How to do it, step by step
- Move the subject away from the background This is the largest single factor you control. A subject one metre in front of a wall gets almost no blur. The same subject with ten metres of street behind them gets a lot, in Portrait mode or out of it.
- Switch to the 2x or 5x lens A longer lens narrows the field of view, which magnifies the background and makes any blur that exists far more visible. It also lets you stand at a sensible distance for a head-and-shoulders frame instead of half a metre from someone's face.
- Get as close as the framing allows Depth of field shrinks quickly at short subject distances. Filling the frame with the subject, rather than cropping in later, is what turns a slightly soft background into a genuinely blurred one.
- Turn on Portrait mode if the subject is a person between about half a metre and two and a half metres That is the range where the depth map is reliable. Wait for the yellow box to confirm a lock before shooting, and check the preview at the subject's outline rather than at their face.
- Set the f-number after the shot, not before Open the photo, tap Edit, tap the f icon and drag. Around f/2.8 to f/4 looks believable; f/1.4 blurs so hard that the segmentation errors around hair and shoulders become obvious.
- Zoom in on the edges before you keep it Check hair, glasses, arms and anything held out toward the camera. If the cut-out is wrong, either move the subject away from the clutter and reshoot, or fall back on optical blur with the telephoto.
What actually creates background blur
Depth of field is governed by the physical size of the opening light passes through, not the f-number printed on the spec sheet. The opening, the entrance pupil, is the focal length divided by the f-number.
An iPhone main camera is about a 7 mm lens at f/1.78, so its pupil is under 4 mm across. A 50 mm f/1.8 lens on a full-frame camera has a pupil of nearly 28 mm. At the same field of view, that f/1.78 phone photo has roughly the depth of field of f/6 on full frame. The f-number sounds fast because the lens is short, not because the opening is large.
Three things determine how much blur you actually get, in order of how much control you have over each:
- Subject to background distance. Free, and by far the most powerful. Blur scales with how far behind the plane of focus something sits.
- Camera to subject distance. Depth of field shrinks sharply as you close in, which is why a macro shot of a flower blurs everything and a full-length shot of a person blurs nothing.
- Focal length. A longer lens magnifies the background, so whatever blur exists is more visible.
Getting real blur out of a phone
Optical blur has one enormous advantage over the computed kind: it cannot go wrong at the edges, because there are no edges. Every strand of hair, every gap in a railing, blurs by exactly the amount its distance dictates.
- Use the telephoto for people. At 2x to 5x you stand two to four metres back for a head-and-shoulders frame. That distance flatters faces, and the narrow field crops the background to a small, magnified, well blurred patch.
- Find depth behind your subject. A long street, an open park, a room with the far wall ten metres away. Standing someone against a wall guarantees a sharp background whatever mode you use.
- For objects, get very close. The main camera focuses to about 10 cm, and a cup at that distance has genuinely shallow depth of field.
- Do not shoot ultra-wide. The 0.5x has the shortest focal length and smallest pupil, so it has the deepest depth of field on the phone.
How Portrait mode actually works
Portrait mode is not a lens effect. The phone builds a depth map of the scene, decides which pixels sit at which distance, then paints a synthetic blur that grows with depth. The map comes from three sources: parallax between two cameras viewing the scene from slightly different positions, the LiDAR scanner on Pro models, which measures distance directly out to about five metres and is what makes low-light portraits work, and a neural network that segments people and objects from the background.
On top of the blur, the phone renders bokeh: bright points in the background are drawn as discs rather than smeared, which is what a real lens does and what makes the effect convincing on a night scene with lights behind the subject.
Two things follow. The f-number slider controls no aperture; it scales the blur radius applied to the depth map, which is why you can change it afterwards and why f/1.4 looks fake while f/2.8 to f/4 does not. And on recent iPhones any photo captured with depth information can be made into a portrait later, with the focus point moved by tapping a different subject.
Where Portrait mode falls apart
Every failure is a segmentation failure: the phone put a boundary in the wrong place, and synthetic blur is unforgiving about boundaries because it is either applied or not.
