Quite a lot has already been written on the site, but the important question of how to fix it in Photoshop was still open. True, I casually touched on this topic in the article about, but in it we considered the Raw format, which may not be available on all cameras. Well, now is the time to talk about how to fix white balance in Photoshop for pictures in Jpg. I will show a couple of options, although in fact there are significantly more of them, but all give similar results, so which one to choose is up to you.
For more than five years I have not taken pictures in jpg, in my opinion it is better to spend 5 minutes to turn raw into an ordinary picture in auto mode, but at the same time be able to fix something in the future than to lose the necessary details and then bite my elbows. This is a small lyrical digression about the benefits of raw, so let's get down to fixing the white balance.
Option to correct white balance using a Levels adjustment layer
The photograph that we will be correcting was taken at night (generally speaking, almost all automatic errors occur precisely in the dark) and has a strong yellow-orange tint. Some people like these options for night photos, but I would make it a little colder to better show the dynamics of the night city.
As always, before starting work, we will make a new layer, but this time not a regular one, but an adjustment layer. Do not be intimidated by the unusual name, everything is very simple - just click on the icon for creating adjustment layers 
in the layers panel on the right at the very bottom, and select from the menu that appears " Levels» (« Levels«). 
A new one will appear in the layers panel, with the icon as at the top, and the panel “ Adjustments» (« Correction"). If you have an earlier version of the Photoshop program, then a separate panel “ Levels» (« Levels"), Which is slightly different in appearance, but performs all the same functions.
First of all, we are interested in the histogram at the top of the panel, or rather the gray slider at the bottom. If you move it from side to side, you can darken or lighten the image.
Now we go directly to the balance of colors, the middle eyedropper of the three in our case located on the side is responsible for them. Click on it and select the area on the image, which, in our opinion, should have a gray color. This changes the color balance of the image. You can try to select different areas and see the result. I used the wall of the building on the left side of the image as the gray object. In my case, the result is like this.
I will not say that it is perfect, but in my opinion it looks more interesting than it was. This method allows you not only to correct the white balance, but also to tint the image. Of course, this is more of a side effect, but it allows you to get interesting results in some cases.
Correcting White Balance Using Color Overlay
The second option, in my opinion, allows you to get more reliable results, but leaves less space for creativity.
Now change the blending mode of this layer to " Color«(« Colour") And decrease" Opacity» (« Transparency"). In our case, I stopped at 29%. By changing this parameter, you can more or less neutralize excessive yellowness (like ours) or blue (if the balance is knocked off in the other direction) in the photo.
So the result is:
In my opinion, the picture is closer to its natural state, but in the previous version it looked more interesting.
The method used to correct white balance depends on the specific image. In some cases, when combined lighting and light sources have different color temperatures, any tricks still lead to an unsatisfactory result. In such cases, there is only one way, even if not to correct the balance of colors, but to make the picture more realistic - desaturate the image.
Of course, the picture changes from color loss, but in some cases this is, alas, the only way to get a good picture.
The emergence of photography made it possible to transfer the concept of "memory" from the virtual-mental world to the physical world. From the yellowed old photos, the faces of our great-great-grandfathers and parents look at us. Yes, and ourselves at the age of "going under the table." The advent of color photography and digital photography sparked a revolution again. The simplicity of the process, ease of storage (thousands of images can fit on a standard hard drive) have given new meaning to the pictures. Today it is not only a memory, but also a mood, sensations, an illustration of a worldview.
All this, of course, is good. But in a barrel of honey, as you know, a fly in the ointment is often found. When we talk about color photography, color also acts as tar. But not the one we would like to see.
Examples? Yes, as much as necessary. Hazy colors of a spring landscape, an earthy "dead man's face" instead of a tan, colored reflections and highlights in the most important parts of the image.
And here you cannot do without correction. The creators of Adobe Photoshop looked like they were in the water, saturating the program with various shade correction tools. Starting from changing the color of the eyes and ending with a global shift in the gamut of the entire image.
Today we will consider one of these tools. It is not as intimidating as levels and curves (when working with channels). Moreover, the palette and setting of properties are implemented very clearly and simply. This, of course, is about a tool called "color balance". We will study its properties on a familiar picture of a lake in the outskirts of Minsk.
