Traditional Culture Encyclopedia - Photography and portraiture - What is the principle of camera anti-shake technology of iPhone 5s?

What is the principle of camera anti-shake technology of iPhone 5s?

In the current photographic system, anti-shake technology is generally divided into digital anti-shake and optical anti-shake 5s uses digital image stabilization like most mobile phones, and a few, such as Nokia lumia92x, the previous 808 and the latest 1020, have optical image stabilization. There are few mobile phones with optical image stabilization, but there are still some.

The reason why this happens is that optical anti-shake, whether it is lens anti-shake or photosensitive element anti-shake, needs to occupy a lot of extra space to place the mechanical parts that compensate for anti-shake. This demand is difficult to meet in mainstream mobile phone design, so even extreme manufacturers like Apple will give up optical image stabilization for comprehensive consideration.

At present, digital anti-shake is mainly embodied in the algorithm. By improving the sensitivity of the photosensitive element to light, the shutter speed is increased, thus avoiding the picture blur caused by the slow shutter. The improvement of sensitivity will bring more noise, and excellent manufacturers can reduce the influence of noise as much as possible through the improvement of algorithm, so as to obtain higher shutter speed in the same dark environment. But in fact, the meaning of so-called digital image stabilization is more reflected in very early digital cameras. I remember that in 2007 or 2008, the highest ISO of digital cameras generally stayed at 400,800, which was not bad. In fact, it is not that the sensitivity of the photosensitive element can't be higher, but when ISO reaches that value, the imaging has reached the point where it is completely unusable. So back to now, the photosensitive element of the mobile phone can also be 3200 at will, and even in a very dark environment, a sufficiently high shutter can be obtained. Nowadays, mobile phones all adopt digital anti-shake technology, so it is better to say that they instinctively have the ability to obtain high shutter speed in low illumination. So, for today, it's not who uses digital anti-shake technology, but everyone is desperately trying to make high-sensitivity noise control more perfect. Compared with other mobile phone manufacturers, Apple's noise control can be said to be first-class.

In addition, the technology of multi-frame superposition is becoming more and more perfect, which can be regarded as a new generation of digital anti-shake technology. 5S has the ability of high-speed camera shooting, which shows that it has the conditions of multi-frame superposition noise reduction on the one hand, and how to know it in detail depends on the market. (Of course, the actual situation is likely to be negative, otherwise Apple's press conference will say that it is a bad release. ). The biggest problem of multi-frame superposition is that the picture momentum between frames is not exactly the same, which is why Sony's hand-held night scene function often appears ghost hands. In addition, superposition will inevitably lead to image quality loss, so multi-frame superposition can only be regarded as an auxiliary energy source, and it is difficult to surpass optical image stabilization to become the mainstream of camera image stabilization in a short time.

Here I briefly talk about the principle of noise reduction. The so-called noise is the electrical noise produced by improving the sensitivity of photosensitive elements, which makes the imaging of pixel points go wrong. The images brought by this kind of electrical noise actually appear randomly on the screen, and most of their forms are "not harmonious" with the surrounding pictures. The algorithm is to detect and guess whether these discordant pixels are noise, and then turn this noise into "harmonious" points with the surrounding through a certain formula algorithm, so as to achieve the effect of noise reduction. Therefore, the noise reduction should not be too strong, otherwise it may kill the wrong person. In fact, the discordant pixels will be harmonious. . . . . . .