AI Prompt Engineering • Published August 23, 2026 • 17 min read

Mastering Prompts for Nano Banana Video Generator: Camera Physics, Lighting & Motion Direction

Master prompt engineering for the Nano Banana Video Generator. Discover camera motion tokens, fluid physics prompts, lighting direction, and 10 production templates.

Mastering Prompts for Nano Banana Video Generator: Camera Physics, Lighting & Motion Direction
A prompt engineering masterclass crafted specifically for the Nano Banana Video Generator. Learn directional tokens, cinematic lighting cues, continuous motion verbs, and multi-shot storyboard prompts.
Film director camera rig and lighting setup demonstrating cinematic direction
Figure 1: Translating professional Hollywood cinematography rigs into neural prompt tokens

Introduction: Thinking in Motion, Time, and Geometry

Most creators entering AI video generation make a fundamental mistake: they treat video prompts like static image prompts. They paste a 100-word paragraph describing a still photograph, press generate, and wonder why the camera drifts aimlessly or the character's limbs dissolve into gelatin.

The Nano Banana Video Generator is an intelligent temporal physics simulation engine. To direct it effectively, you must think not as an illustrator, but as a Cinematographer, Stunt Coordinator, and Director of Photography (DP).

In this masterclass, we break down the definitive prompt grammar for Nano Banana, unpack its camera token lexicon, analyze physical motion dynamics, and supply 10 production-ready templates you can deploy immediately.


1. The Anatomy of a High-Coherence Video Prompt

Every production prompt for the Nano Banana Video Generator must contain 6 modular components arranged in logical order:

+------------------------------------------------------------------------------------+
|                  THE 6-TIER VIDEO PROMPT ARCHITECTURE MATRIX                      |
+------------------------------------------------------------------------------------+
| Tier 1: SUBJECT & POSE      | Physical identity, initial stance, and attire        |
| Tier 2: ACTION & KINEMATICS | Continuous active verbs describing mechanical motion |
| Tier 3: CAMERA RIG & LENS   | Rig type, focal length, trajectory tag, and speed    |
| Tier 4: LIGHTING & SHADOWS  | Light source orientation, temperature, and bounce    |
| Tier 5: ATMOSPHERIC PHYSICS | Mist, sparks, rain, cloth dynamics, and fluids       |
| Tier 6: TEMPORAL METRICS    | Motion scale token, frame cadence, and aspect ratio  |
+------------------------------------------------------------------------------------+

Example of Complete Tier Assembly:

(Tier 1) A samurai warrior in weathered black-lacquered armor (Tier 2) drawing his steel katana in one swift arc through the misty forest air, (Tier 3) [camera:tracking_low_angle_right] on 35mm anamorphic lens with shallow depth of field, (Tier 4) cold moonlight filtering through pine branches casting sharp silver rim highlights on the polished blade, (Tier 5) raindrops slicing through the air with microscopic water splash on armor plates, (Tier 6) [motion_scale:0.65], [speed:cinematic_24fps], [temporal_consistency:high].

2. The Nano Banana Camera Lexicon & Token Dictionary

Nano Banana reads explicit camera bracket tokens to lock virtual camera rigs onto calculated trajectories:

+----------------------------------------------------------------------------------+
|                    NANO BANANA CAMERA RIG TOKEN DICTIONARY                       |
+----------------------------------------------------------------------------------+
| Tag Syntax                      | Motion Behavior                                |
+---------------------------------+------------------------------------------------+
| [camera:dolly_in_slow]          | Moves camera forward toward subject (reveal)   |
| [camera:dolly_out_smooth]       | Moves camera backward away from subject (scale)|
| [camera:tracking_lateral_left]  | Tracks parallel to subject moving sideways     |
| [camera:tracking_lateral_right] | Tracks parallel to subject moving right        |
| [camera:orbit_180_clockwise]    | Semi-circular sweep around subject front-to-back|
| [camera:pedestal_up_reveal]     | Moves entire camera straight up vertically     |
| [camera:pedestal_down_ground]   | Lowers camera straight down to ground level    |
| [camera:fpv_drone_dive]         | Rapid dynamic aerial dive through environment  |
| [camera:dutch_angle_tilt]       | Canted horizon angle creating psychological ten|
| [camera:crane_boom_sweep]       | Arc motion from high altitude down to eye-level|
+----------------------------------------------------------------------------------+

Speed & Velocity Modifiers:

  • [motion_speed:gentle]: Best for ambient backgrounds, nature, and emotional portraits.
  • [motion_speed:moderate]: Standard dialogue, walking, and corporate commercials.
  • [motion_speed:dynamic]: Sports, automotive, dance, and action sequences.
  • [motion_speed:hyper_velocity]: Sci-fi hyperspace, explosions, and drone acrobatics.