A bit of theory
Given the specifics of how color balance works, it is worth repeating the theory of color coding. At least remember what RGB and CMYK color spaces are. More about this is written in the lesson "". For those who do not bother the seed with unnecessary reading, we will briefly repeat.
Both models encode color in a three-dimensional coordinate system (remember algebra and geometry in grade 6 high school). The idea is taken from a physics course. Remember that white color, refracting, breaks down into colored components. Here, too, each axis has its own color. Combining base colors in different saturations gives the resulting hue.
RGB space is natural. The three basic colors (red, blue and green) are mixed at full brightness to give white. An example from life - three spotlights on stage.
However, physically, there is no paint that mixed in one bucket would turn white. Therefore, the CMYK system (Cyan, Magenta, Yellow, Black) is used for printing. Cyan, magenta, and yellow combine to produce a rich dark gray. The fourth component, black, serves to "support the shadows".
There is one more interesting feature. On the color wheel, RGB and CMYK are opposite each other. In other words, they form the following contrasting pairs of blue - yellow, blue-red, crimson-green. By reducing, for example, the presence of blue in the gamut, we will inevitably accentuate the red shades. Removing yellow - blue. Getting rid of the raspberry color, we will strengthen the green color.
This is the principle behind the Color Balance tool.
Practical part
Before starting the practical work, it is worth recalling at least two tips from Adobe.
Perform color work on a calibrated monitor. That is, with settings that ensure that the color displayed on the screen looks the same in reality. Otherwise, there is a danger that on another computer you yourself will be horrified by the results of your work.
Try not to change the color of the base layer. Duplicate right away. And color-correct the copy.
Now is the time to get started.
There are two main ways to invoke the color balance tool:
Through the menu " Image» – « correction» – « color balance»
By pressing the keyboard shortcut CTRL + B
The appeared tool palette is extremely simple. On the right are the standard buttons OK, cancel and set the preview (we strongly recommend that you enable it)
The main block consists of three linear change bars with a slider in the middle. They are built on the principle of RGB and CMYK contrast. The color combinations "blue-yellow", "cyan-red" and "raspberry-green" are interconnected. Moving the slider towards one of them reduces the presence of the second.
There are three points worth noticing below. The tool " color balance»You change the color of shadows, midtones and highlights separately. To select a range, a point is set opposite the corresponding inscription.
That is, you can, for example, increase the presence of green in areas of light and at the same time get rid of it “in the shadows”.
In our case, everything is simple. By emphasizing the green, we reduce the amount of magenta, yellow, and possibly reds. For each range in its own way. In the form of an algorithm, it looks like this:
- Make a copy of the layer
- Go to copy and select the tool " color balance»
- Check the box next to the word " preview". This will allow you to see the results of the change in real time.
- Beginning with " shadows»Change the colors of the image on each of the ranges. Remember, good work is slow work. The first correction may fail. Therefore, start with the shadows, go to the zone of light. And then - the second round. If necessary - the third and so on.
- Finish the job at the touch of a button OK
Comment: Pay attention to the check mark next to the words “ Save the glow". This option "protects" light and shadow transitions from loss of contrast. If the glow preservation option is not activated, there is a danger of getting a “faded” image as a result.
An illustration of this is in the figure below.
After you've changed the color, apply the desired blending modes and adjust the opacity of the top layer.
If necessary, use a soft brush to erase the information on the top layer in the areas where color correction is not needed.
Now it is worth paying attention to one feature, which we mentioned only in passing.
Any tool in Adobe Photoshop works with the selected part of the selected layer. If no selection is specified - with the entire layer. But not with the whole image. It is important.
Take a look at the picture below. The light areas of the forest behind the lake, after a general color change in the photo, have acquired a bright yellow tint.
You can create another copy of the background and work on this area separately. The algorithm is as follows:
- Create two copies of the background
- Turn off the visibility of the top (last copy). To do this, just click on the eye-shaped icon to the left of the layer thumbnail
- Go to the visible copy, change its color
- Adjust layer blend and opacity
- Erase the areas on the copies that you think have acquired an unwanted tint.
- Select the transparent (erased areas) on the layer. For example, the tool “ Fast selection" or " Vmagic wand»
- Set the feathering of the selection. More about this is described in the lesson ""
- Without removing the selection, go to the next layer, turn on its visibility and press the button " DEL».
- Invert selection with keyboard shortcut CTRL + I or through the menu " Highlighting» – « Inversion". Thus, already processed zones will be selected on the upper layer (which you do not need to work with)
- By pressing the buttons " DEL”, Erase them. Only the “problem areas” will remain on the layer.
- Change the color of this layer, adjust the blend and opacity.
Comment: Feathering required. Otherwise, jagged sharp edges of the color transition are obtained. This is a sign of inept correction. And such a photo looks like a C grade.
But there is also an easier way.
It is also based on selection, but avoids creating unnecessary layers.
- Create one (!) Copy of the layer.
- Correct the image using any known method.
- After the problem areas appear, select them. The most suitable tool is. This method allows you to identify even the smallest areas in different parts of the photo. In addition, the "undocumented feature" of the tool is the selection of translucent pixels.
- Set the feathering of the selection area. Choose the feathering radius based on the size of the photo. For sizes up to 1000 pixels along any axis, the radius is 2 pixels. more than enough.
- Call the tool " Color balance"(Or any other tool).
- Perform color correction
- Adjust the blending modes and opacity of the top layer.
- As a result, as you can see in the picture below, the bright yellow color in the background is no longer so annoying. Green shades were added, which were required to get the "May Day" effect.
Comment: do not try to completely get rid of extraneous colors. Otherwise, the photo will appear unnatural.
For example, in the case of our sample, the yellow hues remained in the grass in the foreground and in the foliage of the distant forest. Naturally, our sun is yellow. And on a bright day, the lack of glare would look unnatural.
Advice: actively seek out the best options. Don't stop at a single blend mode. Sometimes the most interesting results are produced by seemingly unnatural regimes.
Successful and fruitful work!
With this article, I want to start a series of posts on simple photo manipulation techniques. Professional photographers are unlikely to learn a lot from here, and this information may be useful for beginners and amateurs. In each article, I plan to analyze the processing of one or two photos, touching on the correction of the most common and difficult mistakes of newbies. The photographs analyzed in the articles are either mine or were sent to me for processing and published with the consent of the author. For the processing of photos will be used programs Adobe Photoshop (for example CS 3 Ru), Adobe Lightroom (for example 3rd Eng version) and some others. Although the articles are intended for beginners and hobbyists, it is assumed that the readers are familiar with the minimum basics of using these programs.
Before starting a discussion, let's make a reservation in advance: we do not touch upon the shooting technique and how it was possible to avoid mistakes at the stage of photographing - many articles and books are devoted to this. But everyone who has ever taken a camera in their hands knows that it happens that the most interesting in content frames have to be thrown away due to technical defects. And sometimes it is simply not possible to reshoot such photos. I hope this series of articles will help you save valuable footage captured with technical errors.
So, in this post we will 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 uncontrolled mode of operation. On the other hand, this tool is rather complicated, which discourages 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 - adjusting the color and brightness of images.
1. Fundamentals 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 and how to change these parameters in order to get the desired result. Of course, everyone in digital photography knows the RGB letters. This is the main color space with which most cameras and digital reproduction equipment work (there are other color spaces, but this is not the point now). All colors in this space are obtained by "mixing" 3 base tones (often called channels) - Red, Green and Blue. The values are usually written sequentially, red-green-blue, for example 125-80-245. The brightness of each tone is specified by 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 is absolutely black, 255-255-255 is absolutely white, in between - all shades of gray. How do you get the rest of the colors, except for the three main ones? Understanding this is also critical to working with colors. Most clearly, color formation in RGB can be depicted as a color wheel. On it, in addition to the three primary colors, there are additional (or rather, opposite) colors: CMY - green-cyan (Cyan), magenta (Magenta) and yellow (Yellow).
