A Practical Guide to Prompting with Veo 3
Achieving Cinematic Consistency
This guide outlines a robust workflow for generating high-quality, narratively consistent video clips using Veo 3. The core principle is treating each 8-second prompt not as an isolated instruction, but as a self-contained universe that carries all the necessary information to maintain continuity with the clips before and after it.
🎥Watch the Full Tutorial
Prefer to learn by watching? Check out the complete video walkthrough of these Veo 3 prompting techniques on YouTube.
Watch on YouTubePart 1: The Foundation – The Master Sheet Workflow
Before writing a single video prompt, you must establish your project's “source of truth.” This involves creating detailed, text-based reference sheets for every consistent element in your story.
1Create Character Master Sheets
For every recurring character, create a detailed paragraph that you will copy and paste. The model has no memory, so this description must appear in every single prompt the character is in.
• Visuals:
Be hyper-specific. Go beyond “a man in a jacket” to include age, specific facial features, hair style and color, skin details, posture, and exact clothing, including textures and accessories.
Example: Dr. Kaelen Reyes, a woman in her late 30s with dark, curious eyes and black hair tied back practically in a simple ponytail. She has a thoughtful, academic face. She wears a lightweight, dark grey Earth Federation field jumpsuit with a visible mission patch on the shoulder.
• Vocal Profile:
Describe the character's voice clearly to guide the audio generation for any dialogue.
Example: Her vocal profile is a calm, thoughtful, mid-range voice with a standard American accent; she speaks with clarity and deliberation.
2Create Location Master Sheets
Similarly, create a detailed paragraph for each recurring location. This description establishes the atmosphere, lighting, and key features that must remain consistent.
• Be Evocative:
Don't just name the place. Describe the mood, the materials, the lighting, and even the smell or feeling of the air.
Example (Interior):
The interior of the Earth Federation Lander Cockpit is a sterile and silent bridge, where the air smells of filtered oxygen and clean metal. The design is minimalist and functional, with two pilot seats made of dark memory-gel polymer... all light originating from the glowing blue holographic displays...
Example (Exterior):
The alien forest of Planet Aethel at dawn. This is a world of breathtaking, wild beauty... The air is thick with mist. Towering, gnarled, ancient trees are draped in thick, phosphorescent moss that glows faintly.
Part 2: The Core Technique – The “Self-Contained Universe” Prompt
Every prompt you write must follow a strict cinematic blueprint and contain all the information needed for that specific shot, assuming the model knows nothing else.
The Blueprint Structure:
Subject:
- • Clearly state the main focus of the shot.
- • If a character is the subject, paste their entire Character Master Sheet here. If two characters are present, paste both full descriptions.
Example: Dr. Kaelen Reyes, a woman in her late 30s... [full description]. AND Commander Eva Rostova, a woman in her 40s... [full description].
Context:
- • This is the most critical field for visual consistency. Fully describe the environment.
- • Paste the entire Location Master Sheet for the primary location.
- • The “Interior + Exterior” Rule: If the shot is an interior with a window or viewport, you must describe both the interior and the visible exterior. Paste the full master sheets for both.
Example: The scene is set within the interior of the Earth Federation Lander Cockpit... [full cockpit description]. Through the large, reinforced forward viewport, the alien forest of Planet Aethel is visible at dawn... [full forest description].
Action:
- • Describe precisely what the subject is doing within the 8-second clip. Be specific and evocative.
- • Separate description from action. The
Subject
field describes who they are; theAction
field describes what they do.
Example: Instead of “A skilled weaver works,” use “Her nimble hands move with practiced grace, passing a shuttle back and forth, weaving a cloth threaded with glowing fibers.”
Style/Composition:
- • Define the visual language of your project. Be consistent.
- • Specify shot type (
Cinematic close-up
,Wide establishing shot
), lighting (high-contrast
,soft natural light
), lens effects (anamorphic
,shallow depth of field
), and overall mood (gritty realism
,atmospheric
).
Example: Cinematic extreme close-up shot, framed so tightly that only her eyes are visible. The focus is razor-sharp on her irises. The lighting is a mix of the cool blue light from the ship's consoles and the soft, grey light of dawn.
Camera Motion:
- • Explicitly state the camera movement, even if it's static. This removes ambiguity.
Examples: Static shot.
,Slow push-in on her face.
,Smooth pan right, following her gaze.
Ambiance/Audio:
- • Diegetic Sound Only: This is crucial. Describe only the sounds that exist within the world of the scene. Do not mention music or narration, as those are post-production layers for different models.
- • Be specific. Instead of “noise,” use “the rhythmic clang of a hammer,” “the low hum of life support,” or “the sharp click of a plastic cover being lifted.”
Dialogue:
- • Keep it short and natural to fit within the 8-second clip.
- • Assign lines using physical descriptions, not names, for maximum clarity.
