- Tonal correction
- Color correction
- Saturation correction
- Color effects
- Reduce noise
- Image size
The script for the Photoshop Basic Workshop is intended exclusively for use in class at the University of Applied Arts Vienna and should serve as a reference work for all course participants. It may not be passed on to third parties.
Like cropping and straightening, the first step in processing an image is to examine its tonal values. In any case, tonal correction should be done after cropping the image to a reasonable format and also after cleanup, i.e. after an initial detail retouch.
The question of tonal values is essentially aimed at two aspects that characterize every image in a genuine way: brightness and contrast.
Can we be satisfied with the brightness and contrast in our sample image, or does an analysis of the tonal values prompt us to intervene?
We find that especially the contrast of our image, needs improvement. The image seems somehow dull and flat. There seems to be a gray haze over it. The brightness, on the other hand, is within an acceptable range, but it could also use a little adjustment.
Of course, recognizing at first glance whether or what the tonal values are suffering from is a matter of practice. Photographers have a certain starting advantage here: for them, dealing with brightness and contrast is obligatory. All others must first acquire this conscious view of tonal values. In the following, we will get to know a few tools that help us to quickly determine what is going on.
For a targeted intervention, the Brightness/Contrast adjustment dialog of the Image menu seems to be the best choice.
> Image / Adjustments / Brightness/Contrast.
In the easy to use dialog you will find two sliders. Use the Brightness slider to darken or lighten the image. The Contrast slider increases or decreases the contrast in the image.
However, this crude tool cannot be used to perform a truly perfect tonal correction. We therefore ignore this dialog.
The Adjustment menu holds a number of tools for us, which make it possible to influence or change the qualitative image aspects brightness, contrast, hue and saturation. Not all of the adjustment options listed here deliver useful results in daily practice. Some tools are only effective in images with a large color depth, i.e. they are oversized for an image like ours. Other tools do not do much more than apply average or standard values and are therefore out of the question. Besides Brightness/Contrast, Equalize is one of the latter.
> Cancel Brightness/Contrast.
> Image / Adjustments / Apply Equalize.
Equalize ensures an even tonal value distribution. Unfortunately, many subtleties are lost due to the leveling. So the dialog is not an option for us either.
> Undo Equalize.
> Open Image menu.
For the same reason, we ignore Auto Tone, Auto Contrast and Auto Color. We certainly don’t want to settle for average applications.
In order to be able to intervene specifically in the tonal values of an image, a tool is needed that offers more differentiated adjustment options. Since the Levels dialog is characterized by just such options and also provides assistance in analyzing the tonal value situation, we like to use it.
> Image / Adjustments / Levels.
At first glance, the dialog seems a bit complicated. However, this is not the case. The histogram is located in the center of the dialog box. The histogram visualizes the tonal value distribution of the image.
Remember: When we talk about tonal values, we are talking about the brightness values of the pixels. In addition to hue and saturation, each pixel also has a brightness value. And this is what we take care of in the context of a tonal correction.
The total spectrum distinguishes 256 single tone values and thus 256 brightness levels. We can see this from the values of the input fields directly below the histogram.
The values indicate the position of the respective slider. The left slider, currently at position 0, marks the black point of the image. The right slider, currently at position 255, marks the white point of the image. Since the zero is a position, it is counted as such. In this way, we arrive at a total of the 256 different tonal values of the overall spectrum mentioned above.
A tone value column rises above each tone value position. The height of the column indicates the frequency of the tone value and thus how often the tone value occurs in the image.
The histogram thus provides the desired assistance in assessing the tonal values of the image, and indeed of the image as a whole. This is because we are currently assessing the tonal values of the composite channel, as the channel entry RGB reveals.
What does the current histogram tell us specifically?
We can see, for example, that the histogram does not span the entire width of the possible spectrum. We can clearly see that there is a piece missing on the left as well as on the right. The histogram does not extend far enough into the shadows range, and it also falls short in the highlight range. In other words, there is a lack of very dark pixels and a lack of very bright pixels in the image. For the image editor, exactly this circumstance is a typical characteristic for too low contrast.
A simple way to improve contrast in an image is to use what is called linear contrast enhancement. The idea is as follows: if the existing tonal values of the image can be stretched so far that the histogram reaches the limits of the possible spectrum, there should be enough tonal values in the dark and light areas to cause an improvement in contrast.
Contrast is the ratio of the dark areas to the light areas in the image, that is, the ratio of the shadows to the highlights. Increasing contrast means making the dark areas in the image darker and the light areas lighter. In other words, to achieve our goal, we need to increase the distance between the shadows and the highlights.
We do this simply by moving the left slider to the left foot of the histogram and by moving the right slider to the right foot of the histogram.
> Set the black point to position 15, set the white point to position 235.
So we reset the black point and the white point. The improvement is immediately visible in a preview. If we want to see the difference again in all clarity, we temporarily deactivate the preview checkbox.
> Deactivate the preview checkbox and activate it again.
The gray haze has dissolved through the simple action of a linear contrast increase. The image is now much richer in contrast than before.
> Apply tonal correction with OK.
> Open Image / Adjustments / Levels again.
If we open the Levels again after applying the tonal correction, we see what the reason is. The histogram of the existing tonal values has actually been spread over the entire width of the possible spectrum. This is why the process of linear contrast increase is also called tone value spreading. Tone spreading has moved the shadows and the highlights away from each other.
In many cases, this simple measure, which can be applied quickly, produces a really useful result.
Using the Input Levels middle slider, we change the contrast in the midtones. This so-called gamma slider is traditionally set to 1.00 and should usually only be moved very gently.
> Move the Gamma slider and set it back to 1.00.
The grayscale wedge at the bottom of the dialog represents the tonal range. The associated sliders can be used to reduce the tonal range. Since we can never have enough tonal values in the image, such a measure seems absurd at first. Sometimes, however, it is actually necessary to reduce the tonal range. But let’s take it one step at a time.
We could also reduce the tonal range by moving the corresponding sliders of the input levels.
> Input levels, set black point to position 50, white point to position 200.
However, this would cause the tonal values to the left of the black point and to the right of the white point to be irretrievably lost. The tonal range would not merely be reduced, but actually clipped. I think it is clear to see that such tonal clipping would not do the image any good overall. The contrast is now much too high. The image loses subtleties and almost seems graphic.
If, on the other hand, we want to preserve the available tonal values in their entirety and only compress their distribution a little, we use the output levels sliders.
> Input levels, set black point to position 0, white point to position 255.
> Output levels, set black point to position 15, white point to position 235 and OK.
> Open Image / Adjustments / Levels again.
Of course I didn’t choose the values 15 and 235 by chance. What we have achieved with the reduction of the tonal range while retaining all available tonal values is nothing more than the reversal of the process of the previous spreading: The histogram looks again almost as it did before we started working with the levels. A small gap on the left and on the right, you may remember.
While manipulating the Input Levels spreads the histogram, manipulating the Output Levels compresses it.
Admittedly, the last step was not very practical. But my point was to demonstrate how the input levels and the output levels work and how they are related to each other.
In fact, you only reduce the tonal range for very specific purposes. A typical occasion would be to remove pure black or pure white pixels from the image. Pure black and pure white are two conditions that can cause problems in printing. You’ll hear more about this in my Photoshop Advanced workshop.
Also, the eyedroppers found in the dialog on the right are rarely used, and when they are, they are used very specifically.
- Pixels that are clicked with the black point eyedropper are set to black. Typical use case: Clicking on the lines of a photographed or scanned drawing sets the lines to true black.
- Pixels clicked with the white point eyedropper are set to white. Typical use case: clicking on the paper background of a photographed or scanned drawing sets the background to true white.
- Pixels clicked with the gray point eyedropper are set to the famous medium gray. Typical use case: a click on the co-photographed gray card takes an exact white balance.
We have not dealt with the Levels in such detail because they are the most popular tool for tonal correction – they are not – but because the Levels can be used to explain the basics of tone value correction so vividly.
Let’s reset the image to its initial state.
> Cancel dialog.
> History palette, click on the first snapshot.
Now that we have learned the basics, let’s reach for a real HiEnd tool.
Tonal correction via Curves
> Image / Adjustments / Curves.
Curves are a versatile tool for editing tonal values. And the professional image editor will appreciate the flexibility and precision of the curves in equal measure.
Curves can be used to accomplish essentially the same tasks as Levels. However, their special mode of operation gives Curves an invaluable advantage over Levels. We’ll take a closer look at this in a moment. Before that, let’s take a look at the setting of the dialog.
In the center of the control panel you will once again find the histogram. Using the Black Point slider and the White Point slider, we could perform a simple linear contrast increase as usual.
> Set the Black Point slider to 15 and the White Point slider to 235.
> Deactivate the preview checkbox and activate it again.
> Reset both sliders again.
Even the eyedroppers are at hand in the dialog.
However, there is nothing to be seen of a curve. Across the panel, from the lower left to the upper right, however, there is a diagonal straight line. Since we have not yet made any manipulation – the input values and the output values are identical – the curve just appears in the form of a straight line.
The operation of the curve is now very simple: it is a matter of bending the curve so that the desired change in the tones of the image comes about.
To curve the curve, points are set. But where on the curve should the points be placed? And in which direction do the points have to be moved to achieve the correct curvature of the curve?
Note that in addition to the grayscale wedge at the base of the histogram, the control panel also has a vertical grayscale wedge. The two grayscale wedges indicate where on the curve the different tonal ranges are located:
- In the lower left corner are the blacks.
- In the left, lower curve area are the shadows.
- In the middle curve area are the midtones.
- In the right, upper curve area are the highlights.
