Earlier, Huawei announced at the press conference that the Sony IMX278 sensor first launched on P8 was a new RGBW sensor. Many people had doubts about what the so-called RGBW sensor meant? What role does it play? How much help is it for shooting? Next, we will analyze through specific samples.

RGB， Red, Green and Blue, the three basic colors that can no longer be decomposed, can get a whole colorful color world through the superposition and mixing of different proportions. When the light passes through the CMOS photosensitive element, each RGB pixel densely arranged on the CMOS will record the real color transmitted and can be detected, and then process, convert it into digital signals, store them on the device, and finally display them on the screen. This is the whole process of the birth of a digital color photo.

In this process, it is the pixels adjacent to "RGBG" on the CMOS that determine the color of the proof. Bryce Bayer, a scientist of Eastman Kodak Company, learned in his research that people are most sensitive to green objects, so he used 1 red, 1 blue and 2 green RGBG arrangement to simulate the colors seen by human eyes. The principle is that when light passes through each pixel, it will be filtered by the corresponding color filter, The rest corresponds to the intensity value of this part of the color. The sensor itself can only get the intensity data of the color, and then the color intensity data filtered with RGB will be returned to the processor. However, each pixel filtered out only one color, and other colors adjacent to this pixel must be filled by "color guessing", that is, the color image can be obtained by interpolation using the anti mosaic algorithm, which is one of the reasons why color restoration is sometimes inaccurate. To go a little further, this color restoration process will cause color interference, which is often referred to as moire, when taking pictures with highly repetitive details. To reduce the effect of moire, you can use a low-pass filter to blur the details, making it a whole color block. However, this will also reduce the resolution of the camera, which is a method that has both advantages and disadvantages. In recent years, some image companies are also introducing new sensor arrays, such as Sigma's X3 and Fuji's X - Trans sensors, but these are the last words, and now they are back to Bayer array.

What about the so-called RGBW? Smart students may have guessed that this "W" is white in color, that is, change the arrangement of RGBG from 1 red, 1 blue, 2 green to 1 red, 1 blue, 1 green, 1 white. What are the benefits of doing this? In theory, because a white filter directly replaces one of the green filters, and the white filter will not filter the light, the amount of light will be more. In theory, the samples taken in low light should have better brightness. However, just because a green filter is removed, the color performance may change slightly.

Next, we can roughly feel this change through the performance of some samples. By the way, we can try our own display device and eyes. The mobile phone models participating in the sample comparison are OPPO Find7 (IMX214), R7 Plus (IMX278) and Xiaomi Note (IMX214). Why choose these three phones as a comparison? First of all, there are not many mobile phones using IMX278 at present. There are only Huawei P8 and R7 Plus on the market, and there is just one R7 Plus on hand. Then we should also add Find7, which is also a member of OPPO and uses IMX214, and then add Xiaomi Note, which is now commonly used to carry IMX214, for comparison. The selection of these three phones is only due to market share, and does not mean that these phones are the best models of IMX214 and IMX278.

Some readers may wonder whether IMX278 is an RGBW sensor. Here are two official Sony documents. At present, only IMX278 is an official RGBW sensor, which is believed to be convincing enough.

(Click the magnifying glass in the upper right corner of the following picture to view the original large picture)

Let's first take a look at the comparison between the two prototypes of Find7 and R7 Plus. First, let's put aside the comparison of focusing, analytical power and other aspects. At first, the color performance between the two prototypes is similar. In order to control the same exposure, the focus and metering points of the two prototypes are set at the black car in the figure. The spot metering of the two mobile phones is relatively sensitive, and the exposure is basically the same.

Then, through 100% magnification, carefully observe the color contrast of the samples. The green performance of the two is basically very close, which is almost indistinguishable to the naked eye. However, in the sample, the gray and red of R7 Plus have a slight decrease in saturation. By observing the backlight edge of the street lamp, the suppression effect of IMX278 on edge purple light has also been greatly improved compared with IMX214.

In contrast to the night scene, the R7 Plus has made great progress compared with Find7. Both of them are shot in automatic mode. The sample of Find7 should be the original night shot level of IMX214, without any optimization. The R7 Plus has obviously sharpened and smeared noise, so this night shot effect is not the original effect of IMX278. Simply speaking, the high ISO of Find7 makes the sample brighter and has more details. Of course, the result is serious color noise. According to reason, the night shot effect with higher brightness that IMX278 should have had did not appear. The less noise is also the result of post smear processing, which does not mean that IMX278 is all responsible.

Then on the basis of such comparison, we add the night shot sample of Xiaomi Note. Xiaomi Note also uses IMX214, but it adds optical anti shake, and we can see that Xiaomi Note will also apply noise to the night shot sample, but it should not add details sharpening. From this point of view, even IMX214 can still get excellent night scene samples with proper post-processing.

In addition, there is no need to tangle too much between the two sensor models in terms of color. Since most manufacturers will post process the samples, the color control between different manufacturers is also very different. For example, the slightly warm color toning like Xiaomi Note will also greatly change the color of the samples, For another example, the sudden white balance misalignment and other problems will affect the color performance of the proof, and the daily shooting process will also encounter large and small problems, which are also intertwined.

So it seems that whether IMX214 or IMX278 is inserted into your mobile phone, as long as the manufacturer optimizes the camera well, their shooting effect will not be bad. The IMX278 using the new RGBW array is not as divine as the legend. Good hardware must be combined with better software to get the best play. Basically, mobile phone manufacturers will adjust their sensors according to their mobile phone needs. The same sensor model, products from different manufacturers, and even products from the same manufacturer but different series may have different color performances, different proof processing, and different imaging styles. You can't just judge whether it is good or bad according to the hardware model. You have to compare it yourself to know whether it is good or bad.