4×4) or 3D Mode on a hero wide to produce the alt
angles from a single seed — this keeps the light and material identity consistent
across views, which is the whole point.
What to Preserve on Reuse
When you bring the location back for a new shot, lock these four properties:
architecture (walls, openings, scale), light quality (source direction, color
temperature, hardness), color treatment (palette, grade), and key environmental
details (signage, damage, specific props that make this place THIS place). Keep
these out of the scene prompt — they live in the location asset. The scene prompt
should describe only what changes: character action, camera behavior, atmospheric
motion, time-of-day shift if any.
Failure Mode This Prevents
Without a location sheet, a six-scene sequence set in "a rain-soaked alley" will
render six different alleys. Each scene prompt re-interprets "rain-soaked alley"
from scratch. With a location sheet tagged as an @ Element or loaded as a
reference, all six scenes share one alley — the puddles are in the same places,
the neon sign says the same thing, the brick wall has the same pattern.
For the camera vocabulary that pairs with location reveals (Pan, Crane, Dolly Reveal),
see ../../vocab.md.
Reference Sheet Types — Beyond Characters and Locations
Reference sheets aren't only for faces and rooms. The underlying principle is
narrower than "character or location" — a reference sheet is any asset built
once to lock a specific property of the production. Each sheet type fixes a
different axis of the work: identity, architecture, motion behavior, wardrobe,
palette, or product. If a property has to read consistently across multiple
shots, give it its own sheet. If it only matters for one shot, write it inline
and move on.
When a sheet pays for itself: if a property will appear in three or more
shots, the sheet is cheaper than re-describing it every time and dramatically
more consistent than relying on prompt text alone. Single-shot details — a
one-time prop, a background extra, a one-off lighting cue — don't earn a sheet.
The character sheet (see ../higgsfield-soul/SKILL.md → Character Sheet
Creation) and the Location Reference Sheet above already cover identity and
architecture. The three sheet types below extend the same pattern to the rest
of the production:
Motion / Camera Sheet
A short reference clip — 3–10 seconds is plenty — that captures a camera path
or motion rhythm you want repeated across multiple shots. The sheet locks the
movement signature, not the content of any single shot. Treat it as a
project-wide style anchor pulled into a generation as an @Video reference, the
way a DP would carry a "look" across an entire film.
Use cases: a music video where the same dolly-in arc punctuates every chorus;
a fight sequence where a signature whip-pan recurs at every climax beat; a
brand piece where every product shot starts on the same slow orbit. Distinct
from a Kling 3.0 Motion Control reference clip — that drives a single
generation's motion transfer. The Motion / Camera Sheet is reusable across the
whole project.
Outfit / Material Sheet
Dedicated wardrobe reference. Locks fabric texture, color values, garment fit,
fastenings, and how the material moves and folds. Generate it once with even
lighting and multiple angles, then reuse it any time the costume needs to read
identically across cuts.
Use cases: a hero costume that appears across 12 scenes; a specific jacket that
must look the same in close-up and wide shot; a uniform that recurs across an
ensemble cast. The outfit sheet is not the same as a character sheet — the
character sheet locks identity (face, build, hair, distinguishing marks), while
the outfit sheet locks what the character is wearing AND how that wardrobe
behaves in motion. Build both when you have a hero costume on a hero character.
The piano test — wardrobe complexity has a generation cost. Every detail of
wardrobe complexity — buttons, ties, jewelry, scarves, layered garments,
intricate prints, fastening hardware — costs the model rendering budget. The
piano test (adapted from the Mr. Core methodology) is the rule: if a wardrobe
element is as visually demanding for the model to render as a piano in the
frame, the model will spend its budget rendering that element instead of the
action. Strip the wardrobe to the simplest silhouette that still reads as the
character. A trench coat reads as a trench coat without the buttons rendering
correctly. A uniform reads as a uniform without the rank insignia drifting
across frames. Design the costume's signature — the silhouette and one
identifying feature — and let the rest go simple. Hero costumes that must
render under close-up scrutiny earn their complexity; background costumes do
not.
Palette / Mood Sheet
A color and tonal anchor for the project. Lock the visual mood across an entire
sequence: shadow density, key/fill ratio, saturation level, signature accent
colors, grade direction. Use a single curated still or a small grid of grade
references — the goal is to give every subsequent generation a consistent
palette to reach for.
Use cases: a noir piece where every scene shares the same shadow density and
warm-amber-on-cool-blue grade; a cyberpunk sequence where neon-to-base-light
ratio stays fixed across 15 cuts; a brand campaign where a signature accent
color appears in every shot regardless of subject. Pairs naturally with the
Soul Hex color system and curated moodboards in
../higgsfield-moodboard/SKILL.md.
Product Reference Sheet
When a product reappears across multiple shots — a hero SKU in a brand
campaign, a recurring prop in a narrative piece, a single item shot from a
dozen angles for ecommerce — give it the same asset-first treatment as a
character or location. Generate the product once as a multi-view reference
sheet, then call it in subsequent shots. Re-describing the product inside
each scene prompt is the root cause of "every shot reinterprets the
product": branding shifts, materials drift, geometry rearranges, and the
hero stops feeling like a single object.
