Professional color correction in Photoshop. Selective Color Correction in Photoshop
Quoting one of the film theory textbooks: “color correction (color setting, color synchronization) is the adjustment of the color components R (red), G (green) and B (blue) to change the overall color and visual style of the image.” And, despite the usual incomprehensibility and intricacy of such theoretical works on cinema, this is one of the most understandable designations.
Many beginners film producers they don’t always understand why color correction is needed - they say, it was necessary before, but now it’s still digital and you can do without it.
They are right about something. Previously, when everything was done on film, color correction was always necessary, because there are no two identical films - in the process of their production, minimal deviations from the standards are still allowed. Now everything is subordinated to the digital world and all processes are subordinated to His Majesty the computer.
However, the need for color correction has not disappeared - the frames still differ from each other, and no one eliminated some of the crankiness of lighting and cameramen (especially on television). The whole need for color synchronization lies in the need to even out the color and hide all the flaws in the shooting so that the final product looks harmonious and complete.
Also, color has a very good effect in cinema - as was written earlier - color perfectly reflects the mood of what is happening in the frame. It’s not for nothing that the expression “set the tone for everything that happens” exists - color directly affects perception.
The color correction process itself is divided into:
- primary color correction
- secondary color correction
If you explain everything in simple words, primary– basic adjustment of colors in relation to “normal” (as intended).
Secondary– this is the process of matching already corrected frames in color with each other. And this is where those additional opportunities open up to hide any shortcomings of the video - adjusting the dark and light areas of the image, shadows, light sources. There is also a more precise “adjustment” of color.
Returning to the statement “color correction is only for film and only manually” - the advancing technical process by leaps and bounds makes it possible to make even the oldest film films automated and digital.
Digital Intermediate
This is a system that scans film and converts it into digital mode. Using it, you can digitize an old film or process a new film in post-production, which was still shot on film. Of course, this process makes color correction much easier and faster.
Why faster? This is because, in the standard case of color correction, frames are corrected one by one. In the case of DI, you can adjust many frames at the same time - if the frames are connected, they are automatically adjusted. The process of restoration and restoration of damaged footage or, for example, chronicles is also possible. This is such an intricate thing.
If we talk about the process of color correction itself, in which a person himself adjusts desired color– this process is complex, but incredibly interesting, because you can influence the entire perception of the film, you can change details. Nothing stops you from painting the tree blue and the people pink...
Color correction is complex and it is necessary to understand the basics, combination and even some symbolism. It all starts with color wheel(wheels), color palettes. It will be familiar to anyone who has ever had experience with three-way color correction. This is a standard tool both in working with color correction, and in color theory in general, where it is necessary to determine a set of combinations of the most suitable colors. In its simplest form, the color wheel consists of 12 colors based on the RYB (subtractive synthesis) color model.
In the RYB color model, the primary colors are red (Red), yellow (Yellow) and blue (Blue). The three secondary colors - green, orange and purple - are derived from mixing primary colors. And the remaining 6 tertiary colors arise from mixing primary and secondary colors.
Notice that warm colors are on the right and cool colors are on the left. Warm colors are bright and energetic. Cool colors have a calming and quiet effect. Then there are several basic color combinations, which can be seen very often in movies. The easiest thing, in my opinion, is to simply have pictures in front of your eyes (to explain to you what and how).
Before you start getting acquainted with color correction, it is worth clarifying that this topic is very broad. To engage in color correction at the proper level, it is better to take special courses where you will be taught the basics of color harmony, taught how to correctly combine colors and use existing methods and methods of color correction. And it is advisable to obtain a certificate. In the age of digital technology, photo processing is a very profitable activity. And the most profitable investment at all times has been and remains education.
We will look at the basic universal method of color correction.
Before you make color correction in Photoshop, you need to understand what it is and for what purpose it is used.
Color correction is a change in the colors, tones and saturation of an image, used either to improve the picture or as a creative technique. The first case may include the need to get more realistic colors or make the photo lighter. After all, with certain settings of photographic equipment, colors can be transmitted distortedly, not the way we see them in real life. This also includes photo correction to increase color saturation to make the photo more attractive. In the second case, color correction will be suggested to you by your own imagination. This could be vintage color correction, fantastic colors of landscapes and the like.
Color correction in Photoshop is done on adjustment layers. If color correction is applied to an image layer, the changes to the image will be irreversible. Adjustment layers work like filters. All effects of the adjustment layer will be visible in the image below this layer. Also, the adjustment layer will allow you to make changes to the final result, if necessary. We discussed the topic of layers in a previous article.
Automatic color correction
The simplest and quick way for beginners - automatic color correction. Open the image in Photoshop, create a duplicate of the image layer ( Ctrl+G). Go to the duplicate layer and click Shift+Ctrl+B. This Photoshop command automatically adjusts the contrast and color of an image, automatically detecting shadows, midtones, and highlights.
