Shots
A Shot is a single continuous camera angle within a Beat or Scene. Shots are the atomic unit of video generation in ACT3 AI — each shot produces one rendered video clip using the AI model of your choice.
Shot Types
ACT3 AI supports all standard cinematic shot types:
| Shot Type | Abbreviation | Description |
|---|---|---|
| Extreme Wide Shot | EWS | Establishes location; subject tiny in frame |
| Wide Shot | WS | Full subject visible; environment prominent |
| Medium Shot | MS | Subject from waist up |
| Close-Up | CU | Face or object fills frame |
| Extreme Close-Up | ECU | Detail shot — eyes, hands, props |
| Over the Shoulder | OTS | Camera behind one character, facing another |
| Point of View | POV | Camera sees what the character sees |
| Two Shot | 2S | Two characters in frame together |
| Bird's Eye View | BEV | Directly overhead, showing space and geometry |
| Dutch Angle | DA | Camera tilted to suggest tension or unease |
Camera Angles
Beyond shot type, you can specify the camera angle:
- Eye Level — Neutral, balanced perspective
- High Angle — Camera looks down, making subjects appear smaller or weaker
- Low Angle — Camera looks up, emphasizing power or dominance
- Over-the-Shoulder — Builds spatial awareness in dialogue scenes
- Worm's Eye — Extreme low angle for intensity or surrealism
Camera Movement
Each shot can include one camera movement:
| Movement | Effect |
|---|---|
| Static | Camera holds still — drama comes from subject |
| Pan | Rotates left or right on a fixed axis |
| Tilt | Rotates up or down on a fixed axis |
| Dolly | Physically moves forward or backward |
| Tracking | Moves parallel to a moving subject |
| Crane / Jib | Rises or descends smoothly |
| Handheld | Slight natural movement for documentary or chaos |
| Steadicam | Smooth movement following a subject |
| Drone | Aerial perspective with sweeping drift |
| Zoom | Changes focal length for emphasis |
Shot Prompting
Each shot has a text prompt field where you describe what the AI should generate. Good prompts combine the shot type, camera movement, subject description, and mood.
What to include in a shot prompt:
- Action — What is happening in the shot and who is doing it
- Camera type and movement — Use standard terms: wide shot, close-up, dolly in, tracking, handheld
- Mood and atmosphere — Tense, serene, chaotic, hopeful
- Lighting description — Golden hour, film noir, studio, night exterior
- Environmental details — Weather, time of day, location character
Good example prompts:
- "Marcus walks slowly through the foggy alley, looking over his shoulder, rain hitting the cobblestones, handheld medium shot, cold blue street lighting, tense."
- "Wide establishing aerial shot of futuristic city at night, drone movement, neon reflections on wet streets."
- "Extreme close-up on protagonist's eyes, static camera, shallow depth of field, cool blue lighting, revealing fear."
- "Tracking shot following character down neon-lit alley, Steadicam smoothness, rain reflections."
Common mistakes to avoid:
- Vague action ("someone walks around") — be specific about who does what
- Contradictory instructions ("bright golden hour" combined with a "night exterior" preset)
- Too many competing elements in one prompt — break complex shots into two shorter shots
Keep prompts under 100 words. Use standard cinematography terms for best AI interpretation.
Shot Settings in ACT3 AI
When you create or edit a shot, you configure:
- Shot type — from the standard types above
- Camera movement — from the movement list above
- Lens / focal length — wide (24mm), normal (50mm), telephoto (85mm+)
- Lighting preset — golden hour, overcast, studio, night, noir, etc.
- Subject description — what or who is in frame and what they are doing
- Background / set — linked location or 2D/3D set
- Duration — target clip length in seconds (2–8 seconds recommended for AI generation)
- AI model — Google Veo 3.1, Grok Video, or Wan 2.1
Generating Video from a Shot
- Open a scene card in the editor
- Click + Add Shot
- Fill in shot type, movement, subject description, and duration
- Click Generate
- ACT3 AI sends the prompt to the selected AI model
- The rendered clip appears in the shot card for review
Use the Render Queue to batch-render multiple shots at once and monitor progress. Apply Render Styles for a consistent visual look across all generated clips.
You can regenerate any shot with different settings or swap to a different AI model without losing your other shots.
Credit Usage
Each generated shot consumes credits based on duration, resolution, and the rendering engine used. Example rates:
- 5-second shot at 1080p using standard engine = approximately 5 credits
- Higher resolution or 4K output increases credit usage
- Draft renders at lower quality consume fewer credits
Use draft renders for creative iteration and reserve high-quality output for approved shots.
Shot Sequencing
Shots are displayed in sequence within their parent scene. You can:
- Reorder shots using drag and drop
- View all shots in Top-Down View to verify camera placement and blocking
- Lock approved shots to prevent accidental changes
Best Practices
- Keep shot duration under 8 seconds for AI generation — longer shots tend to lose coherence
- Use simple, industry-standard film terms in your prompts
- Vary angles across shots to maintain visual rhythm
- Break complex camera moves into two simpler shots rather than one overloaded prompt
- Preview at draft quality before committing to final resolution