The Product Reference Sheet uses a 7-part prompt scaffold. Each part fixes
one axis of the product's identity. Build the parts in order — the locks
come first, then composition, then surface treatment, then capture
parameters, then negative space.
1. Identity Lock — what the product is. Geometry, scale, color
values, defining proportions, distinguishing physical features. The same
discipline as a character sheet's identity block, applied to an object.
2. Branding Lock — placement, scale, color, and treatment of every
brand element on the product: logos, wordmarks, tags, stitched labels,
embossed marks. Specify position relative to product geometry (centered
front panel, left sleeve, lower right corner) and rendering style
(embroidered, screen-printed, embossed, foil-stamped).
3. Layout — the multi-view grid. Standard Product Reference Sheet
layout covers eight orthographic views plus macro close-ups:
Reverse details, secondary branding, construction seams
Left side
Profile geometry, depth on the left
Right side
Profile geometry, depth on the right
Top
Top-down silhouette, crown / lid / opening geometry
Bottom
Underside details, base / sole / footprint
3-quarter
Hero angle — combines front and side reads
Macro close-ups
Branding detail / material weave / construction join / any product-specific detail that earns its own frame
Eight views vs. the Five-View Location Sheet's five — products need more
orthographic angles than locations because the camera is closer and the
geometry is the subject, not the environment. Macro close-ups are the
product analogue of the location sheet's close-up environmental-details
view, scaled up to one per product-specific concern.
4. Background — #DCDCDC light gray, shadowless lighting, no
gradients, no reflections. Studio product-photo convention: the
background contributes zero visual information so the product reads as
the only subject. Specify all four constraints — #DCDCDC alone without
the shadowless + no-gradients + no-reflections trio still leaves the
model room to add atmospheric noise.
5. Realism — the Material Realism block. Reusable template populated
per material; six axes that together make a surface read tactile and
physically grounded rather than rendered-flat:
Raised structure — what stands proud of the base surface and by
how much
Tight density — how packed the surface elements are per unit area
Visible direction — the directionality of grain, weave, thread, or
flow across the surface
Micro shadowing — the small shadows cast by raised elements into
the surface valleys
Surface compression following form — how the material deforms where
it meets seams, edges, or attached elements
Curvature integration — how the surface and its texture follow the
product's overall geometry rather than sitting on top of it
Populate the six axes with material-specific vocabulary. Embroidery,
leather, knit, satin, brushed metal, and suede each have their own value
set; the six axes stay constant.
6. Camera — default product reference package: Canon EOS R5 +
RF 100mm f/2.8L Macro IS USM, f/8, ISO 100. The 100mm macro lens at
f/8 holds the whole product in sharp focus from front-most edge to
back-most edge across all eight views; ISO 100 keeps sensor noise out
of the gray background. Substitute only when a specific shot needs a
different read — e.g. shallow depth of field for a hero brand-detail
macro (f/4), or a wider focal length for an oversized product
(RF 50mm).
7. Restrictions — the content-fidelity block. The Product Reference
Sheet's job is to capture the actual product, not a stylized or
reinterpreted version. This block tells the model what NOT to do with
the source: don't redesign, don't reinterpret branding, don't beautify,
don't add styling. Frame composition discipline (no environmental
context, no props) is enforced here as part of the same content-fidelity
intent; the background/lighting constraints from Section 4 handle their
own scope and don't repeat here.
No redesign
No stylization
No brand reinterpretation
No added elements
No alternate branding
No additional logos
No text overlays
No props
No environment
No lighting effects
No smoothing or beautification
No beauty retouching
The Product Reference Sheet locks the product; narrative use of the
product happens in downstream scene prompts that call this sheet as a
reference.
Hat — Worked Example
A black baseball cap with an embroidered logo, set up as a Product
Reference Sheet:
Identity Lock — six-panel baseball cap, black cotton twill, curved
brim, adjustable strap closure, pre-curved crown, eyelets on each
panel
Branding Lock — embroidered wordmark centered on the front panel,
white thread, 4cm wide, positioned 3cm above the brim seam
Layout — front / back / left / right / top / bottom / 3-quarter,
plus macro close-ups of the embroidered wordmark, the panel seams,
and the strap closure
Background — #DCDCDC light gray, shadowless lighting, no
gradients, no reflections
Realism (Material Realism populated for embroidery):
Raised structure — raised thread structure, thread sitting proud of
the cotton twill base
Tight density — tight stitch density across each letterform
Visible direction — visible thread direction following each letter's
stroke path
Micro shadowing — micro shadowing in the stitch valleys between
thread rows
Surface compression following form — fabric compression around the
stitching where the thread tension pulls the twill in
Curvature integration — embroidery curvature following the
pre-curved front panel rather than sitting flat
Camera — Canon EOS R5, RF 100mm f/2.8L Macro IS USM, f/8, ISO 100
Restrictions — no redesign, no stylization, no brand
reinterpretation, no added elements, no smoothing or beautification
For leather / knit / satin / brushed metal / suede hero products, swap
the Material Realism block's embroidery vocabulary for the material's
own value set on each of the six axes — the scaffold is identical.
The Reference Sheet Family
Sheet
Locks
Lives in
Character
Identity, face, build, distinguishing marks
../higgsfield-soul/SKILL.md (Soul ID, character sheet creation)
Location
Architecture, light, color treatment, key environmental details
Build the sheets the project actually needs. A short ad with one character in
one outfit doesn't need an Outfit Sheet. A 15-shot music video with three
recurring camera moves and a tight color palette absolutely does.
Hero Frame
A Hero Frame is a key image you generate before committing to video — it defines the
visual tone, lighting, color, and composition of your sequence.
Why it matters: Generating a Hero Frame first costs ~1 credit. Adjusting the video
after generation costs full video credits. Get the look right on the image first.
Hero Frame workflow:
Describe your opening shot with full optical stack
Generate as image (Soul 2.0 or Nano Banana 2)
Review: Is the lighting right? Color? Character look?
Iterate on image until it's exactly right
Use as the first frame / style reference for video generation
Hero Frame as style lock:
Once you have a Hero Frame you love, use it as a reference upload for your first video
shot. Describe the match in the prompt as atmosphere and lighting language — not as
camera/genre instructions (those go in the UI settings):
The same overcast midday light as the reference image.
Muted, desaturated tones. The character enters from the left side of frame.
Higgsfield Popcorn — Storyboard Integration
Popcorn is Higgsfield's storyboard tool. Use it to plan a sequence visually before
committing to Cinema Studio generation.
Popcorn → Cinema Studio workflow:
Open Popcorn → describe your story / scene breakdown
Popcorn generates a visual storyboard with shot descriptions
Review shot order, camera angles, key moments
Export storyboard → import into Cinema Studio as shot list
Configure each shot: optical stack, Director Panel, Speed Ramp
Generate in Multi-Shot Manual mode using the storyboard as your script
When to use Popcorn first:
Any sequence longer than 3 shots
When you're not sure of the shot order
Client or collaborator needs to approve the structure before generation
Complex action sequences with many camera angles
Keyframe Interpolation — Start/End Frames
Define exactly where a shot begins and ends. Eliminates morphing artifacts that happen
when the model guesses the transition.
Setup:
Start Frame: Upload or generate the first frame of the shot
End Frame: Upload or generate the last frame of the shot
Cinema Studio interpolates the motion between them
Best uses:
Precise character entrances / exits
Object reveals with exact start and end position
Match cuts between shots (end frame of shot A = start frame of shot B)
Controlled camera moves where position must be exact
3D Mode — Gaussian Splatting
Cinema Studio can build a 3D version of any generated image using Gaussian splatting.
Once active, you can move inside the frame — shift perspective, orbit the scene, and find
a composition that didn't exist in the original 2D generation.
How it works:
Generate an image in Cinema Studio (any model, any optical stack)
Activate 3D Mode on the generated image
Cinema Studio reconstructs the scene as a 3D Gaussian splat
Use the virtual camera to move freely — orbit, push in, shift angle
Capture your preferred composition as a new frame
Use that frame as a start frame for video generation
When to use 3D Mode:
You love the generation but want a different camera angle
You need a reverse shot or over-the-shoulder from the same scene
You're building a virtual camera move through a static scene
You want to explore parallax and depth before committing to video
Workflow tip: Generate a Hero Frame → enter 3D Mode → find 2–3 different angles →
use each as start frames for different shots in your Multi-Shot Manual sequence. One
generation becomes multiple shots.
Grid Generation — Batch Variations
Instead of generating one image at a time, Cinema Studio can produce 2×2, 3×3, or 4×4
grids — up to 16 variations from a single generation, charged as one credit.
Grid sizes:
Grid
Variations
Cost
2×2
4 images
1 generation credit
3×3
9 images
1 generation credit
4×4
16 images
1 generation credit
When to use Grid Generation:
Exploring compositions — generate 16 options, pick the best one
Character sheet creation — multiple poses/angles in one pass
A/B testing visual direction before committing to video
Cost-efficient iteration — spend 1 credit, get up to 16 options
Workflow with grids:
Write your prompt + configure optical stack
Select grid size (2×2 through 4×4)
Generate — all variations appear grouped together
Select the best variation(s)
Use as Hero Frame, start frame, or enter 3D Mode on any individual result
Resolution Settings
Cinema Studio supports explicit resolution control for image generation:
Setting
Resolution
Best for
1K
1024px
Fast iteration, concept exploration
2K
2048px
Standard quality, most workflows
4K
4096px
Final delivery, print, hero assets
Rule of thumb: Start at 1K–2K for exploration and grid generation. Switch to 4K only
for your final selected composition. Higher resolution = longer generation time and more
credits.
One of Cinema Studio 2.5's most powerful workflows is the frame extraction loop. You
can extract the start frame or end frame from any generated video and feed it back into
the image workflow — creating an iterative creative cycle.