This is what the pictures look like before and after automatic color correction.
Hue/Saturation
Open the image in Photoshop. On the layers palette, find the list of adjustment layers by clicking the half-filled circle icon.
Select from the list “Hue/Saturation”/Hue/Saturation.
In the layer settings dialog box, you can change “Hue of flowers”/Hue, "Brightness"/Lightness(make the photo lighter or darker) and “Color saturation”/Saturation(make faded or rich shades).
An image can be divided into color channels. The settings allow you to work with all color channels simultaneously or with one.
When working with a separate color channel, selecting a specific hue that needs to be changed, use the tool "Pipette". Click on the tool icon, move it to the desired area of the photo and make one click. You will see stops on the gradients. On color gradients, you can limit the color range, then changes will occur only in it. By moving the limiters, you set the operating range.
Next, by moving the hue, saturation and brightness sliders, all that remains is to select the settings according to your task. Let's give this photo a purple tint to get a more colorful sunset. To do this, select the blue channel. Drag the range stopper on the gradient to the right to capture the range of purple shades. Move the slider closer to purple Hue, add saturation. When finished, close the settings window.
That's how it happened.
You can learn even more about Photoshop in the course at Fotoshkola.net.
Curves
Adjustment Layer "Curves" has more abilities, which we will look at in the basic method for beginners.
Open the image, call the adjustment layer "Curves" from the list of adjustment layers.
A settings dialog will open. Initially the curve appears straight. We are interested in the tool "Pipette". There are three of them. The first is responsible for shadows, the second for midtones, and the third for highlights.
Now we take the eyedroppers in turn: first click on the blackest part of the photo, second on the gray part, third on the whitest part.
With each dropper you use you will see changes. Curves of RGB color channels (red, green, blue) will appear on the graph. When finished, the curves window can be closed.
In the end it will turn out like this.
Levels
For the adjustment layer "Levels" We will also consider only the basic method of application.
A raster image, and in this case the image of our photos, consists of dots. These dots each have their own color. The black, gray and white points in the image are responsible for saturation, brightness and light. Adjustment Layer "Levels" allows you to change the point value level. Level 0 - black pixels, 255 - white. Level 128 - gray. The remaining levels range between 0 and 255. Redistributing the levels changes the tonal range of the image.
For quick color correction, you need to redistribute the level of midtones. Open the image, select from the list of adjustment layers "Levels".
In the settings dialog box, select the middle eyedropper, which is responsible for midtones. In the image, click on the area where there should be perfect gray. Then close the settings window. In this way, equal values of red, green and blue are selected.
As a result, we get a balanced, rich picture.
Since color correction in Adobe Photoshop is an incredibly broad and multifaceted topic, I want to immediately explain what this article is actually about and who it is addressed to. So, this article is addressed primarily to beginning retouchers, who, I hope, will be able to learn from it useful information O simple methods solving very specific color correction problems. Simply put, this article was written for those who want to quickly master simple but effective methods to immediately try them out in practice. For those who are interested in a deep understanding of the process, I recommend turning to works that are more academic in style. For example, to the excellent articles by Andrey Zhuravlev, published in his LiveJournal and on this website.
Simple skin color correction
I think the place to start is with skin color, since this is usually the biggest challenge for anyone learning Photoshop.
Getting down to business, first of all, I would like to advise you not to forget about the subjectivity of color perception, and not to get hung up on “processing by numbers”. When correcting skin color in a specific photo, you should also not forget about the context, shooting conditions and other features. For example, it is worth keeping in mind that a bright environment (say, dense green foliage) will inevitably give reflexes to the skin, the complete removal of which will make the photograph unnatural.
However, there are certain patterns that we will focus on.
The famous Dan Margulis, a color correction guru who devoted a lot of time to researching this topic, at one time developed the “correct” color ratio, which many retouchers rely on to this day. This ratio (in CMYK color model values) looks like this: in the case of normal skin For a Caucasian person, the value of Y (yellow) should be slightly (very slightly) greater than the value of M (magenta), and the value of C (cyan) should be 3-5 times less than the value of M (magenta). As a rule, if the skin in the photo is not in deep shadow, the K value is zero, since there is no need to add black to create a normal tone - darker (tanned) skin is achieved by increasing the C, M, Y components, maintaining proportions. It is also worth noting that with age the amount of yellow increases slightly. Therefore, in infants, the values of the components yellow (Yellow) and magenta (Magenta) are almost equal, but old people are noticeably yellower.
Now, in order to apply the knowledge of this relationship in practice, we will use the tool Color Sampler Tool. It allows you to mark points on the photo for which we will see the current color value in the palette Info.
To see the average skin color for the selected area, and not the color in one pixel (which for many reasons can differ significantly, which will certainly mislead us), increase the value of the parameter Sample Size from "Pint Sample" to, say, "11 by 11 Average". Now we will see the average value for a square of 11 by 11 points.
To further reduce the risk of accidentally measuring in the wrong place, we will place two or three points in different parts faces. They should be placed on moderately lit areas of the skin, avoiding glare and shadows. The nose with the bridge of the nose and the chin are best suited. Of course, the cheeks fit even better, but it is important for us not to get the blush on. Please note that an already established point can be freely moved around the photo.