This simple-looking circle provides essential information for conscious color grading. 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 decreasing the blue channel, we increase the yellow colors, while increasing the green - we decrease the magenta shades, while decreasing the red and green at the same time - we increase the blue, etc. If we need, say, to make the photo warmer, we need to add orange and yellow. With yellow it is clear - we reduce the intensity of the blue channel. And orange = red + yellow. So we need to add more intensity to the red channel. Gradually, the use of this color wheel comes to automatism, 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. Tool "curves" and semi-automatic color correction by "three points" (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 the light of incandescent lamps in the "daylight" camera mode. The result of the work of the photo automatics is visible to the naked eye. Suppose we didn’t think to shoot in RAW (there the problem of white balance is solved with one click), and now we will 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 proceeding with color correction. Open the "curves" (Ctrl-M), click on the "curve display options". I like the more frequent mesh, but this is not for everybody. Press the "parameters" button on the right, set the cutoff of black and white colors at 0.10%, so that when correcting the brightness, the program does not go into a dull overexposure (255-255-255) and a dull shadow (0-0-0), check the box "use by default ". The rest of the settings can be left alone. 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, which gives us the distribution of points in the photo in terms of brightness. The curve itself runs diagonally - before processing, it is a straight line: original = processed. In the process of work, we can both increase the intensity of certain areas (pull the curve up), and decrease (curve down). You can work both with the RGB curve (gray curve, adjusts the brightness simultaneously for all 3 channels), and with the curve of each channel separately. There are 3 pipettes below the curves window. They are used to define black, neutral gray, and white points in a photograph. In the simplest cases, it is enough to use these eyedroppers, 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 in reality gray. Reflections of objects should not fall on them (for example, in our photo, at the edges of the bottle, the gray foil has a different shade due to reflections). In addition, the brightest and darkest point must be determined. It is important that these 2 points should not be in dead shadow (0-0-0) and dead light (255-255-255), because the program will not receive information for correction from such points. The work algorithm is as follows: we define a white point (poke with a white pipette), then a black one (with a black pipette, respectively), and then poke a gray pipette over potentially gray areas until the result suits us. The program will automatically recalculate 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 possibilities). I slightly dimmed the highlights and increased the contrast a little (this is done with an S-shaped RGB curve). That's all. In 50 percent of cases, such a simple color correction completely helps out.
3. Color correction using reference points
Unfortunately, not always everything is so simple, and sometimes pipettes are indispensable. And in our relatively simple case, the correction with pipettes does not give an ideal result (for example, I do not like the reddish tint of the upper part of the photo). Let's go back to the original photo and try the second option - manual color correction. Here we will move the curves ourselves until we achieve the desired result. But a simple movement of curves "by eye" requires experience and, it would be good, a calibrated monitor that will reproduce the real color of the photo. We will not touch on calibration, this is a topic for other articles, and photographers working with calibrated monitors usually no longer wonder how to do color correction. It remains for us to rely on the same "really gray" points, i.e. dots that were neutral gray when captured. Adobe Photoshop allows you to select the reference points by which we will follow the correction. This is done with the "color standard" tool, a kind of the "eyedropper" tool. In the tool settings, select the average sample size of 3x3 or 5x5 px in order to exclude the dropper from hitting a random pixel, which may also be color noise. Information about color patterns is displayed in the "info" panel. We can select up to 4 reference points.
It is better to choose gray in reality points, preferably of different brightness (shadows, midtones, lights) and from different areas of the photo. In my case, I chose 3 points and everything on the bottle, because there was no longer much of a choice:
The information about the points is reflected in the panel, and now we clearly see that our channels are not balanced. Actually, the result of color correction will be the alignment of the values at the reference points, and at the same time the correction of the color of the entire image. What channel should we take as a basis, to which we will lead the other 2 channels? Usually the channel with the most balanced histogram is chosen, most often it is green. Then from the reference points you can see that we need to lower the intensity of the red channel and increase the blue. Well, let's get started. Let's start with red. First, you can see that we have overexposure in the red channel (the histogram is cut off on the right). This is sad, but not fatal. And this means that we can safely 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 "take" them from the photo (Ctrl + click on the photo area). It is important here that there are no sharp bends on the curve (unless we have several light sources of different color temperatures, for example, an incandescent lamp and daylight, but this is already a much more complicated case). We do not 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 clipped on the left - the blue channel is undershot, and falls short of the right edge. This means that most likely you will have to move point 0 (the darkest) and 255. We aligned the reference points, it became better, but not perfect - the upper part of the photo is clearly reddening:
It is necessary to correct a separate area of the red curve, which is responsible for the points of this brightness. Press Ctrl + click to place a point from the problem area on the curve. In order not to spoil the rest of the areas, put a blocking point on the curve, and move the main point to the desired result.
After that, it remains to correct the brightness and contrast of the image (using the "levels" command, or by correcting the RGB curve), and voila! We did some manual color correction.
Of course, you can still tinker with this image, but the principle, I think, is clear. What is important, the correction by reference points is practically free of the problem with the calibration of monitors - we bring the reference points to gray based on the digital values of the channels, which do not depend on how the monitor is configured.
What to do in more complex cases? Sometimes, with all the desire, it is not possible to adequately correct the photo with the help of one curve. And if the scene was lit by several light sources with different color temperatures, this is almost impossible. It will only help to select individual areas using a lasso with soft edges, and separate correction using the above methods.
This is how, with a little knack, you can “defeat” almost any wrong 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 the correct parameters. I hope this article will help you to make your photos better. Good luck and great pictures!
P .S. In general, this article does not say anything supernova, it is rather a generalization of my personal experience in color correction using "curves". I will be glad to 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 part of it without the permission of the author, or without linking to this page, is prohibited.
Even a little color correction can dramatically improve the overall impression of a photograph. In this article I will tell you one of the ways to use the command ColorBalance Color Balance for light color correction of a photo.
For example, I took a studio photo that has already been processed and is ready for transfer to the customer. However, there is one nuance that can be corrected (nothing needs to be corrected here), namely, to carry out a small color correction of the lighting, so to speak, to make the final "smear".
You can carry out color correction with a large number of tools, methods and techniques for a complete description of the process, you need to write a whole book. I will limit myself to one, but effective technique based on the use of a Color Balance adjustment layer ( Color balance).
Using the Color Balance command for color correction
There are two ways to use Color Balance - directly by the command available by pressing the key combination (Ctrl + B), a destructive method of application, justified only when there is a clear understanding of the command settings and does not imply their change in the future.
The second way that gives the possibility of a flight of imagination and full control over the settings at almost any stage of work on a photo is to use an adjustment layer. This is what I use. I will create a corr. the Color Balance layer by clicking the weights icon in the Adjustments panel
The same operation can be done from the item layers of the main menu of the program, or by selecting the corresponding item of the button for creating adjustments at the bottom of the palette of the same name.
By creating a corr. layer, a panel of its parameters will open, equipped with various sliders and gadgets, before pulling them I will say a few words about the principles of operation and "possible consequences".
Determination of tone and color
I'll start from afar: Without going into details, a digital image can be imagined as a set of pixels of a certain brightness and color.
Definition of tone
Brightness values from 0 darkest to 255 lightest for each of the three channels (RGB model). The brightness scale can be substituted in the form of a gradient ruler from black to white from 0 to 765 - the maximum brightness for three channels, the value can be rounded up to 750, there will be no trouble.
The most complete idea of calculating values for color tone is given by the НSВ model (Hue, Saturation, Brightness - hue, saturation, brightness) which you can read about. The line presented here is nothing more than a part of this model (Brightness).
Conventionally, the ruler can be divided into three ranges:
- Shadows Shadows from 0 to 250;
- half tones Midtones from 250 to 500;
- Highlights from 500 to 750 respectively.
Measuring the image with the eyedropper focusing on the Info palette, calling the F8 palette, we get the RGB color values in the place where the cursor is located, summing these values, we get the brightness of the area, mentally applying a ruler we can imagine in what range of tones it is.
For example, an eyedropper placed over the forehead of the model gives R - 237, G - 212, B - 202 for a total of 651, hence this area belongs to the Highlights range. Now, understanding how to find out in what tonal range a particular part of the image is located, you can understand where the correction will have the greatest impact.
There is no point in doing calculations with absolute precision, it is enough to roughly imagine what the sum of the values will turn out when adding. That's about tone, now let's move on to color.