The loop:
1. GENERATE IMAGE → Hero Frame or grid selection
2. ANIMATE → Turn image into video (any model)
3. EXTRACT FRAME → Pull the start or end frame from the video
4. FEED BACK → Use extracted frame as a new start image
5. REPEAT → Animate again from the new starting point
Why this matters:
The end frame of a video often produces compositions you'd never prompt directly
Match-cutting between scenes using extracted frames as anchors
Object & Person Insertion
Cinema Studio 2.5 can insert characters and objects into a scene that weren't present
in the original start frame. This was essentially impossible before — the model would
ignore new subjects or break the scene trying to add them.
How it works:
Generate or upload your base scene (the environment)
In the prompt, describe the character or object entering the scene
Cinema Studio composites the new subject into the existing environment
The insertion respects lighting, perspective, and scale of the original scene
Best for:
Adding a character walking into an establishing shot
Placing props or vehicles into an existing environment
Building up scene complexity across multiple generations
"What if" exploration — same scene with different subjects
Prompt pattern for insertion:
[Describe the existing scene briefly]. A man in a dark overcoat enters from the left
side of frame, walking toward the center. He carries a briefcase.
Tip: The more specific you are about the entry point (left, right, background,
foreground) and the action, the cleaner the insertion.
Clustering — Automatic Generation Grouping
All generations from the same prompt are automatically clustered (grouped together) in
Cinema Studio's interface. This means:
Every variation from a grid generation stays in one visual group
Iterative re-generations of the same prompt are easy to compare
Long projects stay organized without manual folder management
You can quickly scan clusters to pick the best result from each prompt
Tip: Use clustering to your advantage — generate multiple variations of each key shot,
then pick winners from each cluster before assembling the final sequence.
Cinema Studio Output Format
Core rule: Everything selectable in the Higgsfield UI stays out of the prompt.
The prompt field is for scene description only — pure visual storytelling language.
What goes where
Belongs in UI (dropdowns)
Belongs in Prompt field
Camera body
Character appearance and action
Lens
Environment and atmosphere
Focal length
What happens in the scene
Aperture
Lighting and mood description
Genre
Emotion and tone
Director Panel movement
Everything the eye sees
Speed Ramp
Duration
IMAGE MODE Output Format (Cinema Studio 2.5 only)
━━━ UI SETTINGS (select in Higgsfield) ━━━━━━━━━━━━━━━━━━
Camera: [body name]
Lens: [lens name]
Focal: [focal length]
Aperture: [aperture]
↳ Why: [one sentence — what this stack gives the image and why]
━━━ PROMPT (paste into Cinema Studio) ━━━━━━━━━━━━━━━━━━━
[Scene description only. No camera/lens/aperture language.]
Image Mode example:
━━━ UI SETTINGS (select in Higgsfield) ━━━━━━━━━━━━━━━━━━
Camera: Grand Format 70mm Film
Lens: Classic Anamorphic
Focal: 50mm
Aperture: f/1.4
↳ Why: 70mm grain + anamorphic flare gives instant prestige cinema quality.
f/1.4 puts the harbour out of focus, keeping all weight on the detective.
━━━ PROMPT (paste into Cinema Studio) ━━━━━━━━━━━━━━━━━━━
A weathered detective stands at the edge of a rain-soaked harbour dock at night.
An old leather briefcase sits at his feet, open, papers scattered by the wind.
He stares at the horizon, collar turned up against the driving rain.
Harbour lights fracture on the black water below.
SINGLE SHOT Video Output Format (Cinema Studio 2.5)
━━━ UI SETTINGS (select in Higgsfield) ━━━━━━━━━━━━━━━━━━
Genre: [genre]
Movement: [Director Panel movement]
Speed Ramp: [mode]
Duration: [seconds]
━━━ PROMPT (paste into Cinema Studio) ━━━━━━━━━━━━━━━━━━━
[Scene description only. No movement, genre, speed ramp, or duration language.]
Single Shot example:
━━━ UI SETTINGS (select in Higgsfield) ━━━━━━━━━━━━━━━━━━
Genre: Suspense
Movement: Dolly Out
Speed Ramp: Slow Mo
Duration: 8s
━━━ PROMPT (paste into Cinema Studio) ━━━━━━━━━━━━━━━━━━━
A weathered detective stands at the edge of a rain-soaked harbour dock at night.
An old leather briefcase sits at his feet, open, papers scattered by the wind.
He stares at the horizon, collar turned up against the driving rain.
Harbour lights fracture on the black water below.
He reaches down and slowly closes the briefcase.
MULTI-SHOT AUTO Video Output Format (Cinema Studio 2.5)
Same structure as Single Shot — one UI settings block, one prompt. The user describes
the full scene in the prompt and Cinema Studio breaks it into shots automatically.
━━━ UI SETTINGS (select in Higgsfield) ━━━━━━━━━━━━━━━━━━
Genre: [genre]
Movement: [Director Panel movement — or Auto if varied]
Speed Ramp: [mode]
Duration: [total seconds]
━━━ PROMPT (paste into Cinema Studio) ━━━━━━━━━━━━━━━━━━━
[Full scene description. Let Cinema Studio break it into shots.
