Modern smartphones are already able to compete with full-fledged cameras in terms of shooting quality (for example, we recently did a review of movies shot on smartphones). Even if you don’t plan on doing full-time film production and are just getting into mobile photography, it’s helpful to understand how it all works.
This guide is your go-to guide to help you navigate the world of mobile photography and find the perfect smartphone with a good camera.
The main characteristics of camera phones
A camera phone is a camera that can make calls. Therefore, the matrix is the main parameter. The quality of the pictures primarily depends on it. The main parameter that determines the quality of a matrix is its size. A large matrix is able to accommodate large pixels (we’ll talk about pixel sizes a little later). The size of the matrix is measured in fractions of an inch, denoted by a single quote. For example, like this — 1/3″.
Now to the pixels, their number and size. Let’s start with this example: the camera of the Samsung Galaxy S21 Ultra smartphone (we’ll make a reservation right away — the phone’s camera is good) has 108 megapixels. The full-fledged top camera Sony Alpha A7R III has 42.4 of them. Does this mean that a smartphone can easily take a Japanese camera? Obviously not — the matrix of smartphone cameras is approximately 50 times smaller than in cameras. Namely, the quality of the image depends on the size of the matrix.
We can say that megapixels are physically located on the matrix. These are photosensitive elements. For clarity, let’s imagine them as containers filled with light. The deeper the container, the more light will fit in it, right? It follows from this that if there are many megapixels, and the matrix is tiny, then each individual pixel will be small. This affects the quality of the final photos. When approaching, they will notice a lot of noise.
And if the matrices are the same size, but one of them has fewer pixels, but they are large, and the other has a lot of small ones? Let’s figure it out.
Number of megapixels
Suppose, on average, 4 photons of light fall on each pixel of the matrix. We can calculate the noise of the entire matrix using the Poisson distribution. According to him, noise is the square root of the total number of photons. In our case, the noise will be 2 photons. That is, the brightness of the points in the photo on average will differ by as much as 50%! If we increase the pixel size by 4 times (simultaneously reducing their total number by the same factor), then now one of our pixel can accommodate 16 photons. Applying the Poisson distribution again, we find that the average noise is now 4 photons, or 25%. The photo has become twice as clear. Therefore, large pixels are preferable to small ones.
Another problem with small pixels is crosstalk, i.e. when light from one pixel hits another. Manufacturers are fighting this by choosing more reflective materials for the manufacture of partitions between the pixels.
Modern smartphones use pixel binning. In such matrices, pixels are collected in groups of 4. During shooting, one pair will collect light longer than the other. If all pixels were to collect light for the same amount of time, there would be a risk of overflowing with light from bright spots in the scene. Because of this, in the place of such pixels in the final image, there would be white dots. When using binning, this risk disappears, because the pair that collected less light did not have time to overflow. By combining the information received from these two pairs of pixels, we get an image on which there are no black and white dots, that is, a higher dynamic range.
The main advice that can be given here is not to chase the number of megapixels. Pixel size is more important.
For example, the recognized camera phone Google Pixel 6 is equipped with a wide-angle camera with a large matrix and a pixel size of 1.2 microns. For comparison, the base model iPhone 13 has a pixel size of 1.7 microns on a wide-angle camera. If we compare the quality of the pictures, we find that the Pixel is expected, but not fatally inferior. Google’s algorithms sometimes brighten shadows and also smooth out bumpy textures, making the picture flatter.
The quality of the image does not depend on one matrix. Camera phones in 2022 are no different from older camera brothers: smartphone photos are strongly influenced by the quality of the lens.
If a bad lens is installed in front of a good matrix, no matter how many megapixels and what size they are, the pictures will be fuzzy and with chromatic aberrations.
Most modern smartphones have several cameras that differ just in lenses: a standard module that is turned on by default, an ultra-wide-angle lens and a telephoto lens. But the ultra-wide-angle lens is most likely not as wide as that of full-frame cameras (after all, smartphones have small sensors, which implies a large crop factor). And telephoto lenses can tolerate distortion.
You can fix this by using external lenses. A selection of lenses for smartphones can be viewed here.
On average, the aperture value in smartphone lenses ranges from f/1.7 to f/2.2. The size of the aperture determines how much the lens is able to open, how much light it is able to pass onto the matrix. For example, different apertures are set even within the same iPhone 13 line (more on this in our analysis). Also for comparison: here is how the aperture numbers differ for smartphones of different price categories:
– Budget Samsung Galaxy A12 with f/2.0;
– Mid-price Samsung Galaxy A52 cf/1.8;
– Flagship Huawei P40 Pro with f/1.28.
