The blobs of light that give your photos away
You're at a restaurant. Your friend is sharp, laughing, perfectly lit, and behind her a row of pendant lights dissolves into soft floating orbs. You post the photo. Someone who shoots on a mirrorless camera sees it and knows, immediately, what phone you used. Not from the colors. Not from the noise. From the shape of those orbs.
That shape is not random. It's mechanical. It's the direct fingerprint of how many blades your camera's aperture iris uses to form its opening.
Aperture blades: the iris inside the lens
Every camera lens, including the tiny stacked modules inside a smartphone, controls how much light enters using an iris diaphragm. Thin overlapping blades slide together or pull apart to form an opening of varying sizes. The number of those blades, and how they're shaped, determines the geometry of that opening at any given f-stop.
At maximum aperture the blades retract almost fully, and the opening is roughly circular regardless of blade count. Close down the aperture (a higher f-number, less light), and the blades intrude inward. The opening becomes a polygon. Five blades make a pentagon. Six make a hexagon. Nine make a near-circle with subtle facets. Eleven blades make something your eye genuinely struggles to distinguish from a true circle.
Here's the mechanism that makes this matter for bokeh: out-of-focus points of light in an image are essentially photographs of the aperture opening itself. A tiny, distant light source that falls outside the focal plane gets projected onto the sensor as a blurry disc, and that disc takes the exact shape of the aperture iris. The aperture is, in effect, a stencil.
Five blades? Pentagonal bokeh balls. Nine blades with slightly curved edges? Almost circular, with a faint polygon suggestion. It's geometry all the way down.
The smartphone wrinkle
Most dedicated camera lenses for mirrorless or DSLR systems carry between seven and eleven blades, and manufacturers often curve the blade edges specifically to produce rounder bokeh wide open. It's a deliberate aesthetic choice, and one that costs money in precision manufacturing.
Smartphone camera modules operate under brutal physical constraints. The entire optical stack, sensor included, often sits within a space thinner than a stack of credit cards. Many smartphone cameras use fixed apertures with no iris at all, meaning the "aperture" is just a fixed circular hole. Those produce perfect circular bokeh in the optical sense, which sounds ideal, except it also means you cannot control depth of field mechanically, not even a little.
The phones that do include a variable aperture typically offer only two positions rather than a smooth continuous range, toggling between something like f/1.5 and f/2.4. Two positions, two blade configurations, not a full iris sweep. The blade count in those modules has typically sat between five and seven, producing bokeh that shifts between slightly pentagonal and slightly hexagonal depending on which mode you're in.
Then there's the third category. Computational bokeh, where no iris is doing any work at all. The camera shoots at a fixed aperture, uses depth-sensing data or machine-learning segmentation to identify the subject boundary, and synthetically blurs the background. The "bokeh balls" in Portrait Mode on an iPhone or the equivalent on a Pixel are generated by software applying a circular blur kernel, which is why they look uniformly circular and occasionally slightly wrong at subject edges. The software doesn't know about aperture blade geometry, so it doesn't replicate it, and that is precisely why trained eyes can spot computational bokeh from across a room.
Two photographers, one scene, completely different results
Imagine two people photographing the same subject: a friend standing in front of a restaurant window, with a row of pendant lights receding into the background. One is using a mirrorless camera with a 50mm f/1.8 lens carrying nine curved blades. The other is using a mid-range Android phone with a five-blade variable aperture set to f/1.5.
The mirrorless shooter gets bokeh balls that are almost perfectly circular in the center of the frame, with a slight cat-eye shape toward the corners (a separate optical phenomenon called vignetting of the aperture, worth its own piece). Soft, organic discs.
The Android shooter gets something subtly different. Those same pendant lights render as soft but faintly pentagonal shapes, like someone has very gently pressed a thumb against each circle. Not ugly, not wrong, just geometrically distinct. On a phone screen you'd likely never notice. Crop to 100% or print it large, and the five-sided ghost of that aperture iris appears in every background light source.
Neither is a failure. They're just different mechanical signatures.
What people consistently misread about this
The widespread assumption is that more bokeh, meaning blurrier backgrounds, is always better, and that rounder bokeh balls mark a superior camera. This is wrong on both counts, and it's the kind of wrong that leads people to spend money fixing problems they don't actually have.
Some cinematographers actively prefer the hexagonal bokeh produced by a six-blade cinema lens because it feels less digitally clean and more filmic. Certain vintage lenses with five-blade irises have devoted followings precisely because their angular bokeh reads as characterful rather than clinical. The geometry becomes part of the look.
The more important variable is usually bokeh transition: how smoothly does the image shift from sharp to blurred? A lens with beautiful circular bokeh balls but a harsh, nervous transition zone (called "busy" or "nervous" bokeh) is often less pleasant to look at than a lens with slightly polygonal orbs and a silky smooth falloff. Blade count shapes the highlight discs. The optical formula shapes everything else.
Computational bokeh tends to nail the disc shape (perfect circles, every time) while stumbling on the transition, producing that tell-tale hard edge around hair or glass. It solves the easy geometric problem while leaving the harder perceptual one partially unfinished. Which, when you think about it, is a very relatable way to approach a difficult task.
So here's the question worth asking before you next shoot in Portrait Mode: do you actually know what your phone is doing to that background? If the spec sheet lists a single fixed f-number with no variable option, you're getting computational blur for everything. Not a flaw you need to fix. Just useful to know what the machine is actually doing when it makes your background go soft.
The blur is never free. Something always draws the shape.