No movement, genre, speed ramp, or duration language in here.]
Multi-Shot Auto example:
━━━ UI SETTINGS (select in Higgsfield) ━━━━━━━━━━━━━━━━━━
Genre: Suspense
Movement: Auto
Speed Ramp: Linear
Duration: 15s
━━━ PROMPT (paste into Cinema Studio) ━━━━━━━━━━━━━━━━━━━
A weathered detective pushes open the door of a rain-soaked bar and steps inside.
He scans the room — empty except for a bartender polishing glasses at the far end.
He walks slowly to the bar and sits down. The bartender slides a drink without a word.
The detective picks it up, stares at his reflection in the mirror behind the bottles.
He sets it down without drinking.
MULTI-SHOT MANUAL Video Output Format (Cinema Studio 2.5)
One UI settings block per scene. One prompt per scene. Six scenes = six pairs.
Each scene is fully self-contained — the user configures and pastes them one at a time.
━━━ SCENE 1 — [short scene title] ━━━━━━━━━━━━━━━━━━━━━━━
UI SETTINGS
Genre: [genre]
Movement: [movement]
Speed Ramp: [mode]
Duration: [seconds]
PROMPT
[Scene 1 description only.]
━━━ SCENE 2 — [short scene title] ━━━━━━━━━━━━━━━━━━━━━━━
UI SETTINGS
Genre: [genre]
Movement: [movement]
Speed Ramp: [mode]
Duration: [seconds]
PROMPT
[Scene 2 description only.]
[...continue for each scene]
Multi-Shot Manual example — 3 scenes (same pattern scales to 6):
━━━ SCENE 1 — Arrival ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
UI SETTINGS
Genre: Suspense
Movement: Handheld
Speed Ramp: Linear
Duration: 5s
PROMPT
A weathered detective steps through the door of a dimly lit bar.
Rain drips from his coat. He pauses, eyes adjusting to the dark.
The bar is nearly empty. A jukebox plays quietly in the corner.
━━━ SCENE 2 — The Walk ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
UI SETTINGS
Genre: Suspense
Movement: Camera Follows
Speed Ramp: Linear
Duration: 4s
PROMPT
He walks slowly down the length of the bar, boots on wet floorboards.
A bartender watches without expression. One other patron doesn't look up.
He reaches the end stool and sits down deliberately.
━━━ SCENE 3 — The Mirror ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
UI SETTINGS
Genre: Suspense
Movement: Dolly In
Speed Ramp: Slow Mo
Duration: 6s
PROMPT
A glass of whiskey sits untouched on the bar in front of him.
He stares at his own reflection in the mirror behind the bottles.
His jaw tightens. He picks up the glass, holds it, sets it back down.
Cinema Studio 3.0 Output Formats
3.0 does NOT have: Camera body, Lens, Focal length, Aperture, Color grading, 3D Mode, Grid generation. Never include these in 3.0 output.
3.0 has: Genre (7: General, Action, Horror, Comedy, Noir, Drama, Epic), Director Panel, Speed Ramp (7: Auto, Slow-mo, Ramp Up, Flash In, Flash Out, Bullet Time, Hero Moment), Duration (up to 15s), Audio (On/Off native stereo), Smart shot control, 21:9 ultrawide.
Version guard — values that do NOT exist in 3.0 (never output these):
UI fields: Camera body, Lens, Focal length, Aperture, Color grading, 3D Mode, Grid generation
IMAGE MODE Output Format (Cinema Studio 3.0)
No optical stack in 3.0. Image output uses Soul Cast modes only.
━━━ UI SETTINGS (select in Higgsfield) ━━━━━━━━━━━━━━━━━━
Soul Cast Mode: [General / Character / Location]
Genre: [genre]
↳ Why: [one sentence — what this combination gives the image and why]
━━━ PROMPT (paste into Cinema Studio) ━━━━━━━━━━━━━━━━━━━
[Scene description only. No camera/lens/aperture language — these don't exist in 3.0.]
SINGLE SHOT / SMART Video Output Format (Cinema Studio 3.0)
━━━ UI SETTINGS (select in Higgsfield) ━━━━━━━━━━━━━━━━━━
Genre: [genre — General, Action, Horror, Comedy, Noir, Drama, or Epic]
Shot Mode: [Smart / Custom]
Movement: [Director Panel movement — or Smart for auto camera planning]
Speed Ramp: [Auto / Slow-mo / Ramp Up / Flash In / Flash Out / Bullet Time / Hero Moment]
Duration: [up to 15s]
Audio: [On / Off]
━━━ PROMPT (paste into Cinema Studio) ━━━━━━━━━━━━━━━━━━━
[Scene description only. Use @ to reference uploaded images/video/audio.
No movement, genre, speed ramp, or duration language in here.]