Having placed the dots in the window Info change the color space for them to “CMYK Color”. To do this, click on the small triangle next to each eyedropper. Naturally, our image remains in the same color space in which it was, it’s just that now for the points we selected, the color corresponding to the current CMYK colors will be displayed.
Now all we have to do is correct the color, bringing the CMYK values closer to those described above. To do this, create an adjustment layer with curves.
Since we are working with an RGB image, the Red, Green and Blue channels are available to us in the curves. When raising the red channel curve, we will reduce the value of C (cyan), and when lowering it, we will increase it, since these colors are opposite. The opposite for M (magenta) is green, and for Y (yellow) it is blue. Don't forget that changes in any of the channels affect the other two.
As soon as you start making adjustments, in the window Info A second column with color values will appear. Now in the left column you can see the values before the changes, and in the right column - after.

If for some reason you don't want to work with curves, you can use (also through an adjustment layer) the Selective Color tool. This option can be convenient in cases where you only need to slightly correct the complexion of the face, and you do not want to affect the rest of the image. By setting the Colors parameter to Red or Yellow, you can adjust the tone without affecting the other colors.

In conclusion, I note that the described method helps well in cases where you feel that there is something wrong with the skin color in your photo, but you cannot independently determine what exactly it is. In addition, it often helps (especially novice retouchers) to find a kind of starting point for further manipulations. I do not advise anyone to rely on it as an immutable truth. In real life, we attach considerable importance to skin color, for example, noticing by its change the mood of an interlocutor who is flushed with embarrassment or, conversely, pale with fright. And the perception of this color is always subjective. Remember that the viewer will evaluate the result of your work with their eyes, and not with a Photoshop eyedropper. So you should rely on them first.
Fighting color spots
Often, when processing a portrait, we are faced with areas of skin that differ in their shade. It could be age spots, not properly tinted by a hack makeup artist, or specific reflexes that appeared during the shooting. Often these spots are enhanced by “creative” processing at the raw conversion stage, especially if it includes significant shifts in white balance and active work with tone.
Perhaps, in an article for beginners, one should simply advise creating a blank layer, applying it in Color mode, and painting the interfering spot with a brush, taking a color sample with a pipette, and selecting the optimal transparency. By the way, many people do just that. But I don’t want experienced retachers, if they have the misfortune of reading this text, to immediately throw rotten tomatoes at me. Therefore, it’s better that I tell you about a slightly more pleasant method that allows you not to destroy with one movement all the variety of shades and transitions inherent in the natural structure of living skin. However, the proposed method is very simple, and allows you to save time on manually creating masks (even if you already know how to create them using select color range).
For clarity, let’s take a photograph in which we see the face of a very attractive girl, but next to the corner of her mouth there is a completely unaesthetic smeared spot, the secret of which history carefully keeps silent. This is what we will fight with.

