All along, I thought color correction was difficult, but after reading the book and looking back, I found that the colors I adjusted based on my intuition were not much different from what the book described. It seems that there is not much difference in human aesthetics, and even unconventional methods can lead to the same result!
Color Space#
A color image can be seen as an organic blend of three basic colors: RGB. There are many commonly used color spaces, mainly the following three:
- sRGB: Narrowest color gamut, but used in almost all screens.
- CMYK: Different from general color spaces, used for printing.
- Adobe RGB: Widest color gamut, used for professional products such as digital photography.
Principles of Color Space Selection#
It is recommended to use the color space with the largest gamut, Adobe RGB, because this can retain as much information as possible (even if it cannot be displayed on the screen, it does not mean it does not exist). The method is as follows:
Press Ctrl+Shift+K to open the color settings: Set the RGB working space to Adobe RGB (1988); set the color management scheme to "Convert to working RGB"; click OK.
Color Correction#
Basic Color Correction#
First, let's talk about the three properties of color:
- Hue: Reflects the overall impression of the image.
- Brightness: Reflects the brightness level of the image.
- Saturation: Reflects the vividness of the image.
These three properties are not independent parameters. Any independent change in one of them will affect the overall impression.
Next, let's talk about primary colors and complementary colors.
Primary colors are the basic colors that cannot be further decomposed. They can be mixed together to create any other colors. That is, red, green, and blue.
Complementary colors are a concept. When two colors can complement each other to form neutral gray or black, we call them complementary colors.
Primary colors and their complementary colors are the basis of color correction. They can be clearly seen in the "Color Balance" tool in Photoshop:
The complementary color of red is cyan, the complementary color of green is magenta, and the complementary color of blue is yellow.
Page 107 of the book also explains the accurate color correction of RAW. Since I don't have a device that can generate RAW, I won't discuss it here. It's probably about using a standard color card to calibrate in ACR software.
Side Note#
In Python, to adjust the brightness of an image, you can use the cv2.add, cv2.subtract, and cv2.bitwise (addition, subtraction, inversion) functions in OpenCV. For color images, you can use cv2.split to separate the color channels of the image, and then use cv2.merge to merge the channels. Here is an example using the "snail" image:
import cv2
import numpy as np
# Read the snail image and invert it
img = cv2.imread("snail.jpeg")
bitwise = cv2.bitwise_not(img)
# Separate the color channels of the original image and the inverted image
r1, g1, b1 = cv2.split(bitwise)
r2, g2, b2 = cv2.split(img)
# Merge the color channels in different combinations (6*6*6 combinations)
m = cv2.merge([r1, g1, b1])
cv2.imshow("Merge", m)
cv2.waitKey(0)
cv2.destroyAllWindows()
For more details, please refer to the article Image Processing - Basics (in Chinese).
Principles of Color Correction#
Based on complementary colors, we can easily apply the following techniques:
- If the photo is reddish, increase cyan. If the photo is bluish, increase red.
- If the photo is magenta, increase green. If the photo is greenish, increase magenta.
- If the photo is yellowish, increase blue. If the photo is bluish, increase yellow.
This can enhance color complementarity and achieve color calibration.
Application: Color Balance Tool#
The Color Balance tool can mainly adjust the colors of the midtones, shadows, and highlights. Here is an example:
The overall photo appears slightly blue, so adjust the blue-yellow slider in the midtones to reduce the blue:
The shadows appear slightly green, so increase the magenta:
The highlights are not red enough, so add a little red:
The result looks better.
However, color balance can only adjust the overall color tone and cannot adjust specific areas. Next, we will use the Curves tool to make more detailed adjustments.
Application: Curves Tool and Hue/Saturation Tool#
If you can only choose one tool in Photoshop, undoubtedly it would be "Curves". Because "Curves" can not only be used to fine-tune the brightness and contrast of photos, repair color casts, and create various color effects, but also create black and white, low saturation, high saturation, low tone, high tone, high contrast, low contrast effects. It can be used in conjunction with layer masks to create vignetting, repair "dead white" skies, achieve color separation, complex selections, precise cropping, precise skin retouching, etc.
- Chen Jianqiang, "Photoshop Post-Processing: The Complete Guide to Digital Photography Post-Processing"
The details of the operation are too cumbersome to record 😖. The main idea is to use the "hand tool" in the Curves tool to make local adjustments:
Then use the Hue/Saturation tool to deepen or lighten the colors, and use a brush for local adjustments:
Finally, merge the visible layers and save the image. The final color correction result is as follows: