Midjourney Lens and Depth of Field: Focal Lengths and Bokeh

How Midjourney reads lens and camera words

Midjourney does not simulate a lens. There is no focal length setting, no aperture and no sensor. When you write 85mm f/1.8, the model is not calculating optics; it is recalling what photos captioned with those words tended to look like. In practice that means a tight, flattering portrait with a soft, blurred background.

So treat lens and aperture terms as stylistic cues, not physics. They are useful because photographers label their work consistently, and those labels carry a recognisable look. But the numbers are approximate. The difference between 85mm and 105mm, or f/1.4 and f/2, will rarely be visible. The difference between 14mm and 135mm, or f/1.4 and f/16, usually will be.

Lens cues work best alongside the plain-English description of the effect you want. 85mm, shallow depth of field, background softly blurred gives three agreeing signals. Lens terms also interact with shot type and camera angle, so keep them consistent: a 14mm lens and an extreme close-up of an eye send mixed messages.

Focal lengths and the look they suggest

Tilt-shift lenses are also used by architectural photographers to keep vertical lines straight, but Midjourney mostly associates the term with the miniature effect. If you want straight verticals, ask for straight vertical lines, architectural photography instead.

Aperture words and depth of field

Depth of field is how much of the image, front to back, is in sharp focus. A wide aperture (small f-number) gives shallow depth of field: a sharp subject against a blurred background. A narrow aperture (large f-number) gives deep focus, where nearly everything is sharp.

Midjourney responds to both the f-numbers and the descriptions. The descriptions are usually the stronger signal, so use them together:

• Shallow depth of field: f/1.4, f/1.8, wide aperture, shallow depth of field, background softly out of focus, subject isolated.
• Moderate: f/4, f/5.6, background slightly soft. Useful for two people or a subject with some context.
• Deep focus: f/11, f/16, deep focus, everything in sharp focus, sharp from foreground to background. Good for landscapes and architecture.
• Focus placement: focus on the eyes, blurred foreground leaves, sharp foreground, soft distant city. These help Midjourney decide what should be sharp.

Bokeh: getting attractive background blur

Bokeh is the quality of the blur in out-of-focus areas, especially the soft discs that small points of light become. To get pleasing bokeh, you need two things in the prompt: shallow depth of field and something in the background that can turn into it.

1. Ask for the blur: shallow depth of field or f/1.4.
2. Give it light sources: fairy lights, city lights at night, sunlight through leaves, candles in the background.
3. Name the quality: creamy bokeh, soft round bokeh, dreamy bokeh.
4. Keep the subject well separated from the background, for example standing several metres in front of.

portrait of a woman in a wool scarf at a winter market, 85mm, f/1.4, creamy round bokeh from string lights behind her, focus on the eyes --ar 4:5 --style raw

Copy-ready lens prompts

• 14mm ultra-wide interior photograph of a vaulted library, dramatic perspective, deep focus, everything sharp --ar 3:2 --style raw
• 35mm street photograph of a flower seller in a rainy market, documentary style, natural light, moderate depth of field --ar 3:2
• 50mm photograph of a family dinner table, honest natural perspective, warm evening light --ar 3:2 --style raw
• 135mm telephoto fashion portrait, model on a city street, strong background compression, buildings melting into soft blur --ar 2:3
• macro photograph of a dew drop on a spider web, razor-thin focus, soft green background --ar 1:1
• tilt-shift photograph of a busy harbour from above, miniature effect, narrow band of focus, bright saturated colours --ar 16:9

When lens cues do not work

• The background is still sharp. Add plain-language blur words and give the background some distance. Lens numbers alone are a weak signal.
• Everything is blurry except a tiny area. Midjourney can overdo shallow focus. Try f/4 or background slightly soft instead of f/1.2.
• You stacked camera gear. Long lists of camera bodies, lenses and film stocks dilute the prompt. One lens cue plus one depth of field phrase is usually enough.
• The image looks over-polished. Midjourney’s default aesthetic can override photographic realism. Try --style raw or a lower --stylize value.
• Contradictions. shallow depth of field, everything in sharp focus cancels itself out. Decide what should be sharp.

Doing this faster with Dreamdrive

Dreamdrive saves every prompt you run automatically, with its picture, model and settings, into the project you choose, so you can search your Library for it later, copy it with Copy prompt or drag the picture into another tool.