Let's create an adjustment layer Hue/Saturation:

We select the color range we are interested in and arm ourselves with a pipette Subtract from sample:

Don't forget to hold Ctrl, click on it on our unfortunate spot. In the window Properties this highlights the range we are interested in:

All further manipulations will be applied only to it. To see this range in your photo, move the Hue slider to the far left position. Most likely, you will see how not only the spot we are interested in has acquired a shade, but also large areas of the skin adjacent to it:

To solve this problem, let's refine the range by moving its edges until the influence of the adjustment layer is limited only to the area of interest to us:

After specifying the range, return the slider Hue to the original (zero) position. Now we can carefully move it in the desired direction, compensating for the shade of the spot until it matches the normal skin color. I think there is no need to explain that to combat red spots you should shift Hue to the right, and to fight the yellow ones - to the left. You can also use the slider Saturation to lightly lighten the stain, and Lightness to compensate for brightness changes, if any:

“Accelerating” colors in LAB
I think we paid attention to the skin sufficient attention, so it's time to move on to more interesting manipulations with the image as a whole. I’ll start with the most popular trick, which allows you to literally make a picture much “juicier” in just two clicks. Let’s take this street photo as a source:
We will make it “juicy” in the color space LAB, convert the image into which you can use the menu Image => Mode:

Now all we need is to create an adjustment layer Curves and then in the channels a And b shift the extreme points of the curves, as shown in the screenshot:

 
The more vertical we give our curves (in fact, as you can see, they remain very straight), the more colorful our photo will become.
If you only want to add color, avoiding the shading effect, make sure that the curve, as before your intervention, passes strictly through the center of the grid (that is, the top and bottom points should be shifted an equal distance). On the other hand, no one forbids you to experiment with shades, which is what I did, giving the cloudy photo a little sunny warmth:

This is just a trick. I hope you will understand that it is impossible to clearly explain how this works within the framework of one article. For those who want to understand the issue, I recommend reading the book by the above-mentioned Dan Margulis, which is called “Photoshop LAB Color. The mystery of the canyon and other adventures in the most powerful color space." The publication is very popular, so finding it will not be difficult.
Coloring a photo using curves
No less interesting manipulations can be performed in the more familiar RGB color space, also using just one tool - curve.
As in the previous method, we create an adjustment layer Curves, and start working with individual channels. Only this time the curves will actually bend. In essence, we will simply color the image in the tone we need by manipulating the curves responsible for the red, blue and green channels. The trick is that we will give different shades to areas of the photo that have different brightnesses.
For example, we can raise the blue curve, but not along its entire length, but only in the area corresponding to dark tones. The dark parts of the image will take on a corresponding tint, becoming cooler. At the same time, we will raise the curve responsible for the red channel in the light area. Now the lightest parts - the sun's glare on the ground and the model's skin illuminated by direct rays - will turn reddish, creating the illusion that the photo was taken at dawn.
 

