# How to judge whether the exposure of a photograph is accurate

1. What is a histogram?

One step at a time is needed to determine exposure. When it comes to exposure, I think we have to talk about histogram first. First, let's look at the histogram. Some people may say: What? Histogram? I'll use it when I touch the camera. What can I say about this? Well, it's not too late for you to say that when I've finished. Take this picture as an example: Let's first introduce the most basic knowledge of histogram. The higher the brightness is, the more pixels are represented by the horizontal axis of the histogram from left to right, and the more pixels are represented by the vertical axis from bottom to top. Brightness from 0 to 255, 0 means black, 255 means white. If the peak is higher somewhere, the more pixels there are at this brightness. Take this histogram for example, its distribution is very uniform, indicating that the distribution of pixels in each luminance interval is very uniform. Understand these words, you have a basic understanding of the histogram, but in fact, there are many things about the histogram. 2. How to Read the Parameters of Histogram

I asked a question: Do two pictures with identical histograms have identical pictures? The answer is of course NO, because the histogram records the brightness information of the pixels. In other words, we keep all the above pixels unchanged, just change their relative positions, and the histogram will not change at all, but the content of the picture may be completely different. Understanding this point is very important, and it is very helpful for us to understand the essence of histogram. Okay, all of the above is well-known knowledge, and the following is what I want to talk about, go back to the picture just now. What are the meanings of the colored order, number and percentage on the right side of this picture? When you open the histogram and place your mouse somewhere in the histogram, you will see three parameters, which represent:

Chromatic order: The brightness of the pointer, which is a value from 0 to 255. Number: The number of pixels in this brightness, such as the above image, means that there are 1915 pixels in 138. Percentage: The position of the current color order in the whole color order. Well, there's advanced knowledge in this one. When you hold down the left mouse button and pull right, you will find that they have changed. Chromatic order: The range of color order you choose, such as the one above, is from 115 to 216. Number: The total number of pixels in this range.

Percentage: The percentage here is not the percentage of the position, but the percentage of the pixel in the whole pixel in the range you select. Wait a minute. Some people will say that the total number of pixels in this photo is only 207 284. How come there are 227 7228 pixels in this range? Isn't that self-contradictory? Very good observation, because I choose the RGB channel, the total number of pixels is multiplied by three, later I will talk about each channel. Well, after the above study, you have a better understanding of the histogram, but not enough, there are more complex. Continuing with this histogram, there are several parameters on the left, average, standard deviation, median, and pixels. What do these parameters mean? Average value: The higher the average value, the brighter the photo as a whole, with 128 as the median value. Its algorithm is: the total brightness of the image to the total number of image pixels. Take the picture above for example. Its average value is 117, which is close to 128, so the exposure is normal. Standard deviation: Standard deviation is a statistical term. A measure of the degree of dispersion of data distribution, used to measure the extent to which data values deviate from the arithmetic mean. The smaller the standard deviation, the fewer deviations from the average, and vice versa. The magnitude of standard deviation can be measured by the multiplier relationship between standard deviation and average value. Standard deviation formula: Sample standard deviation S = Sqrt [((xi-x pull)^ 2)/(N-1)], in the formula represents the sum, x pull represents the average of the sample x used, ^ 2 represents the quadratic square root, Sqrt represents the square root. All of these are not important. Just understand. What we need to know is the relationship between the standard deviation and the picture. The bigger the standard deviation is, the more obvious the contrast is, and vice versa. Intermediate value: The brightness value of all the pixels in the image is arranged from small to large, and then the value is located in the middle. The data is divided into two parts, one part is larger than the value and the other part is smaller than the value. (If there are even pixels, there are two numbers in the middle, take the front one.) The meaning of the middle value is to reflect the overall brightness of the picture from the other side, whether it is overexposed or underexposed. It is complementary to the average value, but there is no accurate average value, the specific reasons for my own experience. Pixel: This is not much to say. Everyone is familiar with it.

3. What is the passageway?

After reading and understanding the above, you should have a more comprehensive understanding of the histogram, but you still need some knowledge to really understand the histogram.