- In the upper right corner are the whites.
Let’s go through a few typical application scenarios.
If I want to make the image brighter overall, but keep the contrast unchanged, I raise the curve in the midtone area. To do this, I set a point in the middle of the diagonal and move it to the upper left.
> Set midtone point and move it a little to the upper left.
If I want to make the image darker overall, but keep the contrast unchanged, I lower the curve in the midtone area. To do this, I place a point in the middle of the diagonal and move it down to the right.
> Move the midtone point a little to the bottom right.
If I am not satisfied with the processing, I can undo it by removing the point from the curve again. To do this, simply drag the point beyond the boundaries of the control panel.
> Delete Point.
And how can I use the curve to increase the contrast of the image while keeping the overall brightness unchanged?
Remember the description of contrast as the ratio of the shadows and highlights of an image. If I want to increase the contrast, I need to make the dark areas in the image darker and the light areas lighter. So I have to move the shadows and the highlights away from each other.
And that’s exactly what we can do by manipulating the shadows and highlights of the curve.
We set two points to increase the contrast: A shadow point with which we lower the shadows and a highlight point with which we raise the highlights.
> Set the shadow point and lower it a little.
> Set the highlight point and raise it a little.
Note that the overall brightness of the image has changed only slightly. The brightness is just distributed differently now, if you can put it that way.
Brightness and Contrast
If I want to change the brightness and contrast of the image in one pass, I start by manipulating the overall brightness and then let a contrast increase follow.
To do this, let’s first restore the initial state. To reset the settings we have made, we do not have to delete each point individually. The dialogs in Photoshop offer a hidden function for resetting: If you press the Option key, the Cancel button turns into a Reset button. And clicking Reset resets all settings made up to that point.
> Press Option key and click Reset button.
> Set midtone point and move it a little to the upper left.
> Set shadow point and lower it a little.
> Set highlight point and raise it a little.
- A curve for changing the overall brightness shows a concave or convex shape.
- A curve for changing the contrast has the shape of a slightly curved S-curve.
With that, we have already learned the basic concepts of how curves work and how to use them.
Levels vs. Curves
Let’s now briefly consider what advantages working with Curves has over using Levels.
Obviously, both dialogs have the same goals: to edit the tonal values of an image. However, only a few sliders are available for tone correction with the Levels. So the Levels are more of a crude tool, good enough to do some quick simple manipulation.
The Curves are completely different. I can set almost any number of points in the really convenient dialog and use them to precisely model even the smallest tonal areas. Usually, you can get by with just a few points. But even then, the advantages of the Curves become apparent. In addition to precision, the consistency of the curve, i.e. the smoothness of the curve, also makes an invaluable contribution to perfect tonal distribution.
With the Curves, the image editor has a real HiEnd tool in her hands that can be used in many different scenarios. We will explore this further later.
Let’s now apply the settings we have made to the image.
> Close Curves with OK.
The tonal correction has been made and applied directly to the pixels of the background layer. But we had decided to avoid exactly that. The background layer should always remain in the original view. How can I do a tonal correction without affecting the background layer?
An obvious solution would be to create a duplicate of the background layer and apply the correction to the duplicate. But Photoshop has a much better way of doing this, and that is Adjustment Layers. Let’s take a step back and take a closer look.
> Undo by pressing cmd+Z.
> Layers palette: Open Adjustment menu.
If we click on the button with the contrast symbol at the bottom of the Layers palette, a menu opens that should look familiar to us. It contains almost the same entries as the Adjustment menu in the Image menu.
Here you can find Brightness/Contrast, Levels and Curves, as well as most of the other adjustment features. Here you will find various adjustment options for correcting the tones, the colors and the saturation, as well as for correcting the exposure and for colorization. All essential tools for handling the color values and the tonal values of an image are available.
> Create Curves adjustment layer.
The adjustment layer hovers over the background layer. So the correction we are about to make is not applied directly to the pixels of the original, but is completely separate from any content.
Our manipulations to increase contrast are done in the Properties palette that was expanded when we created the adjustment layer.
> Set shadow point and lower it a little.
> Set highlight point and raise it a little.
The Properties palette of the Curves adjustment layer has the same setting options as the Curves dialog. The result is 100% the same as the Curves dialog.
The fact that the correction is a separate layer brings a number of advantages.
- The correction changes the appearance of the underlying layers, but not their contents.
- The status of the correction remains accessible forever. I.e. the correction made with an adjustment layer can be revised again and again without losses.
- If an adjustment layer that is no longer needed is deleted, the original view of the image is restored. Nothing has been destroyed or lost by the manipulations.
If the adjustment layer is hidden, the image will show up again in its original view.
> Hide and unhide the adjustment layer in the Layers palette.
You can also compare the state before the correction with the state after the correction using the Toggle layer visibility and View previous state buttons at the bottom of the Properties palette.
> Disable and enable the Toggle layer visibility button in the Properties palette.
> Press and release the Before/After button in the Properties palette.
> Close the Properties palette.
If a setting needs to be revised once it has been made, the properties palette of the adjustment layer is opened by double-clicking on its thumbnail. And the correction of the adjustment can be done.
> Double-click adjustment layer thumbnail.
> Increase contrast further.
If you want to undo the last revision of the settings, press the Reset button of the palette.
> Properties palette: Press the reset button once.
If you want to cancel all manipulations, press the reset button once again.
> Properties palette: Press the reset button once again.
After all the advantages of the adjustment layers mentioned here compared to the mere adjustment commands of the menu of the same name, one thing is obvious: The image editor will always reach for the adjustment layers for all color and tonal corrections if possible and largely ignore the adjustment dialogs.
As I have already indicated, the curves can be used to modify not only the tonal values, but also other aspects of an image.
We’ll look at how to use the curves to correct color in a moment. But before we do that, we need to look at some basics of color theory that will guide us.
First, let’s analyze the chromaticity of our sample image. Can we detect a color cast in the image, and if so, what is it?
We could say that the colors of the image tilt a bit to blue. If we look even closer, we notice that it is not so much a blue color cast, but rather a cyan color cast. In this case, the professional image editor talks about the color impression being “dirty”. I think we can understand quite well what is meant by the term when we look at the facade, especially the yellow areas on the first floor. But also the green of the plants makes an unpleasant impression.
> Hide and unhide the Curves adjustment layer.
If we temporarily hide the Curves adjustment layer, we can see that our tonal correction has even slightly increased the cyan color cast. Please pay attention to the gray facade areas. Modifying the tones in the RGB composite channel obviously does not leave the colors untouched. Sometimes color casts only become apparent in all clarity after a tonal correction. It is therefore advantageous to always perform a tone correction before a color correction.
If we want to know which measures we have to take to remove a color cast, we have to know how the colors are distributed on the color wheel.
The spectral colors of white, visible light are distributed on the color wheel. We have already learned about this in another way in the form of the schematic illustration of the Lab color space. With regard to color correction, we are interested in the relationship of the primary colors on the color wheel.
Let’s walk through the color wheel, starting with red, counterclockwise. Red is followed by yellow, yellow is followed by green, and green is followed by cyan. Cyan turns into blue, and blue finally turns into magenta, which turns back into red as we go further. The primary colors are thus distributed in a very specific order on the color wheel.
The immediate neighborhood of colors on the color wheel means that the colors are related in some way. For example, red can be transformed into yellow or magenta in just a few steps, green into yellow or cyan in just a few steps, and so on.
It is striking that a color of the first group of primary colors is always followed by a color of the second group. This results in a scheme that is significant for color correction. RGB colors and CMY colors not only alternate on the circle line, they also face each other on the color wheel.
- Red and cyan are opposite each other.
- Green and magenta face each other.
- Blue and yellow are opposite each other.
RGB colors and CMY colors form pairs of opposites on the color wheel. And such opposite colors are simply called complementary colors.
In addition to the color adjacency and the color opposition, the color wheel also shows another important aspect: The aspect of color neutrality, which forms the center of the wheel and which we have already learned about as the neutral white point.
The cyan color cast of our image can now be understood as the white point of the image being shifted a bit towards cyan. Zones that should appear color-neutral are thus given a certain cyan touch.
To create color neutrality, we have to move the white point back to the center of the color wheel. And we can do that by adding red to the image. So removing a color cast is always done by strengthening its complementary color.
- If cyan is to be reduced, red must be strengthened and vice versa.
- If magenta is to be reduced, green must be strengthened and vice versa.
- If yellow is to be reduced, blue must be strengthened and vice versa.
In order to make color corrections in this scheme, we only need to know the names of the complementary color pairs:
Let’s now apply the knowledge we have acquired to our image.
> Open exercise file 10
Color correction via Curves
While the tonal correction is carried out in the composite channel, the color correction is obviously carried out in the color channels.
> Properties palette: Open the Channels menu.
Since we want to remove a cyan cast, we need to make a correction in the red channel, because cyan and red are complementary colors.
> Bring up the Red channel.
The operation of the curve for a color channel is no different from that of the composite channel. If we want to remove cyan in the red channel, we raise the curve to strengthen red.
> Set midtone point and move it a little to the upper left.
If we want to reduce the amount of red, we do it the other way around: We lower the curve to strengthen cyan.
> Move the midtone point a little to the lower right.
The basic principle of color correction can be applied quickly and easily using the curves.
You’ve noticed, of course, that it’s better to be gentle with color correction. I remove the point from the curve and try to set the new point, which I will now create, more precisely right from the start.
To do this, I use a small, handy tool from the Properties palette and activate the so-called On-image adjustment tool.
> Activate On-image adjustment Tool.