3. Directing Complex Physics & Environmental Dynamics

Neural video models excel when given environmental particles that move naturally with ambient airflow. Particles provide dense spatial reference points for temporal cross-attention layers.

+-------------------------------------------------------------------------------+
|                       ENVIRONMENTAL PHYSICS PROMPT TOKENS                     |
+-------------------------------------------------------------------------------+
| Fluid Dynamics    | "viscous amber honey swirling into tea with micro-eddies" |
| Aerodynamics      | "silk trench coat billowing realistically in 30kt wind"   |
| Particulate Cloud | "drifting golden dust motes floating through god rays"    |
| Combustion & Fire | "sparks crackling and rising upward in turbulent drafts"  |
| Weather Systems   | "heavy monsoon rain bouncing violently off asphalt"       |
+-------------------------------------------------------------------------------+

4. 10 Production-Ready Master Video Prompt Templates

1. Cyberpunk Neon Street Drone Sweep

Cinematic action shot of a sleek armored futuristic motorcycle speeding through a rain-slicked Tokyo alleyway at night. [camera:fpv_drone_dive, motion_speed:dynamic], camera banking tightly around towering holographic billboards, swooping down to match the motorcycle speed inches above the ground. Bright neon magenta and cyan reflections shimmering across wet pavement puddles, water rooster tail spraying from rear tire. Anamorphic lens flare, photorealistic physics, 24fps cinematic film look.

2. Slow-Motion Luxury Watch Craftsmanship

Extreme macro slow-motion commercial of a master watchmaker placing a glowing ruby gear into an open luxury tourbillon movement with fine titanium tweezers. [camera:dolly_in_slow, motion_speed:gentle], 100mm macro lens, razor-thin depth of field. Intricate polished gold gears oscillating with mechanical precision, fine microscopic metallic brush textures, soft warm studio rim lighting. 60fps high-frame-rate fluidity, crystal-clear mechanical detail.

3. Desert Sandstorm Atmospheric Flyover

Sweeping cinematic aerial view of a lone traveler in crimson robes riding a camel across towering golden sand dunes during a raging desert windstorm. [camera:crane_boom_sweep, motion_speed:moderate], camera ascending from dune level to reveal vast desert expanse. Heavy sand particles whipping through air in turbulent swirls, sunlight piercing through dust haze in dramatic golden volumetric shafts, fluid cloth billowing dynamics.

4. Deep Sea Bioluminescent Encounter

Underwater documentary footage of a giant translucent bioluminescent jellyfish pulsating smoothly through dark oceanic depths. [camera:orbit_180_clockwise, motion_speed:gentle], camera floating weightlessly around undulating glowing tentacles. Soft pulsating azure and violet light illuminating floating marine snow particulates in deep ocean abyss, natural fluid water resistance physics, National Geographic 4K cinematography.

5. Architectural Modernist Home Sunrise

Architectural interior cinematic glide through a minimalist glass-walled Scandinavian living room overlooking a misty pine forest at dawn. [camera:tracking_lateral_right, motion_speed:gentle], smooth motorized slider movement past concrete fireplace toward floor-to-ceiling windows. Early morning golden sun cresting horizon, long soft shadows sliding across polished oak hardwood floors, steam rising gently from a ceramic mug on a marble table.

6. High-Energy Action Parkour Chase

First-person FPV action sequence of a parkour athlete sprinting across Parisian rooftop ledges, vaulting over ventilation shafts, and leaping across a narrow alleyway gap. [camera:fpv_drone_dive, motion_speed:dynamic], rapid kinetic camera movements, athletic arm and leg movements in frame. Wind rushing, pigeons scattering in sudden flight, dynamic motion blur, ultra-crisp kinetic impact physics.