Color definition
Directly to my example, determining the color is not a primary task, but "since such a booze has gone, cut the last cucumber."
Let's go up the ruler by the value calculated for tone range we will go to the color wheel at this level of brightness, for whom it is not clear read the article about the HSB link above.
In the center of the circle, the minimum value color saturation moving towards the edge, it gradually increases until it reaches its maximum 255 (edge of the circle). The calculation starts with the largest value in the example, this is Red 237, I plot this value along the corresponding axis. I put the next value 212 for green parallel to the corresponding axis, as you might guess for the blue 202 parallel to its axis. As a result, I get an unsaturated orange color, which is true.
Back to work: Look at the color axes extending from the center of the circle. Don't they remind you of anything, like the color adjustment sliders in the Color Balance options bar? They work the same way as if we were moving along the axes on the color wheel. It is impossible to add any color without reducing the amount of the opposite (complementary) to it.
Working with Color Balance
I want to "liven up" a boring studio background by adding a little blue-green tint to it, guided by the definition of tone, I can safely assume that the object of interest to me is in the middle tones. I will select the corresponding item in the Ton list: Midtones add blue +100, green +46, and cyan -14. The Preserve Luminosity checkbox must be enabled, this will preserve the brightness of the original pixel and affect only its color.
New color shades were treated not only by the background, but also by other areas of the picture lying with it in the same range. I have nothing against this shade on the model's hair, but on her skin it doesn't look exactly "kosher". You can remove the effect of the correction using masks covering parts of the picture with it.
Create a mask application based on a channel
The mask can be made using the color range command from the Selection menu about the bark I wrote in, or you can make a mask based on the image channels. What will I do now.
Go to the Channels palette, the model's face is the lightest in the red channel, I'll make a copy of it by dragging the channel thumbnail onto the leaf icon at the bottom of the palette.
The blank for the mask looks good anyway, but clarify it by adding contrast. I will use the command (Ctrl + L) for this, moving the shadow slider a little closer to the center.
Now let's create a selection based on this blank by clicking on its thumbnail while holding down the Alt key. Back in layers, I will stand on the mask corr. layer I will use the command to invert Ctrl + I.
Remove selection with Ctrl + D. The mask turned out to be not completely black, which means that the correction will partially go to the model's face.
I'll fix it by adding warm tones to it. I will create a new corr. Layer Color Balance in the highlights (Ton: Highlights) add Red +3 and Yellow -2 in the middle tones of the light (Ton: Midtones) add Magenta - 4.
I will leave the correction only on the model's skin, hiding it from the rest of the areas with a mask: Copy it from the previous layer. Holding down the Ctrl + Alt keys and dragging the thumbnail to this layer.
Invert the mask with Ctrl + I to hide the effect from the whole image except the model's skin.
Color correction for shadows and highlights
At the next stage of work on color correction I want to change the color of the highlights on the model's face a little, I will use this technique for this. Create a new empty layer and use the eyedropper to sample the color from the highlight on the forehead. I will set the color proof size value on the eyedropper parameters panel to 11 by 11 Average (11 by 11 Average)
Fill the layer with the selected color, the Alt + Delete combination immediately invert it to get the opposite of the selected color Ctrl + I change the blending mode for this layer to Color (Chroma).
To hide the effect, instead of a mask, I will use the following approach: Double-click to call the Layer Style dialog in the Blend if block in the parameters of the effect on the underlying Underlying Layer, hold down the Alt key, split the shadow slider and set its halves so as to limit the effect in the range of brightness from 133 to 245 approximately in this range lies the glare of interest to me. It is convenient to visually focus on changes in the picture. The numbers indicate here for "general development".
If necessary, you can lower the opacity of the layer, I set it to 50%. Now I will slightly reduce the effect of color correction on the mid tones and use the same technique. Create a new layer, take a sample from the background, fill the layer with it, invert it, call Layer Style in the Blend if block on the Underlying Layer, split and adjust the color slider, the parameters you see in the picture
I lower the opacity value until I get the desired result. The final image, as it can be replaced, differs from the original one as it seems to me for the better.