If the On-image adjustment tool button is highlighted, I do things a little differently when setting a point on the curve. I click a spot in the image where the cyan cast is particularly evident and hold down the mouse button.
> Click the facade area between the windows, keep the mouse button pressed.
A point appears in the curve, exactly in the tonal range where the correction should be strongest.
If I move the cursor slowly upwards in the image, cyan is carefully reduced. I only add so much red to the image that the cyan cast disappears, never more, because then I would cause a red cast.
> Move the cursor to an acceptable position last.
Far away tonal areas will be less affected by the correction. The curve shape takes care of this by itself.
By the way: You should be especially careful when removing cyan. In four-color printing, cyan acts as the carrier of the drawing of the image, i.e. the cyan separation is very important for the tonal value distribution and the texture. Images that contain too little cyan make too colorful an impression. And we want to avoid that at all costs.
Once the view is satisfactory, I finally release the mouse button and the correction is done.
Having removed the main color cast, we examine the image for possible other color casts. As for our image, we can let it go at this point.
Let’s collapse the palette and see the result in a before/after comparison.
> Collapse Properties palette.
> Hide the Curves layer and show it again.
I think we can be satisfied with the result so far. After all, we’ve turned an image of mediocre tonal and color quality into one with good tonal distribution and color balance.
Photoshop has a number of other color correction options available to us. One of them is called Color Balance and is also applicable in the form of an adjustment layer.
We hide the Curves layer and create a Color Balance layer.
> Hide Curves layer, create Color Balance layer.
The properties palette of the Color Balance essentially contains the explained scheme of complementary color pairs and thus represents a conveniently applicable variant for editing a color cast.
Cyan and red, magenta and green, yellow and blue are each linked by a slider. If I move the slider towards one color, I remove the other color at the same time.
Color Balance affects the general color mix in the image. Fine balancing, as with the curves, is not possible here. Nevertheless, we can apply color correction with Color Balance separately, at least in the three major tone value ranges.
> Open the Tone menu.
> Tone: Midtones, Cyan-Red: +15.
Compared to Curves, Color Balance represents the somewhat cruder option. However, Color Balance can be used quickly and intuitively, and in many cases it is sufficient.
One more hint: The inconspicuous checkbox Preserve Luminosity plays an important role in the application of the settings made with Color Balance.
> Check the Preserve Luminosity checkbox to deactivate it and then activate it again.
Preserve Luminosity ensures that the brightness values of the image are preserved in the course of color correction. And thus Color Balance also proves to be an ideal option for color correction following a tonal correction made with the Levels.
We have now learned two excellent ways to edit the tonal values and the color values of an image:
- The fast variant: Levels followed by Color Balance.
- The precise variant: tonal correction and color correction with the Curves alone.
Both ways can also be combined. You start with the fast variant (Levels and Color Balance) and do the final tuning with the Curves.
I would like to introduce you to another feature for color correction.
> Hide Color Balance layer, create Selective Color layer.
> Colors menus Expand and select Reds.
The structure and functionality of Selective Color tells us that the tool is primarily for matching the colors of images that will be printed. Therefore, I use Selective Color not only when I want to selectively intervene in color ranges, but precisely to fine-tune the colors in images that have already been converted to CMYK.
Since the color mixing in Selective Color is done by increasing or decreasing the proportions of the four printing colors, the tool is especially appreciated by classic graphic designers.
This approach to color mixing may take some getting used to. However, once you get the hang of it, everything runs quite naturally.
In order to be able to operate the sliders specifically, you have to know how red is produced in the printing press: In four-color printing, red tones have a very high magenta percentage and a very high yellow percentage. Remember the famous 100-100 red, which consists of 0% cyan, 100% magenta, 100% yellow and 0% black. A more powerful 4c red is unthinkable.
> Properties palette: Colors: Reds, Cyan -30, Magenta +50, Yellow +50, Black 0.
So if we increase the proportions of magenta and yellow and decrease the proportion of cyan at the same time, the red tones in the image will become more intense, clearly seen in the roof tiles.
All other color tones remain completely unaffected by my intervention. And that’s really handy, because Selective Color sometimes saves you from having to make a color selection in advance of editing.
Now we improve the yellow tones of the facade on the first floor.
> Properties palette: Colors: Yellows, Cyan -30, Magenta +15, Yellow +100, Black 0.
To strengthen yellow, we set the Yellow slider to 100% and add a little warmth by slightly increasing the amount of magenta. Since we also want the yellow to shine a little more purely, we finally reduce the cyan proportion.
With this, we have become acquainted with the most important tools for tone and color correction. Of course there is still a lot to learn and in the Advanced Workshop I will introduce you to more fantastic features in this regard. But I can promise you, with the skills you’ve learned so far, you’ll be able to get by in many cases.
We’ve taken care of tonal values, we’ve taken care of color values. And now we want to deal with the third aspect that determines the appearance of a pixel: With saturation.
To edit the saturation, we create a Vibrance layer.
> Hide Selective Color layer, create Vibrance layer.
Its Properties palette offers two sliders that can be used to modify the saturation values in the image in different ways.
The Saturation slider increases or decreases the saturation absolutely, i.e. without taking into account the differences in saturation present in the image. So with the saturation slider I set the general saturation level in the image. I make sure that the higher saturated zones in the image are not overdriven.
> Operate the saturation slider and finally set it back to 0.
When increasing the saturation levels in the image, one should proceed rather cautiously. Excessive colorfulness can be quite nice when creating a look, but most of the time it looks artificial and sometimes causes disappointment in reproduction. Many highly saturated colors simply don’t translate to paper in four-color printing. Especially highly saturated blues and greens, but also mixed colors like orange are sometimes difficult to reproduce.
Vibrance gently increases the saturation values in the image by adjusting the existing saturation values as you increase them. While the saturation slider is used to define the desired saturation maximum in the image, the vibrance slider is used to adjust the less saturated zones in the image.
> Vibrance: +40.
The result is more natural than a general increase in saturation. This is why the image editor often uses the more intelligent Vibrance slider right from the start and foregoes a general increase in saturation via Saturation.
So far we have tried to correct shortcomings. But with this homework done, the editing of tonal values, colors and saturation is in many cases not over yet.
A not insignificant part of image processing deals with artificially creating a color mood, in other words: creating a look.
We’ve all been playfully involved in creating a look at one time or another. Think of the possibilities that relevant apps like Instagram and others offer for this. Some of us just love to change the mood of the natural shot.
I can assure you, Photoshop is the mother of all looks. Especially for the advertising industry, photo editors like to use relevant tools from the Photoshop bag of tricks. In the basics workshop, for understandable reasons, we will only briefly look at the ways to create a look.
One of these possibilities is called Photo Filter.
We hide the Vibrance layer. We fade in the Curves layer again to create a good starting situation for the application of a Photo Filter.
> Hide the Vibrance layer, show the Curves layer.
> Create Photo Filter layer.
The procedure for creating an effect is very simple.
> Open the Filter menu of the Properties palette.
We select one of the presets from the Filter menu and then tune the result using the rest of the palette’s settings.
All the filter effects listed here can also be created manually. But we don’t always have to create everything from scratch. Trying out one or the other photo filter can’t hurt. And often a little twist is all it takes to turn a standard filter application into something exciting.
The first six entries of the Photo Filter menu have names that refer to real existing photo filters. We find here three classic warm filters and just as many classic cold filters that the photographer in the pre-Photoshop era screwed onto the camera’s lens to achieve a very specific photo effect.
> Filter menu: Select Cooling Filter (80).
Nowadays, the photographer usually leaves the creation of photo effects to the image editor. The possibilities in post-production are, of course, many times better and more extensive.
The menu also contains numerous pure color filters and, in addition, a few more special effect filters, such as Underwater or Sepia.
> Filter menu: Select Underwater.
> Filter menu: Select Sepia.
Especially the sepia toning is very popular, because the effect gives the image a certain patina, which is alienating, but at the same time pretends a certain degree of authenticity.
Hardly anything needs to be said about the palette’s setting options. They are virtually self-explanatory.
> Open Color Picker and assign a different color.
The Color Picker allows you to assign any color for toning and with the Density slider you determine how strong the color will be applied.
> Increase Density drastically.
And we meet here again the small Preserve Luminosity checkbox, which ensures that the previously perfectly adjusted tonal values do not suffer from the Photo Filter effect.
> Preserve Luminosity checkbox on.
Black and white conversions are also still in demand. And I think even more than ever since the invention of color photography.
When we’re after black-and-white effects, simply converting to grayscale mode is not an option for us. Grayscale mode doesn’t provide much more than a normal conversion and is really only important in terms of print output.
If we want to open the door to the universe of black-and-white photography, we keep the RGB mode for the time being and reach for Black&White.
We hide the Photo Filter layer and create a Black&White layer.
> Hide Photo Filter layer, create Black&White layer.
If you have ever worked with black-and-white photography, you might know the fascination of this medium. Due to the absence of colors, all concentration is on the tones. This has nothing to do with a simple grayscale conversion. The photographer who has devoted herself to this medium creates tonal compositions.
The history of photography is deeply connected with technological progress. In the heyday of black-and-white photography, each new process brought with it new ways of distributing and treating tonal values.
> Open the Preset menu.
A few of these time-typical tonal value distributions have saved themselves to Photoshop in the Preset menu. And we also find a few filters that play a significant role in analog black-and-white photography.
The filter application gives the shot a completely different character. In this case, the image clearly gains in expressiveness. How did this come about?
Of course, the sliders of the palette play the leading role in the black-and-white conversion. With their help, the six primary colors can be specifically converted into tonal values.
> Manipulate the Reds slider.