Naturally, your creative imagination should in no case be limited to just simulating natural lighting conditions. By using the blue and green channels, I easily got this interesting effect:
 

Don't forget that curves can not only be raised, but also lowered. At the same time, by lowering the curve responsible for one or another color channel, we enhance the color opposite to it (for red it is cyan, for green it is purple, and for blue it is yellow).
This way we can color our photo in opposite colors using just one curve. For example, by giving the blue curve the inverse shape of the Latin S, we will “cool” the dark areas and make the light ones “warmer,” thus obtaining the classic contrasting separate toning:

Notice how much I raised the “dark” part of the curve - cool shadows look quite natural in the photo, so why not. But I omitted the “light” part much less significantly, so as not to yellow the model’s skin.
However, we will return to the topic of toning in another article. And here I will focus on just one more small nuance: if you do not want to affect the brightness and contrast of the original image and strive to work only with color, do not forget to change the blending mode of the adjustment layer to Color.
Selective Color
If you are too lazy to mess with curves, but you want to create crazy (or not so crazy) shades, there is a very basic, but very visual way for you to colorize photos.
It is based on the use of a tool Selective Color, so we'll start by creating the appropriate adjustment layer:

As you, I hope, have already guessed, the simplest way to create a beautiful and contrasting (in the coloristic sense) picture is by giving the light and dark areas of the photo different shades. Therefore, we will perform all manipulations by selecting the Whites and Blacks values in the Color parameter.
I will not dwell on this method in more detail - it is too obvious. All the transformations are so visual that you just have to use your imagination. I will give only one option, made on a quick fix, as an example:

As you can see, the presence of four cyan magenta yellow and black sliders provides a very rich scope for flight of fancy. The main thing is not to be afraid to experiment.
Hello, dear readers. You've probably heard many times about the existence color correction in photoshop. This article will describe in detail the principles and techniques of color correction in Photoshop.
This article will describe:
- what is color correction used for?
- What Photoshop tools can you use for color correction?
- examples of color correction
- how not to spoil the image
Why is color correction done?
There are many tutorials on color correction, but they only describe the principles of image processing. Before you start working with photographs, you need to understand what color correction is and what it is used for.
Color correction is Replacing the original color component, hue, tone and saturation of an image. It is used for several reasons.
1. It often happens that we see real colors that turn out completely different in the picture. This happens due to incorrect camera settings, poor quality shooting equipment or specific lighting. Technology does not have the ability to adjust to lighting in the same way as the human eye. This causes errors in the pictures.
2. Color defects. These include overexposure of the image, excessive darkness, haze or dullness of tones and colors.
3. Creative experiments. Color correction allows you to realize the photographer's plans. You can add fabulousness or brutality, expressiveness or dullness to your pictures. it all depends on the photographer’s intention.
Most often, this is done by decomposing the image into channels. There are several modes for editing:
- RGB - Red, Green, Blue. This is the most popular image editing mode. This is what you encounter most often.
- CMYK - Cyan Magenta Yellow blacK.
In this mode, white in the channels shows the maximum presence of color, and black, on the contrary, the minimum. If, for example, the Blue channel is completely filled with black, then there is no blue color in the image at all.
We briefly looked at what is color correction and what it is used for. Now it's time to move on to practice.
It's not for nothing that Photoshop has received the title of the most powerful tool for color correction. If you look at the Image -> Adjustment tab, you can understand that it is all about color correction.
It will take too much time to analyze all the tools, so let’s highlight the main ones: Levels - called by the keyboard shortcut Ctrl+L, Curves - Ctrl+M, Hue/Saturation - Ctrl+U, Selective Color (Selective Colors) and Shadow/Highlights (Shadow/Light).
Examples of color correction
Let's look at three cases of using color correction using specific examples.
Incorrect color rendering
Look at this photo of a tiger:
There is a lot of red in the photo. For correction we will use curves (Image -> Adjustment -> Curves). Select the red channel and slightly lower the curve just below the middle, as shown in the figure:
Here's what we ended up with:
minimal manipulations significantly improved the quality of the image. Curves are a very powerful tool. It is used in many photo editing tutorials.
Let's conclude: if the image has a high content of one color, you need to select the corresponding channel in the curves and reduce its content, as shown in the example.
Color defects
Take a look at this photo:
There is a lot of white haze and the colors are dull, but that can be fixed.
The Levels tool will help you deal with the haze. To call the tool, press Ctrl+L and set the following settings:
Take a closer look at the left side of the histogram. This is the haze. By moving the slider to the right, we increase the amount of dark tones and get rid of haze.
It has gotten a little better, but the result is still far from perfect. You need to lighten the kitten a little. Make a copy of the layer (Ctrl+J) and open Image -> Adjustment -> Shadow/Highlights (Shadow/Light). We make the following settings for shadows:
We made the photo lighter.
Initially, it was planned to lighten only the kitten, so we create a mask for the lightened layer and hide the background with a black brush, leaving only our model light.
Result:
The photo is already good. however, something is wrong. Let's open Image -> Adjustment -> Photo Filter and add a green or warm filter:
If you close your eyes to the initial quality of the image, we can say that we have achieved an excellent result. And here he is:
Color correction is done, but the photo needs to be retouched. This is another topic. Retouching tutorials are not hard to find.
Here is the result with a little retouching:
Creative approach to color correction
The first example will use a great source photo:
Apply a warm photo filter to it (Image -> Adjustment -> Photo Filter):
We create new layer and fill it with color #f7d39e. Change the blending mode to Exclusion (Exception) and set the opacity to 25%
We get rid of the intermediate layer with the fill and look at the result:
The second example will demonstrate the principle of processing a portrait:
Let's take a picture with the fairly popular "whitening" mode. Create a copy of the layer (Ctrl+J), and press Ctrl+U to bring up the Hue/Saturation filter. Set Saturation to 0. The image will become black and white. Change the blending mode to Overlay and the opacity to 70-80%
Let's add an effect to the photo that is often used in modern cinema. Open our photo again to return to the original image. Press Ctrl+M to open the curve editor. Select the blue channel and add blue in dark colors. Go into editing mode for the blue channel and add blue to the dark tones. In light colors we lower the curve. This will give yellowness and will not allow the skin color to be lost.
This is what the photo looks like after correction in Photoshop:
Let's remove the purple tint by adding green to the dark tones. We also perform these actions using curves.
Minor adjustment of skin color:
At this point the processing can be completed.
How not to ruin a photo?
There are no specific rules on this issue. Color correction depends on the personal preferences of the photographer and his sense of proportion, but there are several recommendations:
- When using different tools, it is worth checking all kinds of settings. The result may be unexpected and unpredictable. And that's not always a bad thing.
- Make copies and duplicates of layers constantly. This will allow you to return to any stage of processing if something goes wrong.