Single Shot 3.0 example:
━━━ UI SETTINGS (select in Higgsfield) ━━━━━━━━━━━━━━━━━━
Genre: Action
Shot Mode: Smart
Movement: Jib Down
Speed Ramp: Slow-mo
Duration: 5s
Audio: On
━━━ PROMPT (paste into Cinema Studio) ━━━━━━━━━━━━━━━━━━━
@CypressLookout packed with cars and people at night. @R34GTR parked
prominently in the center, @240SX and @AE86 visible nearby. Crowd
gathered between the cars, neon underglow reflecting on wet pavement.
City skyline glowing across the water in the distance. Engine noise,
crowd murmur, tension in the air.
MULTI-SHOT MANUAL Video Output Format (Cinema Studio 3.0)
Same per-scene structure as 2.5 but with 3.0 options. Up to 6 scenes, 15s max total.
━━━ SCENE 1 — [short scene title] ━━━━━━━━━━━━━━━━━━━━━━━
UI SETTINGS
Genre: [genre]
Movement: [movement]
Speed Ramp: [Auto / Slow-mo / Ramp Up / Flash In / Flash Out / Bullet Time / Hero Moment]
Duration: [seconds]
Audio: [On / Off]
PROMPT
[Scene 1 description only. Use @ for references.]
⚠ Prompt Character Limit — 512 Characters
Cinema Studio has a hard 512-character limit on the prompt field in both 2.5 and 3.0. However, how those characters are consumed differs by version.
Cinema Studio 2.5 Character Budget
2.5 uses @ Element chips (Characters, Locations, Props). Each chip consumes roughly 80–100 hidden characters for its internal ID and reference metadata.
Max 2 @ Element tags per prompt — each tag eats ~80–100 hidden chars
Keep visible text under ~250 characters when using 2 @ tags
Keep visible text under ~350 characters when using 1 @ tag
Keep visible text under ~450 characters when using 0 @ tags
If a prompt is rejected, remove words — never assume it's a bug
Cinema Studio 3.0 Character Budget
3.0 uses @ references (uploaded images, video clips, audio clips — up to 12 total). The @ reference system is structurally different from 2.5's Element chips. References are attached as media inputs rather than inline metadata, so they may consume less hidden character space than 2.5's Element chips.
The 512-character hard limit still applies
With @ references attached, keep visible text under ~350–400 characters as a safe starting point
With no @ references, you can use closer to the full ~450–500 characters
If a prompt is rejected for length, trim visible text first — do not assume it's a bug
3.0 prompts can be more descriptive than 2.5 prompts since @ references leave more room for text
@ Element Persistence Across Scenes
@ Elements added in Scene 1 persist across all subsequent scenes in Multi-Shot Manual.
You do NOT need to re-add @ tags in every scene prompt.
Recommended pattern for multi-character sequences:
Scene 1: Use @ Elements to establish all characters (e.g. @HERO and @badguy3)
Scenes 2–6: Reference characters by visual description only (e.g. "the man in
the leather jacket") — the Reference Anchor keeps their appearance locked
Why this works: The @ Elements in Scene 1 lock character geometry into the Reference
Anchor. Subsequent scenes inherit that lock automatically. Using @ tags again in later
scenes forces the model to re-process the reference data, which competes with action
prompts and causes issues like character swaps, broken choreography, and missed actions.
Prompting Best Practices
For the full Pre-Prompt Checklist, one-action-per-scene rule, and fast motion trick, see higgsfield-prompt. Cinema Studio–specific additions:
Select camera movement from Director Panel presets (not prompt text)
Select genre from Cinema Studio genre options (not prompt text)
Use @ Elements for character identity; keep prompts to action and scene description only
Fight Scene & Action Design Rules (Tested)
Two-character fight sequences are among the hardest things to generate in Cinema Studio.
These rules come from extensive real-world testing.
Prop-based combat (trays, carts, objects as weapons) ✗
Grappling at close range (often renders as embracing) ✗
Character Swap Problem
When two @ Elements are used in the same action prompt, the AI frequently swaps
which character is the hero and which is the villain. The first character mentioned
tends to get assigned the "protagonist" role regardless of the prompt's intent.
Fixes:
Use @ Elements only in static/slow scenes (standoff, pinned, walk away)
Use plain text for action scenes where choreography matters
If using @ tags, always put the hero character first in the prompt
End each scene by stating who is where, to anchor positions
Fight Sequence Template (Multi-Shot Manual)
Scene 1 (@ Elements) — Standoff: Lock in characters, tension build
Scene 2 (Plain text) — First exchange: Describe fighting energy, not specific moves
Scene 3 (Plain text) — Escalation: Bodies slamming, intensity rising
Scene 4 (@ Elements) — Close-up moment: Pin against wall, intense stare
Scene 5 (Plain text) — The finish: Someone goes down
Scene 6 (@ Elements) — Resolution: Hero walks away from camera
Alternate @ Element scenes (for character faces) with plain text scenes (for action).
The viewer's brain fills in character continuity between cuts — that's how real film
editing works.
Action Design Around AI Strengths
The Fight Scene rules above describe what the model can and cannot render at
any given action moment. The rules below describe how to design the action
itself so the model never has to render the things it fails at. These rules
are adapted from the Mr. Core methodology — a long-form workflow document by a
working creator on the platform — and apply to all action sequences, not just
fights.
Choose locations the model renders forgivingly. Some spaces hide drift;
others expose it. Design action around the forgiving ones:
Forgiving: circular arenas (camera can rotate without a continuity
reference), vague ruins (no architectural lines for the audience to audit),
repeating textures (a forest, a dune field, a colonnade — drift between
frames reads as parallax, not error), fog / dust / smoke environments
(visual noise covers most artifacts).
Punishing: familiar landmarks (the audience knows the building's real
geometry), geometric architecture with strong sight lines (any drift in a
hallway or grid reads instantly), public spaces with text or signage
(lettering drifts across frames), continuous reflective surfaces (mirror or
glass continuity is brutal).
Design choreography around what the model can render. Don't write action
the model fails at and try to compensate with prompt language. Cut the action
to the model's strengths up front:
Standoffs, near-misses, the moment before contact, the aftermath of contact,
bodies in motion through a space — all render reliably.
If the script calls for the second category, restage it as a sequence of
shots from the first category. The cut between shots does the work the
model can't do in a single take.
One transformation per shot. Plan exactly one state change per generation —
one location to one location, one emotional beat to the next, one action to
its consequence. Multi-beat shots ("she enters, sees the body, draws her
weapon, advances on the door") force the model to interpolate between two or
three different intents and the result drifts on at least one of them. Cut the
shot at the first beat. Generate the second as its own shot.
Single-generation strategy. Plan the shot well enough that the first
generation is the work. Generate-and-pick-from-twenty is a budgeting failure,
not a workflow — every additional generation is paying twice for prep you
should have done up front. Treat each generation as the deliverable.
Model Selection for Cinema Studio
Different models perform differently inside Cinema Studio's environment:
Use case
Recommended model
Character-driven drama sequence
Kling 3.0
Clone character from reference footage
Kling 3.0 Omni
Epic scale / action multi-shot
Sora 2
Artistic / stylized sequence
Wan 2.6
Nature / environment sequence
Veo 3 / Veo 3.1
Fast iteration on sequence
Kling 2.5 Turbo
Hero Frame image generation
Soul 2.0 / Nano Banana 2
Quick Decision — Cinema Studio vs Other Tools
Need
Use
Single clip, no sequence
Standard generation
2–6 shot sequence, character consistent
Cinema Studio — Multi-Shot Manual
Sequence but don't need per-shot control
Cinema Studio — Multi-Shot Auto
Don't know shot order yet
Popcorn first → Cinema Studio
Need motion graphics / text animation
Vibe Motion (not Cinema Studio)
Edit existing footage
Kling O1 Video Edit (not Cinema Studio)
Just need audio added to a clip
Lipsync Studio / Kling 3.0
Identity vs. Motion: In Cinema Studio, identity goes in the @ Element definition (or
Soul Cast parameters); motion goes in the prompt field. Never put face/clothing descriptors
in the prompt when @ Elements are active. See higgsfield-prompt and higgsfield-soul
for the full separation rule.
Negative constraints: For Cinema Studio–specific artifacts (prompt rejected, @ Element
character swap, 3D Mode holes, optical stack mismatch) and all general artifacts, see
../shared/negative-constraints.md.
Cinema Studio 3.0 (Business/Team Plan)
Plan requirement: Cinema Studio 3.0 is available exclusively on Business and Team plans. Free and individual plan users should use Cinema Studio 2.5, which remains fully supported.
Version toggle: Cinema Studio 2.5 and 3.0 coexist on the platform. Switch between them using the version selector in the upper-right corner of the Cinema Studio UI.
What's Different in 3.0
Cinema Studio 3.0 is a separate generation engine from 2.5. Key differences:
Higher max duration: 15s (vs 2.5's 12s)
Lower video resolution (for now): Video capped at 720p (vs 2.5's 1080p); Image up to 4K in Character/Location modes (General mode capped at 2K)
Native audio: Audio generated simultaneously with video via unified multimodal architecture — dual-channel stereo, not post-processed
Smart shot control: Model auto-plans camera language based on genre and scene description
Ultrawide aspect ratio: 21:9 added (not available in 2.5)
No optical physics engine: 2.5's camera body + lens stack is not available in 3.0
No color grading suite: 2.5's built-in grading is not available in 3.0
No 3D Mode / Grid Generation: These 2.5 features are not available in 3.0
Resolution note: Cinema Studio 3.0 video resolution (720p) may increase. For 1080p video, use Cinema Studio 2.5. Image resolution in 3.0 varies by mode: Character/Location support 4K, General is capped at 2K.