If I drag the slider for red, the original red tones of the image change, in our case the red of the roof tiles. If I drag the slider to the left, the corresponding tones darken; if I drag to the right, the tones lighten.
And so, for any given base color, an independent tone value conversion can be made.
With the checkbox Tint we bring color in the style of the sepia effect into play.
> Activate the Tint checkbox.
> Open the Tint Color Picker and select a light olive tone.
The way the tonal values are given a color flavor here, as it were, is strongly reminiscent of photos worked out in the historical sepia process of the 1920s and 1930s.
The coloring of the tonal values of an image need not be limited to a single color. With Gradient Map, even complex color gradients can be applied to the image.
> Hide Black&White layer, create Gradient Map layer.
The Properties palette of this popular feature contains not much more than the already known Gradient Editor or Gradient Picker.
> Expand Gradient Picker.
Clicking on the small triangle to the right of the Preview Strip opens the Gradient Picker. And here we can choose as usual from the standard gradients or our own created gradients.
> Gradient Picker: Select Blue, Red, Yellow gradient.
We check the Dither checkbox to smooth the gradient. And of course the Reverse checkbox allows you to reverse the gradient direction.
> Activate the Dither checkbox.
> Activate Reverse checkbox.
> Deactivate both checkboxes again.
Even the few possibilities we were able to explore in the Basics Workshop give you an idea of how wide the field is for creating color effects in Photoshop. And I am convinced that you will enjoy exploring the possibilities on your own when the opportunity arises.
> Open exercise file 1.
Taking color handling one step further, let’s turn to colorizing. One color is replaced by another, parts of the image are recolored.
Since we have already talked about colorizing in the workshop, we already know what it is basically about: The classic technique adds a color to the tonal values of a so-called black-and-white image, which is applied in a certain intensity.
The Color Blending mode of the Fill dialog allows us to do just that: The hue and saturation of the foreground color are applied to the brightness of the image pixels. With the help of a Fill Layer we can also achieve the whole thing without having to take the detour via the Fill dialog.
If I want to recolor the red bucket, I must first select it. Since we have saved selections, we can fall back on one of them and save ourselves the selection process.
> Switch to the channels.
> Activate the selection Bucket_O via cmd-click.
> Switch back to the Layers palette.
Photoshop offers with the Fill Layers a simple and very comfortable way to bring color into the image.
> Layer menu / expand New Fill Layer menu.
Fill Layers are layers that are filled by themselves with color, with a gradient or with a pattern. Solid Color Layers are suitable for coloring. We will not select such a fill layer in the Layer menu, but take a look at the Adjustment Layer menu of the Layers palette.
> Open the Adjustment Layer menu.
Fill layers can also be created via the Adjustment Layer menu.
> Adjustment Layer menu: Select Solid Color.
The special thumbnail indicates that the new layer is a fill layer. As expected, the bucket selection has been converted to a layer mask. And the bucket already has a color fill. The open Color Picker prompts me to define my desired color.
> Choose a strong green and OK.
What has to be done now is clear: We have to activate the right blending mode to achieve a colorization. Because currently the color is applied opaque and without regard to the drawing of the bucket.
To do this, we change the Blending Mode of the Fill Layer itself.
> Open the Blending Mode menu and select Color.
The Blending Mode of our choice can only be Color, of course. With Color we achieve exactly what we wanted to achieve: The hue and saturation of the Fill Layer are applied to the tonal values of the bucket.
A completely different approach, but with the same effect, is offered by Hue/Saturation.
> Hide Fill Layer.
> Adjustment Layer menu: Select Hue/Saturation.
Classic coloring can be simulated via Hue/Saturation by activating the Colorize checkbox in the Properties palette.
> Properties palette: Activate the checkbox Colorize.
Now the hue and saturation of the adjustment layer are applied to the tones of the image. Of course, we still need to define our target color. Also, we wanted to recolor only the bucket and not the entire image. We should have activated a selection to create a layer mask for this, as we did before.
However, we can also use the layer mask already present in the Color Fill Layer by simply dragging it onto the new adjustment layer while holding down the Option key.
> Color Fill Layer layer mask with pressed Option key on the layer mask thumbnail of the Hue/Saturation Layer and press OK for Replace.
Layer masks can be easily copied to other layers in this simple way.
A new color can now be assigned by dragging the Hue slider.
> Hue slider and finally select a blue.
Once we have decided on a hue, we still tune its intensity by manipulating the Saturation slider.
> Increase saturation with the Saturation slider.
Finally, if necessary, I carefully adjust the brightness of the result.
> Minimally increase brightness with the Lightness slider.
However, since I usually have the tonal values of the image perfectly adjusted before coloring, I usually refrain from changing the brightness further here. The Lightness slider of the Hue/Saturation dialog is, by the way, a very coarse tool with which you can quickly overshoot the target.
For simple cases like this, the few colorization steps we’ve done so far are usually quite sufficient. That it can be quite tricky to achieve the goal when coloring is only hinted at here. In the Advanced Workshop we will learn many other approaches to solving the task.
I just want to quickly introduce you to another approach to coloring with the Hue/Saturation adjustment layer. This time we want to colorize a different part of the image, namely the orange structure of the boat.
In many cases, the image editor wants to use a brush for coloring, i.e. color should be applied manually.
> Hide Hue/Saturation layer and create a new Hue/Saturation layer.
> Adjust hue and saturation quickly.
Once the hue and saturation are set, we close the layer mask by pouring black foreground color into it via Option+Backspace. Remember the mask principle.
> Set foreground color to black via D and X.
> Fill the layer mask via option+backspace.
Then we reach for the Brush Tool. Since the outline of the boat is hard, compact and smooth, it is best suited to our clipping tip.
> Activate the Brush Tool and assign a clipping tip.
Now quickly swap the foreground and background colors by pressing the X key. To open the mask we have to apply white. And here we go.
> Set the foreground color to white via X.
> Coloring the hood of the boat.
Now we want to deal with a completely different topic: transforming. Let’s open a new image.
> Open exercise file 11b
> Zoom out a little via cmd-minus.
Transform is an umbrella term for operations that change the pixel position in a specific way. Transforming includes actions like scaling, rotating, tilting, distorting and perspective distorting.
> Open the Edit menu.
The Transform commands are all located in the Edit menu, but are only accessible there if a selection is active in the image or a part of the image is present as a layer.
If we want to transform the whole image, we first create a clone of the background layer.
> Duplicate background layer.
> Hide the original background layer.
> Edit / Expand Transform menu.
The Transform menu is now accessible and has the named operations as entries: Scale, Rotate, Skew, Distort and Perspective.
Let’s start with Scale.
> Edit / Transform / Scale.
Scaling is used to change the size of an image part, i.e. to enlarge or reduce it. The process does not need to be explained. However, let’s take a quick look at the parameters.
Selecting the menu item activates the Transform mode. Transform mode can be recognized by the fact that the object to be transformed now has a frame with eight handles.
> Drag on a corner handle.
We notice: Photoshop scales by itself maintaining the aspect ratio. If I want to change the aspect ratio of the object while transforming, I have to press the Shift key during the process.
> Free scale while holding down the Shift key and give OK.
Clicking the Confirm button on the Options bar applies the transform. Pressing the Enter key or the Return key does the same.
> Finish scaling with OK.
> Undo via cmd+Z.
> Edit / Transform / Scale.
Let’s have a first look at the options bar of the transform mode.
The small button with the link symbol between the two numeric input fields ensures that the scaling is maintaining the aspect ratio. This applies both to manual scaling and to scaling by entering target values.
Since this circumstance causes confusion again and again, I recommend that you never disable the link. By the way, you can hold this as you like. Just know this: Photoshop saves the last state of the link button and keeps it as default from the next program start.
Maintaining the aspect ratio or free, whatever scaling behavior you set as the default for your workflow, the alternate scaling behavior is always available by pressing the Shift key.
Now let’s look at another important aspect of transforming that you should always keep in mind.
Let’s scale the clone towards the center in a first step. To do this, we hold down the Option key during the process. We finish the transform process at last with an OK.
> Scale the clone strongly to the center while holding down the Option key and confirm with OK.
In a second step we want to bring the object back to its original size. We do the transformation again while holding down the Option key and give OK.
> Edit / Transform / Scale.
> With the Option key held down, return the clone to its original size and confirm with OK.
Clearly visible: Transforming, especially if it is done multiple times and in an extreme way, sometimes changes the object very much. Everything seems somehow blurred, the contours have become spongy, artifacts disfigure the image.
If we compare the result with the original state, the full extent of the information loss is revealed. Transforming had a devastating effect on the image quality.
> Layers palette: Hide and unhide background copy.
If we keep in mind what happens to the pixels during the transform process, we should not be surprised by this result. Photoshop has to recalculate the pixels during each transform. When shrinking, pixels that are no longer needed are thrown overboard, and when enlarging, missing pixels have to be calculated based on the existing data. Photoshop performs a so-called interpolation for this purpose, i.e. it uses a complex algorithm for the recalculation.
Each new transformation step starts from the current state of the image. The material for the calculation therefore becomes progressively worse the more often it is transformed.
Photoshop offers a clever solution to this basically unavoidable problem: Smart Objects.
We create a new duplicate of the original background layer and position it in the Layers palette as the top layer. We now turn this new clone into a so-called Smart Object. We achieve this either by selecting the corresponding entry of the palette menu …
> Open the Layers palette menu / Convert to Smart Object.
… or via right-click on the layer by selecting the corresponding entry of the context menu.
> Right-click Background Copy 2 and select the Convert to Smart Object entry.
