Why the race for megapixels in smartphones is absurd

How the smartphone camera works

A camera is a complex thing: it combines a sensor, an optical system, a controller and other auxiliary components, as well as software for photo and video processing. Let's consider each element in more detail.

Matrix

Why the race for megapixels in smartphones is absurd

The matrix is a rectangular microcircuit consisting of light-sensitive elements - pixels. Each pixel contains three subpixels. One sub-pixel transmits only certain wavelengths: for red, green or blue (red, green, blue). This color model is called RGB.

Also, the matrix can be monochrome, without color filters. Three times as many photons fall on each of its pixels. As a result, black and white photos are sharper. Such matrices can be used to enhance the color image from another camera module.

One of the main characteristics of a matrix is resolution. It reflects how many pixels fit on it.

Lens

The tiny smartphone lens is almost a piece of jewelry. A rare system includes 4–5 elements - usually 7–8 or more.

In smartphones with multiple cameras, each matrix will have its own lens. Each of them solves its own problem:

Telephoto lens (telephoto) is needed for shooting from a long distance.
Wide angle (shirik) will help to fit more objects in the frame - this is useful for group photos and architecture photography.
Universal the lens will allow you to moderately well shoot any subject: from a portrait to a landscape.
Varifocal lens (zoom) can bring the subject closer.

Lenses for smartphone lenses are made of glass or special polymers. If their transparency is far from ideal and the elements are not well-fitted, do not expect good photos. Even if the lens moves a few microns, the optical system will defocus.

Diaphragm

The diaphragm is the hole through which light enters the camera. How much light the sensor can receive depends on it. The aperture value is output in f / 1, 7 format.

Stabilization system

Stabilization compensates for blur from camera shake, such as when shooting handheld rather than using a tripod. It can be of two types:

Optical. An honest electronic-mechanical system that physically holds the camera in one position (at least it tries). It gives you sharper photos with minimal noise and virtually eliminates the need for software processing.
Electronic. These are software algorithms. The camera still shakes, but by analyzing several frames, a more or less decent result is created.

Autofocus system

Autofocus itself determines the distance to the object and adjusts the parameters of the camera's optics accordingly. Three types of systems are used in modern smartphones:

Phase. Special sensors collect light rays at different points in the frame. The light is then split into two streams and sent to a light sensor to determine the distance to the object. Advantages: high precision and speed of work. Disadvantages: high price, complexity of the design and its settings.
Contrasting. The contrast of the scene is analyzed. By shifting the lenses, the camera tries to maximize the contrast of the subject against the background. Advantages: compact size and low cost. Disadvantages: The system is slower and is not well suited for dynamic scenes.
Hybrid. Combines phase and contrast focusing to get the best results.

Software

Software can also be considered part of the camera, because it is directly involved in obtaining the result of the shooting. Today, no smartphone gives you frames as they are, without software processing. Sophisticated algorithms, often using a vast database or artificial intelligence technology, edit each shot to "make you beautiful."

Raw images will not be bright or clear enough. The software removes overexposure, draws out dark areas, improves colors, increases sharpness. And it does all this automatically and very quickly.

But there is also a downside to the coin. Aggressive noise reduction can make a photo taken at dusk that looks grainy - as if it contains many small specks. This degrades detail and makes colors unnatural.

What does the number of pixels affect?

The detailed specifications of a smartphone usually indicate the physical size of the camera matrix - something like 1/2, 6 ″. On the manufacturer's website, you can find data on the pixel size in the matrix. This parameter affects the number of points in the frame. The higher the resolution, the better the details are reproduced.

But if the pixels are small, each of them receives little light and cannot accurately determine the color of a point in the real image. As a result, noise appears in the photo.

Noise is dots of random color and brightness scattered throughout the frame. The worse the illumination and the lower the quality of the camera matrix, the more noise there will be in the photo.

Its number in the frame is proportional to the pixel size or the square of the matrix diagonal. If we compare two matrices with points of size 1, 55 µm and 1, 1 µm, then in the frame with the first there will be half the noise.

The dynamic range of the matrix is also important - its ability to capture the entire spectrum of colors and brightness of the surrounding world. The cheap ones have a small range, and the photos turn out to be faded, hazy.

Why smartphone manufacturers are chasing pixels

Because buyers always want the most. Even if in a car for 300 horses you have to stand in a traffic jam or play solitaire on a cool gaming computer.

