Megapixels Don't Guarantee Usable Detail in Phone Photos

In 2012, the Nokia 808 PureView put a 41 megapixel sensor behind a lens that, by a 2012 estimate from the editor of Vision Systems Design, could not focus a point of light smaller than 3.4 micrometres (millionths of a metre), while each pixel was only 1.4 micrometres wide. If you have zoomed into a high-resolution phone photo and found fine textures soft or grainy, that gap is where the story starts.

The claim holds: in a phone camera, a higher megapixel count does not by itself add usable detail, because lens physics, the light each tiny pixel collects and later processing all limit what extra pixels can give.

Even a perfect lens blurs a point of light

Light spreads out as it passes through a small opening. This effect is called diffraction. Cambridge in Colour, a photography tutorial site, explains that it sets a resolution limit that does not depend on the number of megapixels. A flawless lens turns a single point of light into a small disc, called an Airy disk. The width of that disc depends only on the lens’s f-number (the rating written as f/1.8 or f/2.4) and the wavelength of the light.

The same site says resolution starts to be limited when the disc spans about 2 to 3 pixels, and softness can show before that. Here is the check using the Nokia figures.

A disc about 2.4 pixels wide sits inside that 2 to 3 pixel range. This is one phone from 2012, one aperture and one colour of light, so treat it as an illustration. Cambridge in Colour also calls the limit a best case that assumes an otherwise perfect lens.

Tiny pixels also run short of light

Resolution is only half of the problem. A 2016 SIGGRAPH Asia paper by Google researchers says phone cameras have small apertures, which limits how many photons they collect and leaves low-light images noisy. Small sensor pixels also store fewer electrons each, which limits dynamic range, the span from dark to bright detail a pixel can record.

Their answer was to capture a burst of frames, align them and merge them. The authors report that this reduces noise and widens dynamic range, and that it ran in 4 seconds on a phone for a 12 megapixel image.

How phones ended up merging pixels

Makers wanted both high pixel counts and enough light per pixel, and one workaround is to merge neighbouring pixels into one larger pixel. This is called pixel binning. Samsung’s semiconductor division, which sells image sensors, describes it on its own technology page as a way for small pixels, which allow higher resolution but struggle in challenging light, to gather more light together.

The idea is older than it looks. The same Vision Systems Design editor reports that the Nokia 808 combined nearby pixels to give an image of about 5 megapixels from its 41 megapixel sensor. Digit, an Indian technology outlet, described in a 2018 explainer a 2×2 version that merges four pixels into one, turning a 40 megapixel sensor into a 10 megapixel image.

Samsung’s page, which is the company’s own description, extends this to 200 megapixel sensors. In low light, 16 neighbouring pixels merge into one for a 12.5 megapixel image. When light improves, the sensor can switch to a 50 megapixel mode or to the full 200 megapixels.

What merging pixels costs

Binning trades resolution for light. Digit notes that the output has a quarter of the sensor’s pixel count, so a 16 megapixel sensor gives a 4 megapixel image that looks brighter but captures less detail. It also warns that if a maker’s JPEG compression is aggressive, fine detail can degrade until photos look like oil paintings.

The claim also has limits. It does not say smaller pixels are worse. Cambridge in Colour says that once diffraction limits a camera, smaller pixels need not give a worse image. Total resolution is the same, with fewer artefacts such as colour moiré, though noise and dynamic range make the comparison more complicated. It adds that focus errors, motion blur and imperfect lenses are likely to matter more than diffraction in many real shots.

What to check in your own photos

Take two photos of the same scene, one in bright light and one in dim light, and zoom both to 100 percent. Digit says a sensor in low light must raise its sensitivity, which lowers the signal-to-noise ratio of each pixel and shows up as noise. Grain that appears mainly in the dim photo therefore points to a shortage of light at the pixel. Paint-like smearing in both can point to heavy compression or smoothing, the oil-painting effect Digit describes.

Then open the saved file’s details on your phone or computer and read its pixel dimensions. A 200 megapixel sensor that saves about 12.5 megapixels is likely merging 16 pixels into one, as in Samsung’s example. Judge a phone camera by two photos of the same scene taken in the light you usually shoot in.