At first sight nothing has changed. Only the small Smart Object icon of the layer thumbnail tells us that we have successfully created a Smart Object. The advantages of a smart object only become apparent when it is transformed.
We scale down the Smart Object strongly towards the center and finish the transform process with an OK.
> Edit / Transform / Scale.
> Scale the smart object strongly towards the center while holding down the Option key and confirm with OK.
And now let’s scale the Smart Object back to its original size.
> Edit / Transform / Scale.
> Bring the smart object back to its original size while holding down the Option key and confirm with OK.
We can’t see any loss of quality in the transform result. All negative effects have disappeared. The Smart Object seems to stop the information loss during transforming.
Let’s hide the destroyed clone and compare the multiply transformed Smart Object with the original background layer.
> Layers palette: Hide Background copy.
> Layers palette: Hide Background copy 2 and show it again.
Obviously no losses at all. You can’t ask for more than that.
What is behind it?
The Smart Object preserves the state of the pixels of a layer. Each subsequent transform operation thus always reverts to the original state. The basic material therefore always remains the same and does not deteriorate from time to time, as we have to expect with conventional transformations.
So the Smart Object allows a non-destructive transforming. It works so well that we hardly ever do a transform in Photoshop without first turning the pixels into a Smart Object.
We learned the basics of transforming in the discussion of scaling. The command can be found with a few more options to transform images or parts of images in the Transform menu.
> Edit / Expand Transform menu.
In practice, however, I hardly ever use the commands gathered here, but choose a more convenient procedure with Free Transform, which can also be called up via a very catchy key shortcut, namely Cmd+T.
> Edit / Free Transform.
The Free Transform mode gives access to all transform commands, except for the flipping operations. As usual, the Free Transform mode provides us with eight handles for manipulation.
I can scale.
> With the Option key held down, scale towards the center.
If I move the cursor a little away from the object, it turns into a bent arrow. The image is now ready to be rotated around its center.
> Rotate object around the center point.
By moving the reference point, a new rotation center can be determined.
> Move the center point and rotate.
I can also set the reference point in the options bar.
> Options bar: Set reference point back to center and rotate.
The operations of distorting, tilting and perspective distorting can be achieved in the Free Transform mode by pressing simple key commands.
To distort, hold down the Command key. All handles can now be moved completely freely.
> Distort freely while holding down the Cmd key.
Tilting, i.e. distorting in parallelogram style, is caused by holding down both the Command and Option keys.
> Tilt while holding down the Cmd and Option keys.
Perspective distortion, i.e. distortion in trapezoidal style, is caused by holding down the three Command, Option and Shift keys.
> Distort in perspective while holding down the Cmd, Option and Shift keys.
If you forget one or the other key command, you can quickly find the right way by simply trying it out. Simply play through the key combinations with Command, Shift and Option until you achieve the desired result.
Since the parameters of the Options bar can also be accessed in Free Transform mode, there is no reason to call the Transform commands individually in the Transform menu of the Edit menu.
Let’s cancel the process by pressing the Esc key and look at a typical use case for transforming, namely perspective fitting.
> Cancel transform process via esc.
We switch to the background layer, select everything via Cmd+A and copy the whole image to the clipboard via Cmd+C.
> Activate background layer, cmd+A, cmd+C.
Then we open a new document, namely Exercise file 11a.
> Exercise file 11a.
We paste the clipboard content into the target document via Cmd+V. The task is to fit the pasted sign into the existing panel in perspective.
What we do not omit, of course, is to convert the subject to be distorted into a Smart Object before transforming it.
> Layer 1 Convert to Smart Object via right click.
The perspective fitting is really easy in this case. The format of the sign and the format of the panel are proportionally approximately the same. So we can use the Distort option.
For better visual control, we temporarily reduce the opacity of the Smart Object to 50% and we’re ready to go.
> Smart Object, Opacity 50%.
In a first step, we place the vertices of the sign approximately in the vertices of the white area of the panel.
> Place the vertices of the sign in the vertices of the whiteboard.
To fine-tune the alignment, we zoom in on the individual corner points in a second step and make sure that the white edge of our motif comes to lie outside the white area of the panel. This is because the white border is to be removed after the transformation.
> Zoom in on the four corner points and position them exactly.
> Fit on Screen via Cmd+0.
> Apply transformation via OK.
> Set Smart Object, Opactity back to 100%.
We hide the Smart Object layer temporarily and activate the background layer. Using the Magic Wand tool we select the white area of the panel and use the selection to create a layer mask.
> Hide Layer 1, activate Background.
> Activate Magic Wand Tool, Tolerance 50, select white area.
> Show and activate Layer 1 again.
> Create layer mask by pressing the button Add layer mask.
The sign is now exactly transformed, positioned and cut out. To finally give the fit a more realistic look, we use the Layer Style features.
> Activate Smart Object layer.
> Open Layer Style dialog by double-clicking the layer bar.
To make the shield’s clipping lose its hard, artificial character, we create an outline to conceal the problem.
> Activate Stroke and Reset to Default.
> Size 3px, Position Inside, Blend Mode Normal, Opacity 100%, Color Black.
Since the metal frame is three-dimensional, we provide our sign with an appropriate drop shadow.
> Activate Inner Shadow and Reset to Default.
> Blend Mode Multiply, Color Black, Opacity 50%, Angle 135°, Distance 10px, Size 10px, Contour Linear.
And finally we apply a simple light effect in the form of a reflection on the sign area.
> Activate Gradient Overlay and Reset to Default.
> Blend Mode Overlay, Opacity 60%, Gradient Black&White, Style Reflected, Reverse, Angle -10%, Scale 120%.
> Apply Layer Style via OK.
If we now fade the effects out and back in again, it becomes clear how much even simple measures like this can increase the realistic feel.
Fitting in perspective is not always as easy to accomplish as in our example. But we’re just getting to know the basics. In my Photoshop Advanced Workshop I will present much more advanced options to make perspective adjustments, even to change the image perspective itself and much more.
However, we still want to look at a more complex type of transforming in the basics workshop. Let’s take a look at exercise file 11b again.
> Switch to exercise file 11b.
Let’s deselect it via Cmd+D. Let’s bring the Smart Object layer back into view and activate it as well. Hide all other layers.
> Unselect via cmd+D, show and activate the Smart Object layer.
> Hide all other layers.
The Warp command can be found like all other Transform commands in the Transform menu.
> Edit / Transform: Warp and finally close the menu again.
But we can also call the Warp mode via the options bar of the Free Transform mode.
> Free Transform via cmd+T.
> Reduce the sign a little bit to the center while holding down the Option key.
> Press Switch between free transform and warp modes button.
In the warp menu of the options bar we can choose from different types of warping.
> Options bar: Open the warp menu.
Let’s start with the custom type. In its simplest form, the warp mode initially offers us just four control points. As you can see, these are equipped with handles.
> Pull one of the vertical handles slightly inwards.
Dragging a handle causes a local transformation of the pixel mass in real-time preview.
To have more options available for local editing, we can activate a grid.
> Options bar: Grid menu, select 3 x 3.
More control points we can move, more handles we can pull and control grid lines we can move.
> Move one of the control points.
> Pull another handle.
> Move a control grid lines.
The warp mode allows to freely deform the pixel mass manually. This is necessary, for example, when a two-dimensional motif is to be mapped onto a quasi-three-dimensional object.
The Warp mode is also used for local rectification or for manually compressing and stretching image zones.
Working in the Warp mode of the custom type is intuitive, manual and by no means precisely controllable. We’ll take a brief look at a dialog later that allows far more sophisticated intervention in the pixel mass than would be possible in Warp mode.
Now let’s take a look at the standard options of the warp mode.
We reset the warp by clicking on the corresponding button of the options bar and then select a default type from the warp menu.
> Options bar: Reset Warp button click.
> Options bar: Warp menu, select various warp types.
The default types of the warp menu provide normalized deformations of the pixel mass. The image part can be made into an arc shape, oscillated, blown up or squeezed and many more.
Let’s familiarize ourselves with the parameters by selecting the simple arc shape.
> Options bar: Warp menu, select Arc.
We reduce the bend to 25%.
> Options bar: Bend 25.0%.
With a negative sign we reverse the bend.
> Options bar: Bend -25.0%.
We make a horizontal deformation by modifying the value of the second input field and a vertical deformation by modifying the third value.
> Options bar: H 20.0% and V -20.0%.
If we are satisfied with the preview, we finish the warp process by pressing the Enter key, otherwise we exit the warp mode by pressing the Esc key.
> Press Esc key.
Let’s open exercise file 12 and take a look at another important Photoshop feature that allows you to manipulate pixel position.
> Exercise file 12.
We will now learn about Lens Corrections, a method that allows us to correct the typical distortions that result from the choice of certain lenses when taking photographs. Although lens correction is not always desirable or necessary, we need to know how to deal with such distortions and other imperfections.
Again, since we don’t want to edit the original pixels, we’ll create a copy of the background layer and convert it to a Smart Object.
> Duplicate Background Layer, Create Smart Object.
Our image exhibits the typical distortions of a wide-angle shot. Since the straight lines of the subject appear curved, the impression of a slightly convex curvature is created. The impression does not correspond to our natural perspective perception. When we want to edit the distortions of the wide-angle lens, the aim is usually to bring the image close to the normal view, in other words, to straighten the lines and soften the convex impression.
Classic lens corrections are made in the Lens Correction filter.
> Open Filter / Lens Correction.
Let’s first get an overview of the dialog’s options.
In the center is a preview of our image. A fine grid helps us to judge the distortions. We can show or hide the grid and modify it by using the appropriate controls at the bottom of the dialog.
> Enable Grid (if necessary), Size 64.