Cinema Studio 3.0 Quick Specs
See the comparison table at the top for full 2.5 vs 3.0 differences. Key 3.0-specific details:
Video: up to 15s, 720p (may increase), 48 credits/generation
Image (Soul Cast 3.0): up to 4K (Character/Location) · 2K (General), 0.125 credits
Genres (7): General, Action, Horror, Comedy, Noir, Drama, Epic
Speed Ramp (7): Auto, Slow-mo, Ramp Up, Flash In, Flash Out, Bullet Time, Hero Moment
Audio: On/Off — natively generated alongside video (dual-channel stereo)
Shot Control: Smart (auto camera planning) or Custom multi-shot (up to 6 scenes, 15s total)
Image mode: Cinema Studio 3.0 image mode uses Soul Cinema as the default model — see § Image Mode in the Cinema Studio 3.5 section below for the full image-mode model picker (Cinematic models + Featured models), which is shared between 3.0 and 3.5.
Input Limits (@ References)
Type
Max Count
Formats
Size/Duration Limit
Images
9
jpeg, png, webp, bmp
—
Video clips
3
mp4, mov
Combined ≤15s total
Audio clips
3
mp3, wav
Combined ≤15s total
Total files
≤12
@ Reference Patterns for Cinema Studio 3.0
Character identity (first frame):
@Image1 as the main character. She walks through the market, picking up fruit and examining it closely.
Environment / last frame:
@Image1 as the starting environment. @Image2 as the destination. Camera tracks through a doorway transitioning from the first space to the second.
Motion reference / camera cloning:
Match the camera movement from @Video1. A dancer performs on a rooftop at sunset, wind catching her dress.
Audio reference (BGM / dialogue / tone):
Audio @Audio1 plays exactly as uploaded from 0s to end. Do not modify or replace the audio content. Voiceover tone references @Video1.
Multi-image spatial mapping:
@Image1 as first frame, @Image2 as top of frame, @Image3 as left side. Camera slowly pans right, revealing the full scene.
Video extension:
Extend @Video1 by 5s. The character continues walking, reaching the edge of the cliff and looking out over the valley.
Ad recreation:
Mimic @Video1's shot design, pacing, and transitions. Replace all products with @Image1. Match the lighting and camera angles.
Outfit transformation:
@Image1 as the character in casual clothes. @Image2 as the same character in formal attire. A quick-cut transformation sequence with fabric particles.
One-shot continuity:
@Image1 first frame, @Image2 midpoint, @Image3 final frame. No cuts throughout, one continuous shot tracking the subject across all three compositions.
When to Use 3.0 vs 2.5
Need
Recommendation
Highest video resolution (1080p)
Cinema Studio 2.5
4K images (Character/Location)
Cinema Studio 3.0 (Business/Team)
Longer duration (up to 15s)
Cinema Studio 3.0 (Business/Team)
Native audio with video
Cinema Studio 3.0 (Business/Team)
Optical physics (lens, sensor)
Cinema Studio 2.5
Color grading suite
Cinema Studio 2.5
3D Mode / Grid Generation
Cinema Studio 2.5
Ultrawide 21:9 aspect ratio
Cinema Studio 3.0 (Business/Team)
Smart auto-camera planning
Cinema Studio 3.0 (Business/Team)
Free/Individual plan
Cinema Studio 2.5
Cinema Studio 3.5
Cinema Studio 3.5 is the latest Cinema Studio version, sitting alongside 2.5 and 3.0 in the model selector — all three coexist on the platform, version is user-selected, and there is no auto-routing between them. 3.5 is positioned as a simplified creative surface over a deep manual layer: by default everything routes to Auto and the engine handles selection, with manual overrides available at any point in the stack.
What's new vs 3.0:
Three-pill main UI — Genre / Style / Camera collapsed into three primary pills on the main surface
Restored optical physics — camera body / lens / focal length / aperture surface is back, accessible via the Camera Settings panel (different vocabulary from 2.5 — see below)
Style Settings panel — three preset axes (Color Palette / Lighting / Camera Moveset Style) plus a free-form Manual Style mode
AI director toggle — visible in the bottom toolbar (function not yet documented; see § AI Director Toggle)
Operating philosophy: Auto by default, manual with reason. Each pill defaults to Auto and can be overridden — but only override when there is a specific creative reason to do so.
The Three-Pill Main Surface
3.5 collapses creative control into three pills on the main UI:
Pill
What it controls
Surface
Genre
Manual genre catalog
Single-select picker
Style
Color Palette / Lighting / Camera Moveset Style + Manual Style mode
Three-axis preset stacking, or free-form prompt
Camera
Camera Body / Lens / Focal Length / Aperture
Four-axis panel
Each pill defaults to Auto. Open the pill to override.
Cinema Studio 3.5 Genres
Confirmed from the UI: General · Action · Horror · Comedy · Noir · Drama · Epic (7 genres).
The Genre picker is browsable in the UI and may surface additional genres not enumerated here — direct readers to the live UI for the current full list.
⚠ Genre catalog scope: The seven listed genres are confirmed from a 2026-04-25 UI screenshot, which showed scroll arrows above and below the visible list — additional genres exist in the picker but were not confirmed by name and are not enumerated here. Do not invent genre names. When the user names a genre not in the seven, ask them to confirm by checking the UI rather than assuming.
Style Settings — Three Operating Modes
Diese SKILL.md ist sehr gross, daher zeigt SkillsMP hier nur den ersten Abschnitt.Auf GitHub ansehen