- Get inspired by the work of professional photographers. This will allow you to develop your taste and achieve new, interesting results.
This completes the review of color correction examples.
Based on materials from the site:
With this article I want to start a series of notes about simple photo processing techniques. Professional photographers They are unlikely to learn much new from here, but for beginners and amateurs this information may be useful. In each article, I plan to analyze the processing of one or two photographs, touching on correcting the most common and serious mistakes of beginners. The photographs discussed in the articles are either mine or were sent to me for processing and published with the consent of the author. To process photographs, Adobe Photoshop (using the example of CS 3 Ru), Adobe Lightroom (using the example of the 3rd version of Eng) and some others will be used. Although the articles are intended for beginners and hobbyists, it is assumed that readers are familiar with the minimum basics of working with these programs.
Before starting the discussion, let's make a reservation in advance: we are not talking about shooting techniques and how mistakes could be avoided at the photography stage - many articles and books are devoted to this. But everyone who has ever picked up a camera knows that it happens that the most interesting shots have to be thrown away due to technical defects. And sometimes it’s simply not possible to retake such photos. I hope this series of articles will help you save valuable footage captured with technical errors.
So, in this post we’ll talk about color correction of images using the “curves” tool. In general, curves are perhaps one of the most powerful tools in image processing programs. Skillful use of this tool allows you to truly work wonders and do without many plugins with an incomprehensible and uncontrollable mode of operation. On the other hand, this tool is quite complicated, which discourages novice amateur photographers from using it. Let's try to understand the basics of the curves tool and learn how to use it for our tasks - correcting the color and brightness of images.
1. Basics of color theory and color correction
In order to understand color correction, we first need to understand what color is in digital photography, what parameters it is set by, and how to change these parameters to get the desired result. Of course, everyone involved in digital photography knows the letters RGB. This is the main color space with which most cameras and reproducing digital equipment work (there are other color spaces, but this is not about that now). All colors in this space are obtained by “mixing” 3 basic tones (often called channels) - red (Red), green (Green) and blue (Blue). The values are usually written sequentially, red-green-blue, for example 125-80-245. The brightness of each tone is set with a value from 0 to 255. This is a well-known truth.
Another truth, less known, but also important: with equal values of all three channels, we get a neutral gray color. 0-0-0 - absolutely black, 255-255-255 - absolutely white, between them - all shades of gray. How are the rest of the colors obtained, except for the three main ones? Understanding this is also essential for working with flowers. The most obvious representation of color formation in RGB is in the form of a color wheel. On it, in addition to the three primary colors, there are additional (or better yet, opposite) colors: CMY - green-blue (Cyan), purple (Magenta) and yellow (Yellow).
This simple-looking circle provides vital information for informed color correction. With 3 main channels (RGB), we can also control additional colors. Increasing the intensity of any of the main channels (as well as simultaneously weakening the two remaining main channels) leads to an increase in the main color. Decreasing the intensity of each main channel (or simultaneously increasing the two remaining main channels) leads to an increase in the opposite color. That is, by reducing the blue channel, we enhance the yellow colors, by enhancing the green, we reduce the purple shades, by reducing red and green at the same time, we enhance the blue, etc. If we need to, say, make the photo warmer, we will need to add orange and yellow. With yellow it’s clear - we reduce the intensity of the blue channel. And orange = red + yellow. This means we need to add more intensity to the red channel. Gradually, the use of this color wheel becomes automatic, but for now you can draw something like a cheat sheet. So, the theory is more or less clear to us, now let’s move on to practice.
2. “Curves” tool and semi-automatic “three-point” color correction (simple case)
In order not to delve completely into the theoretical jungle, let’s take a photo with a deliberately incorrect white balance: taken under incandescent light in the camera’s “daylight” mode. The result of photo automation is visible to the naked eye. Suppose we didn’t think of shooting in RAW (where the white balance problem is solved with one click), and now we’ll fight with JPEG.
First of all, let's figure out what tools the “curves” procedure offers us, and what settings need to be made before starting color correction. Open “curves” (Ctrl -M), click “curve display options”. I like a more frequent grid, but it's not for everybody. Click the “parameters” button on the right, set the cutoff of black and white colors to 0.10% so that when adjusting the brightness the program does not go into dull overexposure (255-255-255) and dull shadow (0-0-0), check the box “use by default." The rest of the settings can be left untouched. So we see the curve window. The X axis is the original brightness, the Y axis is the brightness after processing, from 0 (full shadow) to 255 (full light). The background shows a histogram that gives us the distribution of points in the photo by brightness. The curve itself runs diagonally - before processing it is a straight line: original = processed. As we work, we can either increase the intensity of certain areas (pull the curve up) or decrease it (pull the curve down). You can work both with the RGB curve (gray curve, adjusts the brightness of all 3 channels simultaneously), and with the curve of each channel separately. Below the curves window there are 3 eyedroppers. They are used to set the points of black, neutral gray and white in the photo. In the simplest cases, it is enough to use these pipettes; the smart program will perform color and brightness correction for you.