In the right dialog area we find a panel that currently shows the Auto Correction options.
The practical Auto Correction uses the EXIF data that was embedded in the image when it was taken. Since the EXIF data usually also contains information about the lens used, Photoshop applies corrections in Auto Correction mode based on this information.
Unfortunately, our image does not provide any information about the lens used. At least the camera model and a few camera settings can be recognized: NIKON E5400, focal length 5.8 mm, maximum aperture ƒ/2.8. We can read that on the left below the preview field.
In some cases we can use Auto Correction despite the lack of EXIF data. In the Search Criteria area, Photoshop makes suggestions to complete the missing data. Here we can choose a camera manufacturer as well as a suitable camera model. And we can even apply the parameters of certain lenses.
> Auto Correction-Panel: Open and close Camera Maker menu.
> Auto Correction-Panel: Open and close the Camera Model menu.
> Auto Correction-Panel: Open and close the Lens Model menu.
If no reasonable replacement can be found, we reach for the tools and make the corrections manually.
At the top left we have access to a few relevant tools.
Remove Distortion Tool
Let’s try to remove the distortions with the Remove Distortion Tool.
> Activate the Remove Distortion Tool.
> Click in the center of the image and drag to the right and then drag to the left.
If we click in the center of the image and drag to the right, the so-called barrel distortion is reduced. If we drag to the left, the opposite pincushion distortion disappears.
Since the preview is recalculated in Life mode, we sometimes have to be a little patient. As long as the blue progress bar indicates that the calculation is still running, we have to hold back with the evaluation.
Our approach should be cautious in any case, so as not to go from the frying pan into the fire.
> Perfect the rectification with the Remove Distortion Tool.
We will use the Remove Distortion Tool to remove the distortion only until we have straightened the lines to a certain extent. The grid provides us with the necessary clues for this.
What we find after rectifying our image, however, is that the image as a whole needs to be straightened. And of course we do this task with the Straighten Tool.
> Activate Straighten Tool and apply it to the top edge of the D to L keys.
II apply the reference line to a horizontal line in the middle area of the image during straightening. This way I avoid as much as possible to amplify a problem in the upper part of the image while removing it in the lower part and vice versa.
Let’s look at the preliminary result.
> Disable preview checkbox and enable it again.
All in all, we could be satisfied with the rectification done. But the dialog offers much more than these, admittedly quick to apply, but still quite simple options.
> Bring Custom panel to view.
The Custom panel offers a whole range of genuine options for lens correction. We will not go through all options, many of them are self-explanatory. However, we will take a closer look at a few aspects.
First, we notice that two values already have a modification, which doesn’t surprise us, since these are exactly the aspects that we have already edited manually: The geometric distortion and the angle.
Let’s now improve the result with a small perspective transformation. As you can easily see, the horizontal lines in the upper and lower parts of the image converge in a trapezoidal shape. With a minimal increase of the value of the horizontal perspective, this deficiency can be quickly corrected.
> Custom panel: Horizontal Perspective +2.
A short word about vignetting: A typical problem in some wide-angle shots is the so-called vignetting, i.e. the shading of the image section in the corners. By increasing the corresponding value in the Lens Corrections, this phenomenon can be counteracted in a simple way.
> Custom panel: Vignette Amount +10.
Chromatic aberrations represent a completely different insufficiency. Since we will deal with this in more detail in the Advanced Workshop, here is just a brief word about it.
Chromatic aberrations occur in the form of color fringes that sometimes appear at certain contours in the image. The phenomenon owes itself to the curvature of the lens’ glass body and is difficult to avoid.
Fortunately, we can quickly and easily fix this unwanted effect in Lens Corrections.
Let’s go in search of these color fringes and zoom in strongly on the D key of our subject.
> Zoom in on the D key.
The vertical contours show clear color fringes in the green/magenta area. They can be easily removed with the slider of the same name.
> Custom panel: Chromatic Aberration: Green/Magenta Fringe -70.00.
We successively reduce the value of the Green/Magenta Fringe until the phenomenon disappears.
We return to the Fit on Screen view and complete the Lens Correction by pressing the Enter key.
> Fit on Screen via cmd+0 and close dialog with OK.
A look at the Layers palette tells us that the Lens Correction filter has been applied to the Smart Object in the form of a Smart Filter.
Smart Objects force Smart Filters. And with that, we’ve learned about another top feature. Smart Filters make it possible to readjust the filter settings over and over again without any loss of quality.
If we hide and unhide the Smart Filters or, more specifically, the Lens Correction, we can examine the result in before/after style.
> Layers palette: Hide and unhide Smart Filters.
By double-clicking on the Lens Correction entry, we return to the Lens Correction dialog and can readjust the settings made there. A Smart Filter application remains freely editable.
> Double-click the Lens Correction entry.
> Bring Custom panel to view and cancel last.
So the Smart Filters live up to their name. Incidentally, all filters can be applied as Smart Filters to Smart Objects. We’ll come back to this when we get into a few more filters during the course.
Behind the seemingly harmless term Liquify lies a powerful tool for local manipulation, or more precisely, for deforming parts of an image. Why we count Liquify among the techniques of transforming becomes immediately clear when we call up the dialog.
But first, let’s open a new image.
> Open exercise file 13.
> Create Duplicate Background Layer, Smart Object.
> Open Filter / Liquify.
The extensive Liquify dialog offers a large collection of tools and features, all aimed at one thing: moving pixel mass locally.
Let’s reach for the Forward Warp Tool of the toolbox to illustrate this with a first operation.
> Activate Forward Warp Tool.
As with the classic Brush Tool, we first need to take care of the Options. We set the size of the brush tip and the pressure via crtl-option-drag.
> Size 300, Pressure 75 via crtl-option-drag.
The pressure determines the strength with which an effect is applied. The lower the pressure, the more precise the work.
> Brush Tool Options: Density 60.
The density value determines the strength of the effect within the brush tip. A low value reduces the effect in the peripheral area of the brush tip. To get smooth transitions when moving the pixels, you usually choose a value between 50 and 75. You could compare the density with the hardness of a brush tip.
> Edit the nose and mouth with the Forward Warp Tool.
Using the Forward Warp Tool, pixels can be moved locally.
This small manipulation makes it clear what the Liquify dialog is all about. We are making corrections here in the sense of modeling. An inconspicuous intervention changes the character of the image without a viewer being able to recognize the manipulation as such. In other words, we intervene in the reality of the image in order to change it according to our wishes.
It is important to understand that in doing so we do not primarily achieve a qualitative improvement, but an improvement in content. And that’s what makes this kind of intervention problematic in certain contexts.
Before we think about this a bit, let’s try a few other tools of the dialogue. Let’s activate the Bloat Tool and enlarge the – from our point of view – left eye of the subject a bit.
> Activate Bloat Tool, Size 350, Pressure 75, Density 60, Rate 20.
> Enlarge left eye by two precise clicks.
You can achieve the modification by single tool clicks or use the build-up effect by keeping the mouse button pressed. In this case, the rate value comes into play.
The higher the value for the repetition rate, the faster the effect builds up.
> Further enlarge the left eye with the build-up effect.
With the Pucker tool we cause the opposite effect. The pixels are pushed together towards a center.
> Activate Pucker Tool, Size 350, Pressure 75, Density 60, Rate 20.
> Reduce left eye back to original size.
The Twirl Tool causes a twisting movement in the pixel mass.
> Activate Twirl Tool, Size 350, Pressure 75, Density 60, Rate 20.
> Rotate the right eye.
If you hold down the Option key during the process, the rotation is counterclockwise.
> Rotate the left eye with the Twirl Tool while holding down the Option key.
Let’s now check how strong our intervention has turned out. We already know the procedure: Deactivating and reactivating the preview checkbox at the bottom of the Properties panel provides a before/after comparison.
> Deactivate the preview checkbox and activate it again.
If you have done too much of a good thing, you can use the Brush Reconstruct Options to soften the overall modeling. To do this, we open the Reconstruct dialog and drag the slider to the left until we are satisfied with the result.
> Brush Reconstruct Options: Open the Reconstruct dialog.
> Set the slider several times to 0 and then back to 100; finally set to 50 and OK.
The Restore All button of the Brush Reconstruct Options cancels all modifications caused by brush tools.
> Click on Restore All.
You already guessed it: The Liquify dialog gives you powerful tools for deforming the human body.
There is one special tool set we want to deal with briefly.
> Open Face-Aware Liquify.
The sliders in the Eyes section can be used to edit various aspects of the eyes. The left column contains the sliders for manipulating the appearance of the left eye, and the right column correspondingly contains those for manipulating the right eye.
> Perform manipulations on the eyes.
With the Eye Distance slider we can even modify the distance between the eyes.
> Move Eye Distance slider to the left and to the right.
We can manipulate the length and width of the nose.
> Perform manipulations on the nose.
We can put a smile on the subject’s face and modify various aspects of the mouth.
> Mouth: Drag the Smile slider to the right.
> Modify other mouth aspects.
And finally, the Face-Aware feature even allows you to change the shape of the face to a certain degree.
> Face Shape: Move the Jawline slider (chin area) and Face Width slider to the left.
If used skillfully, the modifications are drastic but unobtrusive. Let’s check the preview checkbox again for comparison.
> Deactivate the preview checkbox and activate it again.
> Close the Liquify dialog with OK.
It is already clear that there is much more to the Liquify dialog than has been explored so far. However, due to time constraints, a discussion of other important tools and features cannot take place until the Photoshop Advanced Workshop, which shouldn’t stop you from spending a rainy afternoon manipulating the faces of your friends, relatives, and acquaintances before then.