In addition, it depends on the aperture whether bokeh is obtained. But since the size of the matrix also affects the bokeh, on smartphones the background will be blurred only when macro shooting small objects.
To compensate for this, smartphones implement portrait modes — creating bokeh programmatically. This approach has its own problems. Sometimes algorithms blur hair along with the background, or, for example, the temples of glasses. Algorithms are getting better from model to model. So, for example, they work better on the iPhone 13 than on the 11th model.
Viewing angle and zoom
Among smartphones, there are models with cameras that can provide a wide angle. Usually this is an additional camera in addition to the main one (more on this below).
As for the zoom, in most smartphones it is implemented digitally, that is, the picture is simply cropped and scaled. This method will work well only on a steep matrix. Full zoom, which allows you to shoot a larger object without losing quality, is rare and does not give more than two / three times the magnification. But there are also external lenses.
Different smartphones may have different auto focus technologies:
- The most common option is contrast autofocus, which is found in most smartphones. Its principle of operation is as follows: the camera looks for contrast zones, then moves the lenses, comparing the contrast at their different positions. Such focusing is slow, works better when shooting static objects, and in poor lighting, which means poor contrast, it may not work at all.
- Laser autofocus works due to the emitter located next to the camera. The emitter emits a beam, which is reflected from objects in the frame, and the smartphone calculates the distance to them. This kind of focusing only works if you are shooting close up. When focusing on distant objects, the camera will switch to contrast or phase focus.
- Phase-detection autofocus works based on sensors that analyze the doubled image of the object, adjusting the focus so that they match. This focus is faster than contrast, in daylight it will be able to focus on fast moving objects. But it doesn’t work well at night.
Different manufacturers may have their own implementation of phase detection autofocus. For example, Apple calls it Focus Pixels, while Samsung calls it Dual Pixel. These implementations differ in that special sensors are built right into the pixels on the matrix. In Dual Pixel, they are built into literally all pixels.
Smartphones have three main types of stabilization:
— digital stabilization in most smartphones is used when shooting video. Since the video recording uses only the central part of the camera matrix, the rest of it is designed to compensate for shaking by shifting the image in one direction or another beyond the original frame.
— optical stabilization works due to a special mechanism that shifts the lenses in the opposite direction of shaking, allowing you to maintain sharpness.
— hybrid stabilization implemented, for example, in the Google Pixel 4 smartphone. This is a combination of optical and digital stub, which is controlled by special algorithms that analyze what is happening in the frame.
In addition to the quality of the matrix and lens, video shooting on smartphones depends on the central processor. The video has three main parameters:
- number of frames per second (determines smoothness);
- bitrate (determines the quality of the image).
In order to shoot a good video, in which all these indicators will be high, you need a powerful processor. Also, if the device is prone to overheating, it will turn off every 15 minutes of video recording — you can’t shoot that much.
This problem is solved, for example, in OnePlus 9 Pro, as the smartphone is equipped with an additional cooling system.
Number of cameras
As we wrote above, different cameras can have different viewing angles, allowing you to shoot, for example, with a zoom on one, and wide-angle on another.
Also, increasing the number of cameras improves the quality of photos. For example, a smartphone will take a series of pictures on them with different shutter speeds. Then the algorithms glue the photos together — the technology works identically to pixel binning.
Flashes in smartphones illuminate only a small object well at small distances. They also shine directly. Therefore, strictly speaking, smartphone flashes are more suitable for use as a flashlight.
Modern smartphones use LED flashes. Either with a single diode, or paired, where the diodes have different color temperatures to improve color reproduction and reduce the likelihood of overexposure.
Why physical characteristics are not always the most important
Most of the shortcomings of mobile photography, which are of a purely physical nature — the small size of the matrix, for example — are solved by algorithms. The processing power of modern smartphones allows neural networks to complete photos.
Thanks to this, the smartphone is able to cope with night shooting, in which professional cameras can fail. Neural network algorithms in smartphones are also used for other tasks: zoom, detail enhancement, noise reduction.
The smartphone starts taking pictures when you just launch the application. And in order, for example, to achieve bokeh when shooting a person (and above we wrote that this requires a wide matrix and a fast lens), neural networks in a smartphone perform complex calculations.
First, they segment the image, that is, determine which objects are in the image. This is necessary to separate objects from the background. Then she builds a depth map, and only then blurs the background.
Phones are a full-fledged combat unit of a professional photographer. A duet of pumped iron and algorithms allows you to take great pictures. You just need to be mindful of their limitations and make informed choices.