iPhone Portrait mode tips that fix the usual problems

How Portrait mode actually works, and why that matters

Portrait mode is not a lens effect. The phone builds a depth map of the scene, decides which pixels belong to the subject, and then blurs everything else by an amount it calculates from the f number you chose. On Pro models with a LiDAR scanner, that depth estimate gets help in low light. Everywhere else it is computed from the camera data.

Understanding that explains every failure you have seen. Hair fails because individual strands are thinner than the depth map's resolution. Glasses fail because the frame is at one distance and the lens shows the face at another. A hand held toward the camera fails because it crosses several depth planes at once.

It also explains every fix. Anything that makes the depth decision easier makes the photo better: more separation, even light, simple outlines, a subject that fills a sensible part of the frame.

The blur itself has improved a great deal, and the modern implementation renders out-of-focus highlights with a shape rather than a smear. The mask is still the weak point, not the blur.

Distance, the fix for almost everything

Two distances matter and people get both wrong.

Subject to background. This is what the effect is measuring. Two metres is a reasonable minimum, more is better. A person against a wall produces a flat cut-out with a visible outline, because there is no depth difference to render. Move them away from the wall and the same phone produces a convincing photograph.

You to subject. Portrait mode wants roughly one to two and a half metres, and it will tell you when you are outside that. Closer than about half a metre it usually refuses. Further away and the subject is too small for the mask to be accurate.

For two people, get their heads onto roughly the same plane. If one stands half a metre behind the other, the depth map will blur the one at the back, and it will look like a mistake because it is one.

The f number, and why the default is wrong

The dial at the top of a Portrait photo runs from about f/1.4 to f/16 and defaults to around f/4.5. That default is not the amount of blur a real 50 mm lens at f/4.5 would give you, because the effect is simulated, so treat the numbers as a strength slider rather than as physics.

Around f/2.8 is the sweet spot on most phones: enough separation to read as deliberate, not so much that the mask errors become the subject of the photograph. Below f/2 the background goes to soup and every stray hair gets clipped.

The setting is non-destructive. You can change it after the fact, endlessly, in Photos, Edit, and the f icon. So shoot at the default and tune it later if you would rather not fiddle in the moment.

One rule worth keeping: the busier the background, the higher the f number should be. A hedge two metres back at f/2.8 looks like a professional photograph. A crowded street at f/1.6 looks like a badly cut-out sticker on a smear.

Portrait Lighting, honestly assessed

Portrait Lighting sits on the dial beside the shutter: Natural, Studio, Contour, Stage, Stage Mono and High-Key Mono. It is relighting applied on top of the depth map, so it inherits every weakness of the mask and adds its own.

Natural Light is the only one that is reliably safe, because it applies no relighting at all. Studio Light brightens the face and is usually acceptable. Contour Light adds shadow to the sides of the face and works on strong bone structure and looks strange on soft features.

Stage Light and Stage Light Mono drop the background to pure black, and this is where the cut-out is most brutally exposed, because a pure black field shows every edge error. It can look striking on a shaved head in even light and it fails often enough that it is not worth relying on.

High-Key Light Mono, which drops the background to white instead, is the most forgiving of the dramatic ones and the most useful for anything you will actually keep.

All of them, like the f number, are reversible afterwards. Nothing about Portrait Lighting is baked into the file.

Adding or removing Portrait after the shot

On iPhone 15 and later, you do not have to be in Portrait mode at all. When the camera detects a person, cat or dog in ordinary Photo mode, it captures depth automatically, and a small f symbol appears in the corner of the viewfinder to tell you it has done so.

Later, open the photo, tap Edit, and Portrait appears as an option alongside the usual adjustments, with a Depth Control slider. You can also change the focus point by tapping a different part of the frame, which is the closest thing the phone has to refocusing after the fact.

The practical consequence is large: stop deciding in the moment. Shoot in Photo mode, which is faster, keeps the full field of view and does not lock you to one framing, and add the Portrait effect only to the frames that deserve it. Portrait mode files are also larger and slower to shoot, so this is a speed gain as well.

The limit is the detection. Depth is only captured when the phone spots a person, a cat or a dog. Photograph a bicycle in Photo mode and there is no depth data to use later, so for objects you still have to choose Portrait mode at the time.

When to stop using Portrait mode

Portrait mode is the wrong tool more often than its popularity suggests. Skip it when:

• The background is already far away. A subject in front of a distant landscape at 2x is naturally soft behind them. Simulating more just adds artefacts.
• It is dark. The depth estimate degrades with the light, and Night mode does not run in Portrait mode on most models, so you lose the one feature that would have saved the frame.
• There is fine detail at the edges. Curly hair, fur, lace, chain-link fences, wire glasses. The mask will find all of them and fail.
• Anyone is moving. The processing needs a stable subject, and Portrait mode has a longer capture than Photo mode.
• You want the place in the picture. If the background is why you took the photo, do not blur it.

The real alternative is optical: stand further back, zoom to 2x or to 4x on a Pro, and let the longer focal length compress and soften the background on its own. It is not as strong an effect, and it never produces a halo.

Worth one disclosure since retouching comes up here: ShutterAI is our own app. Its Glow tool is skin-safe retouch that keeps freckles and texture and does no reshaping, which is a different job from Portrait mode. It is not out yet. It reaches iPhone this fall, free to start, and you can join the waitlist.