Let's try to use this technique: first we need to understand where the neutral points in the photo are those points that were gray in reality. Reflections of objects should not fall on them (for example, in our photo on the edges of the bottle, the gray foil has a different shade due to reflections). In addition, you need to determine the brightest and darkest point. It is important that these 2 points should not be in dead shadow (0-0-0) and dead overexposure (255-255-255), because from such points the program will not receive information for correction. The algorithm of work is as follows: we determine the white point (poke with a white pipette), then the black one (respectively with a black pipette), and then poke with a gray pipette in potentially gray places until the result suits us. The program automatically recalculates the balance of the curves, which can be seen in the corresponding window.
After we are satisfied with the white balance, we can adjust the brightness balance either with the “levels” tool or by correcting the RGB curve (essentially the same levels, but with much greater capabilities). I turned down the brightness of the highlights a little and increased the contrast a little (this is done by the RGB S-curve). That's all. In 50 percent of cases, such a simple color correction completely helps.
3. Color correction using reference points
Unfortunately, everything is not always so simple, and sometimes pipettes are not enough. Yes, and in our relatively simple case correction with eyedroppers does not give an ideal result (for example, I don’t like the reddish tint of the top part of the photo). Let's return to the original photo and try the second option - manual color correction. Here we will move the curves ourselves until the desired result is achieved. But simply moving curves “by eye” requires experience and, it would be nice, a calibrated monitor that will convey the real color of the photo. We will not touch upon calibration; this is the topic of other articles, and photographers working with calibrated monitors usually no longer wonder how to do color correction. We can only rely on the same “really gray” points, i.e. dots that were neutral gray when photographed. Adobe Photoshop allows you to select reference points by which we will monitor the correction. This is done with the “color reference” tool, a variation of the “eyedropper” tool. In the tool settings, you need to select an average sample size of 3x3 or 5x5 px to prevent the pipette from hitting a random pixel, which can also be color noise. Information about color standards is displayed in the “info” panel. We can select up to 4 reference points.
It is better to choose gray points in reality, preferably of different brightness (shadows, midtones, highlights) and from different areas of the photo. In my case, I chose 3 points and all on the bottle, because... There was no longer much choice:
The information about the points is reflected in the panel, and now we can clearly see that our channels are not balanced. Actually, the result of color correction will be the alignment of values at reference points, and at the same time correction of the color of the entire image. Which channel should we take as the basis to which we will bring the other 2 channels? Typically the channel with the most balanced histogram is selected, most often green. Then from the reference points you can see that we need to reduce the intensity of the red channel and enhance the blue one. Well, let's get started. Let's start with red. Firstly, it is clear that we have overexposure in the red channel (the histogram is cut off on the right). It's sad, but not fatal. And this means that we can easily move point 255 (the brightest) to achieve the result. Now we move the curve so that for all reference points we reach the values R = G (approximately ±2 units). One point is not enough for this. Points on the curve can be placed either directly with the mouse or “taken” from a photo (Ctrl + click on the photo area). It is important here that there are no sharp bends in the curve (unless we have several lighting sources of different color temperatures, for example an incandescent lamp and daylight, but this is a much more complex case). We don’t pay attention to the color of the picture - the work is not finished yet.
We do the same with the blue channel. Here the histogram is cut off on the left - the blue channel is underexposed, and does not reach the right edge. This means that you will most likely have to move point 0 (the darkest) and 255. We aligned the reference points, it became better, but not perfect - the top part of the photo is clearly red:
It is necessary to correct a separate area of the red curve responsible for the points of this brightness. Using Ctrl + click we place a point from the problem area on the curve. In order not to spoil the remaining areas, we place a blocking point on the curve and move the main point to the desired result.
After this, all that remains is to adjust the brightness and contrast of the image (using the “levels” command, or correcting the RGB curve), and voila! We did manual color correction.
Of course, you can still tinker with this image, but I think the principle is clear. What's important is that reference point correction is virtually free of problems with monitor calibration - we bring reference points to gray using digital channel values that do not depend on how the monitor is configured.
What to do in more complex cases? Sometimes, no matter how hard you try, you cannot adequately correct a photo using just one curve. And if the scene was illuminated by several light sources with different color temperatures, this is almost impossible. The only thing that will help here is to select individual areas using a lasso with soft edges, and separate correction using the methods described above.
This is how, with some skill, you can “defeat” almost any incorrect white balance. The matter, as you have seen, is not very simple, which once again proves that it is better to shoot in RAW and with correctly set parameters. I hope this article will help you make your photos better. Good luck and great photos!
P.S. In general, this article does not say anything super new; it is rather a generalization of my personal experience for color correction using “curves”. I will be glad to receive criticism and clarifying comments that will help improve the article, and I will also be happy to answer questions.
P.P.S. All rights to this article belong to the author. Any copying of the article or its part without the permission of the author, or without a link to this page, is prohibited.