This entry explains the three factors that determine depth of field, how to dial in shallow depth of field or deep focus according to narrative needs, and where the focus should land. It suits shooters who can’t tell where the blur comes from, and also those writing AI video prompts who want precise control over sharpness and blur in the frame.
When to Use Which
Shallow depth of field works for: close-ups and interviews (isolating the subject from a cluttered background); product shots (keeping the eye on the product alone); emotional passages (background dissolves into bokeh, the image turns “soft”—the so-called “cinematic look”).
Deep focus works for: landscapes and establishing shots; group scenes with layered blocking (multiple people sharp front to back); moments where the environment itself carries information (room decor telling the story). Don’t blur away the context you need to convey just for the sake of a cinematic look.
How to Do It
- Set your sharpness goal first: ask yourself whether this shot is “about the person” or “about the environment” → person means shallow depth of field, environment means deep focus → blur is a narrative choice, not a default filter.
- Adjust the aperture: want more blur → open up the aperture, lower the f-number (e.g., f/4 → f/1.8) → depth of field narrows noticeably; want more sharpness → stop down to f/8–f/11.
- Change focal length, adjust subject distance: if the aperture is wide open and still not blurry enough → switch to a longer focal length and shorten the subject distance → these two have a stronger effect than aperture: depth of field is roughly inversely proportional to the square of focal length and proportional to the square of subject distance.
- Lock where the focus lands: once the composition is set → place focus by “eye > face > body”; for portraits, focus on the eye closest to the lens → where the focus sits matters more than how much blur you have; if the eyes are sharp, the shot works.
- Need to shift attention: if foreground A and background B both carry action → use a rack focus, moving the focal plane smoothly from A to B within 1–2 seconds → the viewer’s eye follows the focus.
Parameters / Key Points
- One-line rule: wider aperture (smaller f-number), longer focal length, closer subject distance → shallower depth of field; the three compound each other
- Aperture ranges: f/1.4–f/2 extremely shallow, reserved for single-person close-ups; f/2.8–f/4 the portrait safe zone, roughly sharp from eye to ear; f/8–f/11 deep focus, common for landscapes and group shots
- Combo A: 85mm + f/1.8 + subject distance 2 m → on full frame, depth of field is only about 5–8 cm; eyes sharp, ears already softening, background fully melted away
- Combo B: 24mm + f/8 + subject distance 3 m → already past the hyperfocal distance, roughly sharp from 1.3 m to infinity; a classic deep-focus setup
- Bokeh: the soft discs formed when background point lights are blurred; the wider the aperture and the farther the background sits from the subject, the larger and rounder the discs
- Phones and small-sensor cameras have inherently deep depth of field: for true shallow depth of field, rely on the trio of “telephoto end + get close to the subject + push the background farther away”; don’t count on algorithmic blur
AI Prompt Writing
- Core terms:
shallow depth of field/bokeh/deep focus/rack focus - Chinese template: shallow depth of field, focus on [subject]‘s eyes, background blurred into soft bokeh; for deep focus write: foreground and background both fully sharp
- English template:
shallow depth of field, sharp focus on [subject]'s eyes, creamy bokeh background, 85mm portrait look; deep focus:deep focus, foreground and background both sharp, 24mm wide-angle; rack focus:rack focus from [A] in the foreground to [B] in the background - Key point: always specify where the focus lands (focus on …); dropping a lone “bokeh” hands the focal position over to randomness; you can add focal length and f-number (e.g., 85mm f/1.8) as a stylistic cue
- Response differences across models: unverified (to be filled in after real-world testing)
Common Mistakes
- Shallow for shallow’s sake, f/1.4 across the whole film → all environmental context is mush, and the audience never knows where the story takes place
- Shooting a two-person portrait at a wide aperture → the two aren’t on the same focal plane, so one is sharp and the other soft; stop down to f/2.8–f/4, or place both at equal distance from the lens
- Focus lands on the ear, the tip of the nose, or behind the lashes → the eyes are slightly soft; the viewer can’t say what’s wrong but just feels it’s “not locked”
- Prompts pile on bokeh without stating the focus position → the blur lands randomly, and the subject may end up blurred instead