How you frame your subject dictates psychological dominance, narrative intimacy, and spatial scale. Learn the essential 10 camera shot types and angle prompts that control OpenAI Sora, Kling AI 1.5, and Runway Gen-3 with director-level precision.
AI video generators do not automatically understand director shorthand like "make it dramatic". To guarantee a specific framing, your prompt must specify three interdependent variables: Shot Distance (e.g. Extreme Close-Up vs. Establishing Shot), Camera Elevation (e.g. Worm's Eye vs. Bird's Eye), and Optical Lens Pairing (e.g. 14mm wide angle vs. 85mm portrait prime).
| Camera Angle | Elevation Relative to Subject | Subconscious Audience Perception |
|---|---|---|
| Bird's Eye (Top-Down) | 90° directly above subject (overhead nadir) | Omniscient, detachment, fate, geometric patterns |
| High Angle | 30° – 60° above eye-level looking down | Vulnerability, weakness, smallness, surveillance |
| Eye-Level (Neutral) | Parallels subject's eye line (standard) | Human empathy, direct equality, natural reality |
| Low Angle | 30° – 60° below eye-level looking up | Heroism, power, dominance, imposing stature |
| Worm's Eye View | Floor level pointing directly upwards | Overwhelming scale, monstrous terror, towering majesty |
| Dutch Angle (Canted) | Optical Z-axis rotated 20° – 35° | Disorientation, dread, madness, imminent collapse |
Subject is tiny or barely visible; environment dominates 90%+ of frame
Isolation, immense scale, geographic context, grandeur
Extreme wide establishing shot of a solitary hooded traveler walking across vast salt flats at dawn, expansive epic scale, 14mm rectilinear ultra-wide lens, horizon line in lower third
Subject's full body from head to toe is visible with ground and ceiling clearance
Physical action, spatial relationships, posture, full wardrobe presentation
Wide full-body shot of a cybernetic knight standing in a dimly lit cathedral, full body visible from boots to helmet, boots kicking up dust, 28mm cine prime lens, balanced headroom
Framed from waist or mid-thigh up to top of head
Dialogue, natural human interaction, weapon holsters, relatable intimacy
Cinematic medium cowboy shot framed from mid-thigh up of a frontier gunslinger reaching for revolver holster, sharp eye contact, 50mm spherical lens, f/2.8 natural background falloff
Subject's face fills the frame from collarbone to forehead
Empathy, emotional intensity, psychological drama, micro-expressions
Dramatic cinematic close-up shot of an astronaut's face through reflective helmet visor, intense focused gaze, visible sweat beads and heavy breathing, 85mm portrait prime, f/1.8 shallow depth of field
Isolated detail: single eye, lips, watch gears, insect wing
Intense sensory focus, claustrophobia, revelation of fine craftsmanship
Extreme macro close-up framed tightly on an antique mechanical clock escapement wheel ticking, brass gears meshing with jewel bearings, 100mm macro lens, ultra-shallow depth of field
Camera placed below subject looking upwards toward the sky
Power, dominance, heroism, towering intimidation, vertical authority
Stark low-angle worm's eye view looking up at a towering monolith in a storm, camera tilted steeply upward, dark ominous clouds swirling overhead, 20mm wide lens exaggerating vertical scale
Camera positioned high above subject pointing directly or steeply down
Vulnerability, omniscience, strategic pattern recognition, entrapment
High-angle bird's eye view looking straight down at an intricate circular stone labyrinth at midnight, solitary lantern light illuminating the central path, 90-degree nadir perspective
Camera rotated 20–35 degrees on roll axis; horizon is intentionally tilted
Psychological instability, panic, moral ambiguity, suspense, intoxication
Dramatic Dutch angle shot with camera tilted 25 degrees along the optical Z-axis, canted horizon line, neon-lit subway corridor tilting diagonally, disorientation, 35mm anamorphic prime
Camera looks over the back of one subject's shoulder at another subject
Conversational pacing, confrontation, perspective sharing, narrative connection
Over-the-shoulder (OTS) shot framed behind a detective's shoulder in soft focus foreground, looking directly across interrogation table at suspect sweating under harsh lamp, 50mm lens
Camera acts as the exact physical eyes of a character in the scene
Total first-person immersion, visceral adrenaline, immediate empathy
First-person POV shot from the eyes of a downhill mountain biker flying down steep alpine ridge, handlebars visible in lower frame, rapid peripheral motion blur, head-mounted 14mm action cam perspective
Hollywood cinematographers rarely hold static angles; they chain angles with movement trajectories to reveal story information dynamically. In Sora and Kling 1.5, you can prompt compound trajectories that shift angles across the shot timeline:
"Stark low-angle shot starting 3 feet off the wet pavement, slowly pushing in toward the detective as he stands under streetlamp glow, camera maintaining an upward 30-degree tilt throughout the move."
"Overhead top-down 90-degree nadir shot of a spiral staircase, smoothly pedestaling vertically upwards 10 meters, spinning slightly to reveal labyrinthine architectural geometry."
Rather than just writing "close-up", specify the anatomical boundary or object isolation: for example, "macro extreme close-up framed tightly on the subject's left eye, eyelashes and iris reflections occupying the entire frame, 100mm macro lens, shallow depth of field". Also add "wide shot, medium shot, torso" to your negative prompt.
Use the term "Dutch angle shot, camera tilted 25 degrees along the optical Z-axis, diagonal compositional horizon line, high visual tension". This instructs models like Kling and Runway Gen-3 to rotate the horizon line rather than just distorting the subject geometry.
Generative diffusion models default to medium shots because the vast majority of web training data consists of eye-level human portraits. To break this default, you must explicitly declare the camera elevation, distance, and lens focal length within the first 10 tokens of your prompt.
Specify foreground occlusion and depth: "Over-the-shoulder (OTS) shot framed behind subject A's dark silhouette on camera right, looking past their shoulder toward subject B in sharp focus at camera left, 50mm lens, dialogue framing".
Upload any cinematic movie clip or commercial to Video to Prompt. Our optical timeline engine automatically identifies shot elevation, camera angle, and focal length for instant AI prompt reproduction.
Open Free Dropzone Tool →