Example: The woman with the blonde bun, her voice a crisp, commanding alto with a clipped, Pan-Slavic accent, says: “Miracles aren't in my mission parameters.”
Complete Example: Putting It All Together
Here's a full example demonstrating how to apply the master sheet workflow and self-contained universe approach to create a single 8-second video prompt.
Master Sheets (Created First)
Character Master Sheet
Dr. Kaelen Reyes: A woman in her late 30s with dark, curious eyes and black hair tied back practically in a simple ponytail. She has a thoughtful, academic face with defined cheekbones and a small scar above her left eyebrow from a childhood accident. She wears a lightweight, dark grey Earth Federation field jumpsuit with a visible mission patch on the shoulder, practical cargo pockets, and reinforced knees. Her posture is confident but contemplative. Her vocal profile is a calm, thoughtful, mid-range voice with a standard American accent; she speaks with clarity and deliberation.
Location Master Sheet
Earth Federation Lander Cockpit: The interior is a sterile and silent bridge, where the air smells of filtered oxygen and clean metal. The design is minimalist and functional, with two pilot seats made of dark memory-gel polymer that conform to the occupant's body. The control surfaces are smooth, touch-sensitive glass panels that glow with soft blue holographic displays showing navigation data, system diagnostics, and communication channels. All light originates from these glowing blue displays, casting cool shadows across the occupants' faces. The walls are brushed titanium with subtle panel lines. Through the large, reinforced forward viewport, space is visible as an infinite black void dotted with distant stars.
The Complete Prompt
Subject:
Dr. Kaelen Reyes, a woman in her late 30s with dark, curious eyes and black hair tied back practically in a simple ponytail. She has a thoughtful, academic face with defined cheekbones and a small scar above her left eyebrow from a childhood accident. She wears a lightweight, dark grey Earth Federation field jumpsuit with a visible mission patch on the shoulder, practical cargo pockets, and reinforced knees. Her posture is confident but contemplative. Her vocal profile is a calm, thoughtful, mid-range voice with a standard American accent; she speaks with clarity and deliberation.
Context:
The scene is set within the Earth Federation Lander Cockpit. The interior is a sterile and silent bridge, where the air smells of filtered oxygen and clean metal. The design is minimalist and functional, with two pilot seats made of dark memory-gel polymer that conform to the occupant's body. The control surfaces are smooth, touch-sensitive glass panels that glow with soft blue holographic displays showing navigation data, system diagnostics, and communication channels. All light originates from these glowing blue displays, casting cool shadows across the occupants' faces. The walls are brushed titanium with subtle panel lines. Through the large, reinforced forward viewport, space is visible as an infinite black void dotted with distant stars.
Action:
Dr. Reyes sits in the pilot's seat, her fingers dancing across the holographic interface with practiced precision. She reaches forward and taps a sequence of glowing icons, her brow furrowed in concentration. A new display materializes before her, showing a three-dimensional star map. She leans back slightly, studying the data, then reaches up to adjust a control above her head with deliberate, careful movements.
Style/Composition:
Cinematic medium shot, framed from her waist up, showing both her hands and face clearly. The shot is static with high-contrast lighting between the blue glow of the displays and the dark shadows. Shallow depth of field keeps Dr. Reyes in sharp focus while the background controls are slightly soft. The mood is serious and contemplative, with clean, minimalist sci-fi aesthetics.
Camera Motion:
Static shot. The camera remains perfectly still to emphasize the precision and focus of her work.
Ambiance/Audio:
The low, constant hum of life support systems. Soft electronic beeps as she touches the interface controls. The subtle whoosh of the air recycling system. No music or narration.
Dialogue:
The woman in the pilot's seat, her voice a calm, thoughtful, mid-range voice with a standard American accent, speaks with clarity and deliberation: “Navigation systems show three possible routes. Calculating fuel efficiency now.”
đź’ˇKey Takeaways from This Example
- •Complete Character Description: Every detail about Dr. Reyes is included, from physical appearance to vocal qualities
- •Immersive Environment: The cockpit description includes sensory details (smell, materials, lighting)
- •Specific Actions: Rather than “she works,” we see exactly what she does with her hands and body
- •Clear Technical Specs: Shot type, lighting, and camera movement are explicitly defined
- •Diegetic Audio Only: Sound effects exist within the story world, no external music
- •Self-Contained: This prompt could generate a consistent clip without any other context
Want More AI Video Creation Tips?
Join the Robin Blocks mailing list for exclusive guides, techniques, and insights on AI video generation and creative technology.
By subscribing, you agree to receive emails from Robin Blocks. You can unsubscribe at any time.
Ready to Master Veo 3 Prompting?
Start implementing these techniques in your next video project and achieve the cinematic consistency you've been looking for.
Back to Robin Blocks