Now is the time to think a little about image manipulation in general.
In this respect, one topic is right in front of our eyes: the modeling of the human body.
With the Liquify tools, it’s easy for the image editor to optimize the human body. Here, eyes are enlarged, noses are made smaller, lips are injected, breasts are enlarged, butts are brought into shape, body lines are idealized, in short: aesthetic interventions of any kind are performed. You could actually call Liquify a tool box for digital cosmetic surgery.
The modeling of the human body is indeed controversial. However, a sustained public discourse leading to clear agreements or regulations is only being conducted selectively and, in my opinion, too timidly in many aspects.
What is at stake here is nothing less than the image we have or would like to have of ourselves. Because the even faces and perfect bodies function as projection surfaces for our desires.
Certainly, an alert mind questions this aesthetic dispositive where it is instrumentalized by advertising, for example. But the small, inconspicuous optimizations still do their work subliminally. We are exposed to the pressure to optimize and can only resist it with an enormous effort of will. We have internalized the aesthetic compulsion too much.
We cannot discuss here how the aesthetic compulsion arises in the form of an ideal of beauty. But I would like to briefly present to you the circle of responsibility in image editing:
- The image editor claims to follow the art director’s instructions in cleaning human bodies of their alleged blemishes.
- The art director refers to the client of the advertising campaign, who prefers a certain idealized image.
- The client thinks that with this idealized image he has to meet the wishes of a potential customer.
- And the potential customer takes her cue from the body that has been freed of all flaws in the image processing and tries to adapt her own appearance to it.
What are the consequences of this?
The viewer admires the allegedly flawless face and unconsciously considers it to be real, even though he or she is aware of the idealization. On a subliminal level, then, a confusion takes place. And this confusion is not a lapse, but has been deliberately worked towards: The image editor knows exactly how far to turn the screw in order to implement the aesthetic constraint without causing cognitive dissonance.
It is the moderate, unobtrusive modeling that can sometimes do the very most damage to our conception of the human body.
The idealization of the body is something deeply human. Every era, every community pays homage to its own specific ideal of beauty. I think that idealization can become a problem precisely when it demands irrevocable, physical deformations. The emphasis here is on the little word can. For no one wants to deny an adult human being, who is in full possession of his mental powers, sovereignty over his body.
Since difficult, also biopolitical and psychopolitical problems emerge on the horizon that has now become visible, which we cannot go into here, I would like to conclude by only pointing out that every ideal is relative, even if it makes the absolute claim to be achieved.
The dialogs Lens Correction and Liquify are counted among the filters. We will now deal with a few more filters and open a new screen for them.
> Open exercise file 14.
> Open the Filter menu.
Filters are basically program routines with which certain image processing tasks can be solved more or less comfortably.
The Filter menu contains various filter sets, each of which contains the filters of a relevant topic.
Blur filters can be found in the Blur Set and in the Blur Gallery.
> Open Filter / Blur Set and Blur Gallery.
Sharpen filters can be found in the Sharpen set, stylize filters in the Stylize set and noise filters in the Noise set.
> Expand Filters / Sharpen Set, Stylize Set and Noise Set.
It is, of course, impossible to explore all filter applications within the scope of the workshop. But we have to deal with a few important filters that are relevant in daily work. So we will pick out a few filters that we want to get to know better. The rest of the filters will be dealt with by yourselves on a rainy afternoon.
I make a rough distinction between filters that are used for classic image processing tasks and those that are used to create effects.
The Filter Gallery offers a whole range of such effect filters. And although we have to refrain from dealing with them here, I would like to show you how the Gallery works by means of a concrete example.
> Filter / Open Filter Gallery.
The moment we open the Filter Gallery, a large dialog appears that is divided into four zones:
- In the left zone we have a generous preview panel.
- On the right, a selection menu containing numerous filter sets.
- On the top right, the same filter menu, but this time in a list view, as well as the setting options of the active filter.
- And below that, something like an effect layer palette that shows all filters that are currently applied to the preview.
Selecting a filter is done by activating a filter button directly in the filter sets or via the entries in the filter list.
> Activate various filters in the sets.
> Expand and collapse the filter list.
I grab the cutout filter from the Artistic set completely at random.
> Open the Artistic set and activate the cutout filter.
The filter application can be controlled using three sliders.
> Cutout filters: Number of Levels 4, Edge Simplicity 6, Edge Fidelity (degree of precision of implementation) 2.
Working with such effect filters is done according to the trial-and-error principle. The specific setting options are mostly self-explanatory.
It does not have to remain with the application of a single filter of the Gallery on the picture. We can add more filter applications to the current one by clicking the New effect layer button at the bottom of the palette.
> Click the New effect layer button.
> Expand Brush Strokes set and activate Crosshatch filter.
> Crosshatch filter: Stroke Length 50, Sharpness 20, Strenght 1.
It is worthwhile to play around with the Filter Gallery in search of new effect combinations. You will find that useful results can always be found in the middle between too simple and too complex.
We’ll break off the process and now turn to the filters for everyday use.
Let’s start with the blur filters.
> Cancel Filter Gallery.
> Open Filter / Blur Menu.
We will pick out a few important blur filters from the Blur menu and take a closer look.
> Filter / Blur / Gaussian Blur: 20 Pixels.
The Gaussian blur is our standard blur filter when diffuse blurring is needed. Its application causes a uniform dissolution of the contours. The image becomes blurred.
> Move the radius slider to the left and to the right.
The simple dialog offers only a radius slider with which we determine the strength of the blurring effect. The preview field is used to view the blur at different zoom levels.
The goal of blurring is of course not to create a blurred image. We use the Gaussian blur primarily to smooth out the texture, i.e. to remove noise.
Another processing goal could be, for example, to create depth of field blur in order to enhance the spatial impression in the image or to focus on the objects that remain in the area of sharpness. For this task, however, Photoshop provides a few more specialized blur filters that give better results.
The Gaussian Blur filter, of course, is the equivalent of the Blur Tool, which we’ve already learned about.
- We reach for the Blur Tool when we want the blur to be local and manual.
- We use the Blur Filter when we want to blur evenly over a wide area.
> Cancel Gaussian Blur.
Other blur filters do not provide a diffuse dispersion of the contours, but a dispersion in certain directions, e.g. the Motion Blur filter.
Since we want to apply filters as Smart Filters if possible, in order to protect the original and maintain free editability, we duplicate the background layer as usual in a first step and create a Smart Object.
> Duplicate Background Layer.
> Convert to Smart Object via right click.
Let’s first create a selection of the car to create a realistic scenario in the example. The selection doesn’t have to be too precise. To do this, we’ll grab the Brush Tool, give it a soft, relatively large tool tip, and activate Quick Mask Mode by pressing the Q key.
> Activate Brush Tool, tool tip size 300, hardness 30.
> Activate Quick Mask Mode via the Q key and select the car.
> Return to normal mode via Q key.
> Invert selection via cmd-shift+I.
If we now apply the Motion Blur filter, the main subject is spared and only the background is blurred in a typical way.
> Filter / Blur / Open Motion Blur.
> Modify Distance and Angle.
The Distance slider is of course our throttle, i.e. the Distance value determines how much blur is applied. And by setting the Angle, we define the direction in which the blurring occurs.
> Motion Blur: Angle 10, Distance 50 and apply with OK.
So the Motion Blur filter can be used to simulate motion. In fact, it doesn’t matter whether it blurs the subject or the background – psychologically, both work. We’ll take a look at that in a moment.
You may have noticed that the selection has been converted to a layer mask when the filter is applied. So we can edit the Smart Filter mask just as we would edit a normal layer mask.
If we want to apply the effect to the subject rather than the background, we activate the Smart Filter mask by clicking on the thumbnail. Now we select the entry Invert in the Adjustments of the Image menu and invert the tone values of the mask. Immediately the effect is applied to the main subject, the background is spared from the motion blur.
> Activate Smart Filter Mask and invert via cmd+I and invert again.
Since the Invert command is used frequently, we should memorize its shortcut. The key shortcut for Invert is cmd+I.
Now we want to work on one more detail that will further increase the realistic feel. We want to make the front wheel rotate. The radial blur will be our tool of choice for this.
Since the radial blur is to be applied to a narrowly defined area, we separate the front wheel by selecting it in the usual way and turning it into its own Smart Object layer.
> Activate Background Layer.
> Activate Brush Tool, Tool Tip Size 150, Hardness 30.
> Activate Quick Mask Mode via Q key and Brush Front Selection.
> Return to normal mode via Q key.
> Create clone via cmd+J and place layer in the palette on top.
> Create Smart Object via right click.
Once the preparatory work is done, we select the Radial Blur filter. The filter unfortunately doesn’t offer a preview, so we have to guess a bit when it comes to setting the strength of the application.
In our case, a moderate rotation will do. As for the output quality, we don’t compromise and choose the best option.
> Filter / Blur / Radial Blur: Amount 5, Blur Method Spin, Quality Best and apply via OK.
This example should show us that filter applications can also be used for creative purposes. With Motion Blur and Radial Blur we not only simulated motion, but also brought a certain dynamic into the image.
As far as blurring is concerned, the filters of the Blur Gallery, which we will deal with in the Advanced Workshop, are particularly suitable for such and similar tasks.
The smoothing of textures, noise reduction, the softening of surfaces in general define the other large task area in which blur filters are used. Once again, we will approach this important topic with the help of a concrete example.
Let’s open a new image and create a Smart Object Layer of the background layer.
> Open exercise file 15
> Duplicate background layer.
> Smart Object via right click.
