Cinematic Portrait Lighting Studio
Generates a moody, film-still-quality portrait with Rembrandt lighting, shallow depth of field, and rich color grading.
Cinematic is a look you can engineer: light ratios, lens language, grain, and pacing. These prompts bake film-grade direction into copy-paste templates — from trailer music and product ads to moody portrait photography. Paste one in and your output stops looking like AI and starts looking like cinema.
Generates a moody, film-still-quality portrait with Rembrandt lighting, shallow depth of field, and rich color grading.
Sweeping aerial opener — golden hour, slow reveal over any landscape or city.
A 6-second cinematic product spot for Veo 3: hero shot, detail macro, logo resolve — the commercial structure that converts.
Turn any still into a cinematic opener: slow push-ins, parallax drift, and living atmosphere from a single image.
Braams, risers, and the three-act hit: trailer-grade cinematic music with real structure, not just loud strings.
The Sora 2 format that actually goes viral: one subject, one camera move, three timed beats — establish, reveal, payoff — written so the model holds the shot together.
Cinematic is not a filter you name — it's a stack of decisions you describe. This hub's prompts specify light ratio (a bright key against a dark room), lens behavior (shallow focus, slow push-in), texture (grain, haze, halation), and pacing (a held shot instead of cuts). Video models like Veo and Runway render those decisions faithfully; image models like Midjourney read the same language as style direction. The fastest way to learn the stack: take one prompt from this hub, generate it, then change exactly one decision — swap golden hour for overcast noon — and watch what the word was carrying. Within an hour you'll write cinematic prompts of your own that outperform preset filters, because you'll be directing instead of decorating.
One gorgeous clip is a screensaver. Three clips that belong together are a film. The difference is a shot list and a style block. Before you generate anything, write the sequence down in plain words: the wide shot that places us, the detail that tells the story, the face or object we end on. A thirty-second piece usually needs three to five shots, and generating them in story order — rather than whichever prompt excites you first — exposes the gaps while fixing them is still free.
The practical trick is the style block: a short paragraph of decisions you paste, unchanged, at the end of every prompt in the sequence. Same grade ("muted teal shadows, warm highlights"), same texture ("fine 35mm grain"), same light logic ("low sun from frame left"), same lens family. Video models treat that repeated block as the film's identity, so a forest dawn and a cabin interior suddenly share a world. Change the block and you have visibly switched films — which is exactly how you should handle a deliberate time jump: new block for night, old block for morning, and the audience reads it as hours passing.
Shoot the hardest shot first. If the sequence depends on one hero image — the reveal, the product, the face — generate it while your standards are fresh, then build the supporting shots to match it, pasting its description of light and color into their prompts. Save the easy connective shots (a road, a sky, a door closing) for last; they flex to fit whatever the hero shots demand. Finish with sound in mind: a Suno trailer cue or even room tone laid under the cuts hides small visual mismatches that the eye would otherwise catch. Consistency is rarely about better prompts. It is about refusing to let each prompt start from zero.
Professionals increasingly split the job between two tools: still first, motion second. Generate the frame in an image model, where you can afford twenty attempts at the exact light and composition, then hand that frame to a video model as the opening image and prompt only the movement — "camera pushes in slowly, mist drifts left". You have separated the two hardest problems in AI video, art direction and motion, and each tool now solves the one it's best at. The failure rate drops by more than half, and your shot list becomes a row of approved frames before a single second of video exists. Direct the stills like a photographer, direct the movement like a camera operator, and never ask one prompt to do both jobs at once.
Veo 3, Runway, and Kling for video; Midjourney and DALL-E for stills; Suno for trailer-style music. The camera and lighting language is shared across all of them — each page names the tool it was tested on.
Usually too much light and too much motion. Cinematic shots are darker than feels safe, with one clear light source, and the camera moves slowly or not at all. Add 'low-key lighting, static tripod shot, subtle film grain' and compare.
2.39:1 (anamorphic widescreen) is the classic film frame. For social platforms, generate 16:9 and crop, or describe a 'cinematic 2.39:1 composition' inside a 16:9 frame for letterboxed output.