The two front-most layers in the prompt: Director Instruction sets the tone, Constraint establishes the rules, Baseline anchors image quality. Together they answer what tone is this shot and what bottom-line rules must be followed.
The soul field of a single-shot prompt, governing the style orientation of all sub-fields. Must cover four elements simultaneously, none can be omitted:
| Element | Business Meaning | Standard Example | Common Mistake |
|---|---|---|---|
| Style Positioning | Quality tier and aesthetic reference | Hollywood director quality / documentary realism | Vague adjectives like good-looking |
| Realism Requirement | Exclude surreal elements | Realistic style, no special effects, no science fiction elements | Missing negative declaration, model auto-adds sci-fi |
| Mood Baseline | Emotional atmosphere | Calm and professional / Thrilling | Mood word conflicts with image (tragedy written as cheerful) |
| Technical Direction | Imaging medium and optics | Movie-level picture, 35mm film look, natural light effect | Contradiction: film grain paired with digital sharpness |
Standard format: {style positioning}, {technical direction}, {realism requirement}, {mood baseline}, {supplementary technical direction}
Good example: Hollywood A-list director quality, cinematic visuals, realistic style, no special effects, no sci-fi elements, calm and professional tone, natural lighting
Informs the model of underlying technical boundaries (aspect ratio, resolution, format, frame rate). Director instruction defines the vibe; constraint defines the specs.
Standard format: Aspect ratio: 16:9, Resolution: 1920x1080, Format: MP4, Frame rate: 24fps
Vertical short-video delivery must include platform-specific rules (e.g., 9:16 or specific encoding). This field is auto-injected by the system based on delivery specs, not generated creatively.
Injects image-quality anchor words to establish the minimum quality standard for output. Pick 1–2 words per category and usually place them at the very front of the prompt.
| Category | Recommended Vocabulary | Rendering Effect |
|---|---|---|
| Resolution Anchor | 8K resolution, ultra high definition, 4K crisp detail | Constrains minimum pixel density, prevents blur |
| Style Quality | cinematic quality, photorealistic, hyperrealistic, film grain | Steers toward cinematic / photorealistic look |
| Detail Enhancement | highly detailed, intricate textures, fine details, sharp focus | Enhances texture, edges, material detail |
Standard baseline phrase: 8K resolution, cinematic quality, highly detailed, photorealistic