When you want the audience to hold their breath by stretching out, a few seconds at a time, a situation where the audience knows the danger and the character doesn't. Use it to show a spot where something is about to happen, like behind a door or under a desk.
In Shot 1, first show a device stuck under the desk with a blinking red light. In his interview with Truffaut, Hitchcock explained that suspense arises when only the audience knows about a bomb under the table, and this shot plays exactly that role.
To recreate it: Generate the device insert separately, feeding in the same desk image as the background start frame as a reference.
In Shot 2, the accountant seen from above types without knowing anything. From above, the person looks small and defenseless, which raises the audience's anxiety.
To recreate it: Write 'high-angle, looking down' together with the fact that the person doesn't know, like 'unaware of anything beneath it'.
In Shot 3, approach very slowly toward the face that has heard the ticking. The frame tightens and creates a suffocating feeling.
To recreate it: AI follows a slow approach in one direction well, so write just one: 'very slow push-in toward his eyes'.
In Shot 3, one hard side light buries half the face in deep shadow. Light and dark split sharply, giving the feeling that something is hidden.
To recreate it: Write 'single hard side light, chiaroscuro shadows', and fix any drained-looking color all at once with color grading at the end.
In Shot 4, the person looks down under the desk with the frame tilted. The slanted horizon makes you feel the moment balance collapses.
To recreate it: Give only the tilt and keep the camera fixed, as in 'static tilted camera', so the frame is less dizzying.
Alternate the device insert and the person shots, shortening the shot length each time to 2s, 1s, 0.5s. As the cuts get shorter, the audience's breath quickens.
To recreate it: Don't leave the short-cut rhythm to AI. Make clips of about 3s and cut them down in editing.
Lay a riser under the cross-cutting and build it louder, then cut the sound dead at the moment the person looks under the desk and move to the next scene. Sudden silence is scarier than the loudest sound.
To recreate it: Mute the generated clips and align the end of the riser exactly to the frame where the last cut changes.
Cinematic multi-shot sequence, 10 seconds total. Shot 1 (2s): static insert of a small black device with a blinking red light taped under a wooden office desk; Shot 2 (3s): static high-angle medium shot of a tired accountant in a gray suit typing at the same desk late at night, unaware; Shot 3 (3s): very slow push-in on his face as he stops typing and hears a faint ticking; Shot 4 (2s): static Dutch angle medium close-up as he slowly lowers his head to look under the desk, sweat on his temple. Consistent accountant and dark office, low-key chiaroscuro with a single hard lamp, desaturated bleach bypass grain, cold claustrophobic dread.
Alternate between a dangerous object and a character who knows nothing, and the audience knows first, so the waiting itself becomes tension. As the cuts get shorter and a low sound grows, the body reacts first.
Insert shot of a small black device with a blinking red light taped under a wooden office desk. Static camera, low-key lighting from a single hard lamp, deep shadows, desaturated high-contrast film look, cold and ominous.
High-angle medium shot of a tired accountant in a gray suit typing at the same desk late at night, unaware of anything beneath it. Static camera looking down, pools of hard light in a dark office, desaturated bleach-bypass colors, quiet dread.
Close-up of the accountant's face as he stops typing, hearing a faint ticking. Very slow push-in toward his eyes, single hard side light, chiaroscuro shadows across his face, gritty film grain, cold desaturated tones, tense and claustrophobic.
Dutch angle medium close-up of the accountant slowly lowering his head to look under the desk, sweat on his temple. Static tilted camera, low-key light from below, deep blacks, high-contrast desaturated film look, breathless suspense.
Alternate device inserts and character shots with cross-cutting, and shorten the cut length each time: 2 seconds, 1 second, 0.5 seconds. Lay a riser sound effect (sound design: hits, whooshes, risers) under it and let it grow, then at the moment the character looks under the desk, cut the sound off and move to the next scene.
Make one background start frame so the same desk and lighting appear in every shot, and generate the device inserts separately, using that image as a reference. A slow push-in in one direction is a move AI follows well, so make the tension shots with it, and make the short, choppy rhythm in editing. Colors that look drained of light tend to wobble from shot to shot, so match them all at once with color grading at the end.