- Hair. Flyaway strands and curls are finer than the depth map's resolution, so they are assigned to the background and vanish, leaving a helmet-shaped outline. Backlit hair is the worst case.
- Glasses. Frames read as one depth while the lenses show a displaced version of the face behind them, so you get a sharp frame, a blurred eye and a hard edge across the temple.
- Fences and railings. The gaps are background but sit inside the subject's silhouette, so Portrait mode either blurs the bars or sharpens the gaps. Both look wrong.
- Anything reaching toward the camera. A hand held out, a drink, a dog's nose. These sit at a different depth from the face, so they blur when they should not.
- Transparency. Glass, veils, water, smoke. There is no single correct depth for a transparent pixel, so the map guesses.
The practical answer is to give the segmentation an easy problem: clear space between subject and background, a simple silhouette, no railings, and check the outline in the preview rather than the face.
Adding blur to a photo you already took
| Method | Real or synthetic | Best for | Where it breaks |
|---|---|---|---|
| Telephoto plus distance | Real optical blur | Portraits, anything with hair or glasses | Limited blur, needs space behind the subject |
| Macro distance on the main lens | Real optical blur | Objects, food, flat lays, details | Only works within about 30 cm |
| Portrait mode | Synthetic, depth mapped | People and pets at half a metre to two and a half metres | Hair, glasses, railings, hands toward the camera |
| Editing a Portrait photo later | Synthetic, depth mapped | Changing focus point or blur strength after the fact | Only works if depth was captured at the time |
| Third-party segmentation blur | Synthetic, mask based | Non-portrait photos that need separation | Same edge cases, usually with a coarser mask |
| Lens Blur or tilt-shift tools | Synthetic, not depth aware | Landscapes and flat scenes | Obviously fake on people, blurs a shoulder and a wall equally |
Apple Photos can change the blur and focus point of a photo captured with depth data, but it cannot add blur to a photo that has none. Snapseed's Lens Blur lays down a radial or linear gradient with no idea what is in the frame, which reads well on a table top and badly on a person. ShutterAI's Blur It applies background blur after the fact with a tappable focus point, so a photo shot in the normal camera can still be separated later.
Which method for which photo
A short decision rule that avoids most disappointment.
- A person with space behind them: 2x or 5x lens, no Portrait mode. Real blur, no edge artefacts, full-resolution file.
- A person against a close background: Portrait mode, f/2.8 to f/4. It is the only thing that will separate them, and at moderate strength the segmentation errors stay small.
- Curly or backlit hair, glasses, a railing: optical only. This is where synthetic blur shows its working.
- Food, objects, details: main lens, as close as it will focus. No mode needed.
- Low light with lights behind: Portrait mode on a Pro model. LiDAR still measures depth in the dark, and the rendered bokeh discs on background lights are the best part of the effect.
Questions people ask
How do I blur the background of a photo on my iPhone?
Either use Portrait mode, which builds a depth map and blurs by software, or get real blur by switching to the 2x or 5x lens, standing your subject well in front of the background and moving close. Optical blur is weaker but never breaks at the edges.
Why does my Portrait mode photo have a hard edge around the hair?
Portrait mode blurs using a depth map, and individual strands of hair are finer than the map can resolve, so they get classified as background and blurred away. Front-lit hair against a plain background segments far better than backlit flyaways.
Can I blur the background of a photo I already took?
If it was shot in Portrait mode or with depth data, yes, and you can move the focus point too. If it was a normal photo, you need a tool that segments the subject afresh. Generic lens blur filters are not depth aware and look fake on people.
Which iPhone lens blurs the background most?
The telephoto, in practice. It has a larger physical opening than the ultra-wide and it magnifies the background, so a head-and-shoulders portrait at 2x to 5x separates better than the same framing on the main lens. The 0.5x ultra-wide blurs least of all.
Why does my Portrait mode blur look fake?
Usually the f-number is set too low. At f/1.4 the blur is stronger than any lens of that size could produce and every segmentation error becomes visible. Drag it to f/2.8 or f/4 and it reads as a photograph again.