There are many kinds of channels: RGB, red, green, blue, lightness, color. First of all, we need to understand that the number and the pixel in the histogram are not the same concept. When we select the RGB channel, the maximum number is equal to the pixel value * 3. When we choose other channels, the maximum number is equal to the pixel value. For example, in RGB channel, the number is 3119 when the color order is 100; in R channel, the number is 945 when the color order is 100; in G channel, the number is 1610 when the color order is 100; in B channel, the number is 564 when the color order is 100. You will find that the number of RGB channels is R + G + B. That is to say, RGB channel is actually the sum of R, G, B channel values. And what are the pixels? We take the final mix of RGB three colors as a color. This is what we call the pixel. I believe you can understand why the maximum number of RGB is equal to the pixel value in a single R, G, B channel. Similarly, RGB channels and intensity channels are different. This is the lightness channel: Lightness channel

Maybe you have a question. Isn't histogram just a reflection of brightness information? Why is the histogram of RGB channel different from that of explicit channel? This is due to the calculation method. Lightness statistics is the composite value of each pixel, and the calculation method of the brightness value of the pixel is: 30%*R+59%*G+11%*B. This corresponds to the previous pixel, which is a composite value. Looking at the front so many, I believe you have some dizziness... Hold on, there's one last thing about the histogram. Let me talk about it in terms of the histogram of the red channel. the red channel

What can you think of when you see this histogram? It shows that the red information is mainly distributed in the middle and dark parts, but not in the bright part. The histogram information of a single channel plays an important role in color matching and correction of color deviation.

4. Cache Level of Histogram

Finally, we will talk about the cache level of histogram. What does this mean? Let's look at the picture first.

I will not talk about its calculation principle, but I will talk about it directly: the higher the cache level, the faster the histogram generation, but the more inaccurate (briefly speaking, the higher the cache level, it will not calculate the value of each pixel, but will merge several pixels into one pixel calculation). If you need to change the cache level to 1, just click on the triangle in the upper right corner. Cache Level 3 Cache level 1

Histogram is very helpful for us to understand exposure. For example: Different exposure levels

The first picture is obviously over-exposed and the ground is obviously under-exposed. The third exposure is right, high light and low light. It's the right exposure. Well, if you believe me above, you still don't understand the essence of histogram. As I said earlier, the histogram records the brightness information of the pixels. In other words, we change the relative position of all the pixels above, and the histogram will not change at all, but the content of the picture may be completely different. It's important to remember this sentence: Histogram records the brightness information of the pixels, and nothing else represents it. The accuracy of exposure is not necessarily related to the uniformity of brightness distribution. The three histograms above correspond to the following pictures: Samples corresponding to histogram

The above point is the limitation of histogram. Histogram only reflects the brightness information, nothing else represents, and it is not necessarily related to whether the exposure is correct or not.

5. Application of Histogram

I mentioned the limitations of histogram before. The reason why I want to talk about the limitations first is that we should not be too superstitious about histogram. We should not form what kind of histogram in our minds is accurate exposure and what kind of histogram is not accurate exposure. So how do histograms work? The answer is in combination with the shooting environment. The role of histogram in photography is obvious. Especially when the sun is too strong to see the screen clearly, it is difficult to judge whether the exposure is accurate. At this time, the combination of histogram can make a rough judgement of exposure. For example, when you take a snowscape, you have to have a lot of pixels in low light and middle tone, which is unrealistic in general. For example, when you shoot dark clouds, you insist that the high-light part have a large number of pixels, which is unrealistic in general.

You need to have a rough estimate of the shape of the histogram in combination with your shooting environment, rather than blindly pursuing low light. Both halftone and high light have pixels. Of course, this requires a certain accumulation of photographic experience. One common quick way to improve is to look at histograms of photos in typical environments. Histogram knowledge, if mastered skillfully, is very helpful for color matching and map recognition (late means of recognition). Another: For exposure, it is not suitable to digitalize it completely, and there is no absolutely correct exposure value. Not that the average must be how much, the percentage must be how much, histogram must be how accurate exposure, it is only a reference, is a tool. Whether the exposure is accurate or not depends on your shooting intentions. For example, if you want to shoot a LOMO, you can't use the normal standard deviation to measure it, because the standard deviation of LOMO is generally higher. For example, if you want to shoot the Japanese system, you can't use the normal average value to measure, because the average value of the Japanese system is generally higher, so whether the exposure is accurate must be discussed in conjunction with your shooting intention.