7. Gourmet Espresso Pouring Commercial

Commercial macro shot of rich dark espresso pouring steadily into a clear double-walled glass cup, forming thick golden-brown crema layers on the surface. [camera:pedestal_up_reveal, motion_speed:gentle], 90mm macro lens with soft backlight. Rich liquid viscosity, delicate steam swirling upward in soft illumination, gentle micro-bubbles forming along glass perimeter, commercial advertising lighting perfection.

8. Sci-Fi Spacecraft Orbital Docking

Epic sci-fi establishing shot of an interstellar cargo starship executing a docking maneuver with a rotating space station above Earth. [camera:dolly_out_smooth, motion_speed:gentle], camera pulling back to reveal the curvature of Earth with glowing atmospheric blue rim below. RCS thrusters firing short bursts of white gas, solar array panels reflecting harsh unfiltered sunlight, zero-gravity physics, Interstellar IMAX visual scale.

9. High-Fashion Runway Model Walk

High-fashion runway tracking shot of a model walking with confident rhythmic stride down a mirrored runway in a sculptural iridescent metallic dress. [camera:tracking_lateral_left, motion_speed:moderate], 85mm portrait prime lens, smooth motorized dolly track. Runway spotlight beams following movement, iridescent fabric shifting hues from emerald to violet with each step, camera shutter flashes popping in dark audience background.

10. Wildlife Arctic Wolf Chase Sequence

Wildlife documentary tracking shot of a white Arctic wolf sprinting across pristine powdery snow in a Siberian blizzard. [camera:tracking_lateral_right, motion_speed:dynamic], low-angle camera matching the wolf's stride. Powdery snow flying upward from paws with realistic particle dispersion, wolf's fur rippling in freezing wind, intense focused amber eyes locked forward, BBC Planet Earth cinematic style.

5. The Universal Video Negative Prompt

To suppress common video diffusion defects, apply this calibrated negative prompt template:

morphing faces, shifting identity, extra limbs, mutated hands, duplicated bodies, floating disconnected heads, camera jitter, strobe flicker, sudden scene cut, low frame rate, choppy animation, blurry texture loss, unnatural physics, rubbery limbs, melting objects, pixelation artifacts, watermark, timestamp text.

Summary & Creative Workflow

Writing prompts for the Nano Banana Video Generator is an art form rooted in directorial clarity. By pairing active motion verbs, explicit camera trajectory brackets, atmospheric particle dynamics, and realistic temporal speeds, you unlock Hollywood-grade AI video clips on demand.

Ready to build your next video storyboard? Generate customized camera prompts with our free AI YouTube & Video Prompt Generator and polish your ideas with our AI Prompt Improver.

Storyboard sequence showing multi-shot video consistency across prompt chains
Figure 2: Multi-shot storyboard continuity pipeline maintaining character and lighting tokens

Frequently Asked Questions

Q1. Why do normal image prompts fail when used in the Nano Banana Video Generator?

Image prompts rely heavily on static descriptive adjectives (e.g. "photorealistic, sharp focus, 8k"). Video diffusion models require temporal motion vectors, continuous action verbs, speed coefficients, and directional camera commands. Static adjectives often cause the video model to freeze motion or create flickering micro-shifts.

Q2. How do I prompt for continuous, unbroken character movement without random cuts?

Use progressive kinematic verbs (e.g. "stepping steadily forward across the bridge", "raising coffee cup to lips in one continuous motion") combined with explicit camera tracking tags like [camera:tracking_front_medium] and low-to-medium motion scale [motion_scale:0.45].

Q3. What is the best way to handle dramatic lighting shifts (like sunrise or lightning flashes)?

Explicitly describe the temporal light transition in the prompt: [lighting_transition:dim_amber_to_bright_white_flash] or "morning sun gradually cresting the horizon over 4 seconds, cast shadows shortening as golden rays flood the interior".

Q4. Can I generate slow-motion or high-speed hyper-lapse videos in Nano Banana?

Yes. Use tags like [speed:slow_motion_120fps] for fluid liquid and action shots, or [speed:hyper_lapse_smooth] for clouds moving over skylines and fast urban traffic streaks.

Generate Production Video Prompts in Seconds

Use our AI Video & YouTube Prompt Generator to automatically apply camera movement tags, lighting ratios, and motion physics tokens.

Open Video Prompt Generator