Which smartphone would you buy for the same price: with a 12MP camera or a 48MP one? Choosing the second, you get four times more megapixels for the same money. But your photos won't get four times better.

A sensor with a lot of small pixels is cheaper and better than a sensor with large pixels.

Large matrices take up more space inside the smartphone. The optical system for them should also be larger. Accordingly, there will be less space for the remaining parts in the body. The smartphone will get thicker or the camera will stick out. It will have to be protected with tempered or sapphire glass. And this is also money.

Selling a fat, expensive smartphone is difficult. It is easier to order matrices with a large number of small pixels and conduct a loud marketing campaign: add an automatic stamp “shot with a 48MP supermegaflagman” to the camera photo, so that everyone knows that someone has bought a new smartphone. And let fans and pros use DSLRs.

Although Nokia, for example, took a chance and got the legendary smartphones Lumia 1020 with 41 megapixel cameras. And this is in 2013!

What does the quality of the photo really depend on?

Matrix and pixel size

If you take two matrices of the same resolution, then the photo of better quality will potentially be obtained with the larger of them. There, the pixels are larger, which means that more photons fall on each one when shooting. As a result, subpixels can more accurately determine the color of a specific point.

It would seem that if in one matrix the pixels are 1, 4 microns in size, and in the other - 1, 2 microns, they are practically the same. But 17% is a tangible difference that will definitely show up in the quality of photos and videos, especially if you shoot in low light.

Another important point is the distance between adjacent pixels. In small matrices, manufacturers frankly save on it. In larger ones, they can allow you to qualitatively separate neighboring pixels so that they do not affect each other.

Production technology

New methods make it possible to more accurately determine the intensity of the light flux from fewer photons, which means that you can ensure low noise and good color rendition, even if you shoot at dusk without a flash.

But you need to read and analyze. For example, the HTC One (M7) smartphone offered UltraPixel technology. The manufacturer promised a serious increase in the quality of photos and videos.

In fact, the UltraPixels turned out to be just larger 2 micron pixels. Can this be considered a new technology? Unlikely. For comparison: the Google Pixel, which HTC also assembled and which at one time was considered one of the best camera phones on the market, had a matrix with pixels of 1.55 microns. The size of the camera was not increased so as not to increase the thickness of the smartphone. The matrix resolution of 5 megapixels was small even for 2014. As a result, there were no queues for HTC One (M7).

Another example is technologies like Super Pixel or Quad Pixel. Four adjacent pixels of a large matrix are combined to obtain a photo of a lower resolution, but of better quality. The solution is purely software. If the matrix is so-so, the efficiency will be low.

Stabilization

Optical stabilization is always better than digital. Post-processing algorithms will still be applied to the frame, and it is better if it is sharp initially.

Zoom

To get closer to the subject in the frame, the optical zoom shifts the lenses, and the quality of the photo is practically not affected. Digital zoom stretches part of the picture to fill the entire frame. This feature is available in any photo editor, often even in a standard camera application. Therefore, it doesn't make sense to pay for digital zoom.

Autofocus system

Contrast AF is an inexpensive system for mediocre cameras. Phase detection autofocus is suitable if you are shooting fast running children, cats or athletes. But the ideal option is a hybrid system that combines the advantages of phase detection and contrast detection autofocus.

Diaphragm

Since a smartphone is used for shooting in a variety of situations, a camera with a larger aperture will benefit: f / 1, 7 is better than f / 2, 0. The higher the value (or the lower the number after the slash), the higher the lens aperture and the more effective it will be. work at dusk or indoors.

Brand name

Yes, it is not only an advertising tool. It happens that the same matrix is installed in the Chinese smartphone and the flagship of the A-brand. But the output shots are very different.

If the manufacturer does not invest effort and money in the development of components, technologies and software, you should not expect beautiful, clear frames. If he saves on everything, for example, puts on cheap lenses with poor transparency, then this will affect the result.

What to remember

Dozens of megapixels are primarily marketing. The quality of photos and videos does not directly depend on them.
Even 5 or 8 megapixels is enough to print a good quality picture on a landscape sheet. The 4K screen resolution of the leading TV is about 8-9 megapixels. Full HD - only 2 megapixels.
Larger pixels collect more light. The result is a crisp, well-detailed frame with natural color reproduction and no noise.
If you don't want to bother with theory, go to practice. Comparative reviews of smartphones and photos from cameras (full-size and cropped - cut and enlarged fragments) will make it possible to understand the real state of affairs.