Same prompt, generated with each model separately — click one to see how it holds up.
Why this works
“High-angle CCTV view from a building corner” and “slight fisheye distortion” establish an impersonal surveillance frame, while “looking cautiously over their shoulder” supplies the tense, suspicious action. The “narrow urban alley” and “graffiti-covered concrete wall” compress the scene around the figure, and “blue-green night-vision tint,” “heavy digital noise,” and “low contrast” create the eerie security-camera palette and degraded texture. “Wet pavement with faint reflective puddles” adds small, subdued highlights without weakening the deliberately grimy realism of “photorealistic CCTV surveillance snapshot.”
FAQ
→How do I make the hooded person more prominent and central?
Replace “standing beside the concrete wall” with “standing centered in the alley, occupying most of the frame from the waist up,” and add “subject sharply prioritized in the middle of the surveillance image.” Keep “high-angle CCTV view” if you want the person to remain observed from above.
→How do I shift the scene from eerie surveillance to a warmer, more cinematic night mood?
Replace “stronger blue-green night-vision tint” with “muted amber sodium-vapor streetlight with warm orange highlights,” and replace “low contrast” with “moderate cinematic contrast.” Change “heavy digital noise” to “subtle film grain” to retain texture without the harsh security-camera degradation.
→How do I create a series of variations while keeping the same surveillance identity?
Keep “Photorealistic CCTV surveillance snapshot,” “slight fisheye distortion,” “heavy digital noise,” and “realistic security-camera grit” fixed, then vary the location phrase “narrow urban alley” with options such as “underground parking garage,” “rainy service corridor,” or “empty train-platform stairwell.” Change the subject action from “looking cautiously over their shoulder” to “walking past the camera,” “pausing beneath a flickering light,” or “partially obscured behind a dumpster.”
Learn the technique behind this
- Why can't AI spell, and how do I get readable text in an image? — Older diffusion models had no character-level representation of text, so they produced letterform-shaped texture instead of words.
- How should a prompt be structured, and does word order matter? — A prompt that behaves predictably names one subject first, then what it is doing, then where, then the light, then the lens or medium, then the style.
- Do negative prompts work, and what should go in one? — Negative prompts work on models that support a separate negative conditioning channel — mainly Stable Diffusion and FLUX-family models.
Related prompts
Grotesque Claymation Gangsters in Gritty Alley
Charcoal coat with tan-handled tote
Woman in Teal Jacket Capturing Twilight Reflection
Red Gummy Bear Monster Attacks
