Candid crowd shot at esports tournament AI Image Prompt
Create a realistic candid event-photography image shot from within the seated audience at a dark indoor {argument name="event type" default="esports gaming tournament"}. Use a wide 16:9 landscape frame, cool blue and purple stage lighting, shallow depth of field, high ISO concert-photo atmosphere, and a large blurred LED screen glowing in the background. In the foreground are three main seated spectators: on the left, a young man in a plain white long-sleeve shirt with a black lanyard, relaxed posture, hands low in frame; in the center, a young woman with long dark wavy hair wearing a pale pink ribbed wrap-style knit top, seated with hands folded in her lap; on the right, a young man in a light gray sweatshirt turned slightly toward the stage and holding a smartphone or small camera. Behind them is a dense crowd of spectators, some standing, some cheering. Include exactly 3 visible inflatable cheering sticks: two long white sticks with red tips held upright near the center-left background, and one blue-and-white stick partly visible behind the right spectator. Add privacy anonymization: place exactly 8 soft-edged rectangular purple blur blocks over visible faces, varying in size and opacity, covering the left foreground man, the top-left background person, a small center-left background face, the center-back person in red plaid, the central woman, a small mid-right background face, a bald mid-right background face, and the right foreground man. Make the image feel like an accidental TV broadcast crowd shot at a lively tournament, with natural poses, no readable text, no logos, no watermark, and photorealistic lighting.
Content: Formula: foreground specified spectators (position, clothing, pose) + background crowd with cheering sticks + precise privacy blur blocks + stage lighting/shallow DoF/high ISO photography parameters + natural candid vibe. Structure: subject→scene→lighting→color→composition→privacy treatment→style.
Pros: Extremely detailed, every visual element explicitly described; privacy blur seamlessly integrated without breaking realism.
Cons: Too many elements with exact counts (e.g., 8 blur blocks, 3 sticks), hard for model to comply precisely; subjective style like 'accidental capture' depends on model interpretation.
Reference image: No obvious reference-image dependency
Source materials are collected from publicly accessible web pages. Contact us if a rights issue needs review. Copyright & Privacy
