Module 3: Cinematic Motion
3.1 Advanced Camera Prompting
Mastering the ‘AI Director’ language: Pans, Tilts, Dolly Zooms, and Rack Focus.
Introduction: The Digital Camera
Welcome to Module 3, where we bring our creations to life. Up until now, we have built the story foundation (Module 1) and the sound foundation (Module 2). Now, we need motion. We need to shoot our “principal photography.”
In AI Filmmaking, **video generation models** like Runway (Gen-3), Kling, or Luma Dream Machine are our “digital cameras.” They do not natively understand emotion; they understand specific visual instructions. Novice users make the mistake of prompting the *subject* (e.g., “A robot runs”) and letting the AI decide *how* to shoot it. A Director must dictate **how the camera moves** to tell the audience what to feel.
It is the process of defining the specific relationship between the camera and the subject. By using professional cinematography terminology, we move from passive observation to active visual storytelling.
Decoding the AI Camera
To master camera moves, you must first ensure you are still using the tools we built in Module 1. Every professional video prompt has four cascading layers. Do not forget them:
Directing with Sliders & Language
We are focusing on two control methods. We use **explicit language** in the prompt text, and we also leverage **slider interfaces** provided by advanced video models (especially Runway Gen-3 and Luma).
The Basic Pivot: Pan & Tilt
These moves pivot the camera from a fixed point. A **Pan** moves horizontally (left/right). A **Tilt** moves vertically (up/down). They are best for scanning an environment (Phase 3.1.1).
A sweeping, dramatic TILT UP from a detail of mossy footsteps on rusted metal, slowly revealing the massive post-apocalyptic lighthouse against the stormy ocean [Visual Style Guide] –ar 16:9
Advanced ‘Director’ Moves
These moves physically move the camera body, requiring the AI to recalculate the world’s perspective and depth in real time. These define the “awesome” look.
The Dolly (Push/Pull)
The camera physically pushes into or pulls back from the subject. Best for increasing focus or revealing a surprise.
Prompt: Close-up of the weathered lighthouse keeper. A dramatic, slow DOLLY PUSH-IN focuses on the motivation in their eyes. (Phase 3.1.2)
The Dolly Zoom (Vertigo)
The ultimate tension move. The camera physically pushes in *while simultaneously zooming out,* causing the background to warp or compress while the subject stays consistent. (Phase 3.1.3)
Prompt: Medium shot of the keeper on the edge. A slow DOLLY ZOOM creates immense vertigo as the ocean compresses behind them.
AI video struggles with complex focus pulls (Rack Focus). But advanced models have interfaces for this. In Runway Gen-3, look for the **”Camera Controls”** interface. Do not type “shallow depth of field” into the prompt. Instead, **click on the area you want sharp** (e.g., the foreground flower) and use the **focus slider** to blur the background. Generate. Then re-generate that same scene, but **click on the background lighthouse** as the sharpness point. Later, you can cross-dissolve these two clips in post-production for a perfect focus pull (We cover this in Module 5!).
Lesson Assignment
Your task is to take the 10 story beats you identified in Module 1 and “direct” them. **We do not need to generate video yet. We are writing the final, dynamic video prompts.**
- Review your story beats from Lesson 1.2.
- For at least **5 of your key beats**, draft the **full professional video prompt** (Subject/Action + Style Guide + Cinematic setting + Advanced Motion).
- Your list must include at least one prompt using an advanced camera technique: Dolly Push, Dolly Zoom, Pan, or Tilt.
- Submit your final prompt list in the assignment submission area below.