Since we always have to deal with surfaces that have specific textures in photos, we could choose any image to trace the effects of smoothing. However, the example of human skin is particularly well suited for this purpose, because this sensitive motif shows, as hardly any other, the boundaries that separate a successful intervention from an unsuccessful one.
> Filter / Blur / Gaussian Blur: 3 Pixels.
Even minimal blurring of the skin will cause a general blurring of the image. But we want to avoid the latter at all costs. So we have to proceed in a more differentiated way. Direct application of the Gaussian blur does not yield satisfactory results here.
> Cancel Gaussian Blur.
Much better results can be achieved with the Smart Blur filter.
> Open Filter / Blur / Smart Blur.
Smart Blur enables selective blurring. Our goal is to smooth the skin areas, but at the same time preserve the sharpness of the image.
In order to better understand the process, and thus the filter settings, we must first understand what image sharpness means, or rather where sharpness is located in the image: image sharpness is created by contrast in the edges. To put it simply: wherever clusters of lighter pixels meet clusters of darker pixels in the image, the impression of sharpness is created.
We can see this wonderfully when we look at the eyelashes and their immediate surroundings, or how the contour of the dark iris contrasts with the white of the right eyeball. Where the contour is softened, as in the area of the left eye, the impression of sharpness is lost.
If we manage to exclude the large areas of high contrast from smoothing and capture only those areas whose low contrast we want to reduce further anyway to get smoothing, we will have achieved our goal.
In other words: We want to blur the areal zones in the image, but at the same time protect the edges.
For exactly this purpose we set the entry Normal as application mode.
> Smart Blur: Mode Normal and Quality High.
As usual, we do not compromise on quality. And now we go on to set the parameters of the blur.
- We already know the Radius slider, with which we determine the strength of the effect. Strictly speaking, we use it to determine how far the pixels should be dispersed locally.
- And with the Threshold slider we separate the zones that should be smoothed from those that should remain untouched. If we set the threshold value to 0.1, only the contrast of the smallest edges will be reduced, if we set the value to 100, large edges will also be blurred.
> Smart Filter: Radius 10; Threshold 0.1 and then 100.
Threshold and Radius thus play together.
I select a section of the image in the preview box that contains zones I want to blur, but also zones that should remain sharp.
I usually start the adjustment by entering a relatively high Radius value, in this case about 5.
> Smart Filter: Radius 5.0.
The effect should be overly visible at first. This facilitates the subsequent definition of the threshold value. And also this threshold is initially set to a large value, I choose 25.
> Smart Filter: Threshold 25.0.
My goal is to soften the pore structure a bit without completely destroying the natural texture of the skin. A careful approach is therefore called for.
By successively decreasing the threshold, I now approach that point where the zones that should remain sharp appear sharp and the zones that should be smoothed show the blur.
I do this in steps of 10 using the Down arrow while holding down the Shift key.
> Reduce Threshold via Shift+Down arrow to 5.0.
And now I also reduce the Radius value until the desired smoothing is achieved. Zooming in the preview to 200% is helpful for the inspection.
> Zoom the preview to 200%.
> Reduce radius via down arrow to 2.0.
In the absence of a preview checkbox, I temporarily click directly in the preview field to make a before/after comparison.
> Click on the preview box and release it.
As you can see, the differentiated, selective blurring only takes effect where I want it to. The hair and the eye area hardly lose any sharpness.
> Apply Smart Blur with OK.
However, blurring is only one of the ways of editing textures, i.e. smoothing them. Photoshop also offers us completely different options in the form of noise filters.
We start again by creating a Smart Object from the copy of the background layer.
> Create a duplicate background layer and position it as the top layer.
> Create Smart Object via right click.
> Open the Filter / Noise menu.
Noise is the term used to describe contrast in the smallest edges of an image, which can be of natural or artificial origin. Smallest structures like the pores of the skin, larger structures like freckles, dust and scratches, but also artifacts like the digital color noise in the shadows or the classical film grain are counted to it.
Noise reduction is not about softening areas or zones, but about removing such microstructures in a targeted and careful manner. This is why certain noise filters are sometimes even better suited for smoothing textures than the blur filters.
Let’s take a look at the larger structures first.
Dust & Scratches
Dust and scratches, or even freckles, can be removed quite well with the noise filter for the coarse, i.e. with Dust & Scratches.
> Open Filter / Noise / Dust & Scratches.
The dialog has two sliders whose functions are already familiar to us.
- With the Radius slider we determine how much processing is done.
- With the Threshold slider we determine what size the structures must have that they are processed.
Since I first want to find out how much I have to intervene to make the structures I’m targeting disappear, I set the Threshold slider to 0 and successively increase the radius until the freckles dissipate.
> Threshold 0.
> Radius to 1 and then raise it to 15 via the Up arrow.
The freckles have disappeared, but so have all the finer skin structures. We now get them back by setting the threshold. To do this, we pull the Threshold slider to the right until all the structures are visible again and then move it successively to the left.
> Set the threshold to 70 and then reduce it to 20 in steps of 10 using the shift-down arrow.
> Fine-tune the threshold via the down arrow in single steps down to 12.
We stop reducing the Threshold value as soon as the large structures, in our case the freckles, have disappeared. Smaller structures like the fine pores of the skin are spared from the application.
> Apply Dust & Scratches via OK.
The Dust & Scratches filter has been applied to the entire image, as expected. Since we want to remove the structures only in certain zones, we now need to mask those areas that should be spared from the application.
We activate the Smart Filter mask and, in a first step, fill it with black foreground color to close it. Now we reach for the Brush Tool, provide it with a suitable tool tip and open the mask by applying white foreground color exactly where the effect should take hold.
> Background Copy 2: Activate the Smart Filter mask with one click.
> Set foreground color and background color to black/white by pressing the D and X keys.
> Fill mask with black via option+backspace.
> Activate Brush Tool: Size 150px, Hardness 0%.
> Set the foreground color to white by pressing the X key.
> Apply white in the mask on the skin areas to be treated.
Please make sure to spare important edges or structures like hair or eyelashes etc. when brushing and only brush the skin areas.
Once we have removed the freckles, we still take care of the finer structures. We also want to reduce these carefully. And we can do that with the Reduce Noise filter.
Since we want to reduce the noise in the same zones from which we removed the larger structures, we use the Smart Object with the already existing Smart Filter mask.
> Activate Layer Thumbnail.
> Open Filter / Noise / Reduce Noise.
Before we start, we again think about what we want to achieve as precisely as possible. We want to soften the finer skin structures, i.e. the pores, a bit, but not remove them completely. If we go too far with the noise reduction, the skin will get an unnaturally smooth texture. The processing would stand out as such. Our demand for a halfway realistic impression would no longer be met.
We make the essential adjustments in Basic Mode.
A zoom level of 200% is helpful when editing the fine textures.
> Activate the Basic radio button if necessary.
> Zoom Preview to 200%.
We set all sliders to 0, activate the Strength input field and raise it successively until we reach the desired degree of smoothing.
> Set all sliders to 0.
> Activate Strenght and raise it to 5 via the Up arrow.
Now, by increasing the values for Preserve Details and Sharpen Details, we bring back the fine structures a bit to achieve a more natural look of the skin texture.
> Activate Preserve Details and increase it to 20% via the Shift-Up arrow.
> Activate Sharpen Details and raise it to 50% via the Shift-Up arrow.
Since we find neither significant color noise nor blocky JPEG artifacts in the image, we leave both options disabled.
We finish the process with an OK and examine the result in 100% view.
> Apply Reduce Noise via OK.
> Zoom out to 100% via cmd+1.
We have gently smoothed the textures and the structures of the skin using a combination of removing larger structures and carefully softening fine structures.
Let’s do the before/after comparison by clicking the eye icon of the background layer while holding down the Option key. This will hide all the other layers and also show them again.
> Click the background layer several times while holding down the Option key.
Not bad. But the professional image editor won’t be completely satisfied yet. She would still remove the pigment or color spots of the skin and probably also bring back the shine on the tip of the nose a little. But we’ll leave it at that for the time being in the basics workshop.
Almost more important than blurring and smoothing image structures is the opposite task. Sharpening is an indispensable process in image processing and is usually performed at two neuralgic points in the workflow: At the very beginning of image processing and at the very end.
After opening the image, the image editor’s first look is at the image sharpness. In most cases, a careful improvement of the basic sharpness is called for. Adjusting the basic sharpness at the beginning of image processing is called input sharpening.
Once the image processing is complete, the required output files are created, e.g. CMYK TIFFs for printing and RGB JPEGs for the web. With the creation of the final format, a certain amount of blurring often creeps in again, which then has to be corrected in a very last step. Restoring the required basic sharpness at the end of image processing is called output sharpening.
In addition, it is sometimes necessary to sharpen individual parts of the image during the image processing itself.
- Locally limited zones are processed with the Sharpen Tool.
- If you want to sharpen larger areas, you can use one of the Sharpen filters.
A good choice for the many tasks of sharpening is given to us with the Unsharp Mask filter.
Let’s hide all the Smart Objects we have created so far and take a look at how the Unsharp Mask filter works using our example.
> Hide all other Smart Object layers.
> Create duplicate background layer, Smart Object.
> Open Filter / Sharpen / Unsharp Mask.
The strange name of the filter comes from a classic darkroom technique. In analog unsharp masking, the negative is exposed together with an unsharp negative of the negative. The effect is an increase in contrast and thus sharpening of the edges of the image.
And the Unsharp Mask filter in Photoshop also sharpens an image by increasing the contrast along the edges of the image.
The USM dialog features three sliders whose function we already know from other contexts.
- With the Amount slider we control the contrast increase and thus the strength of the sharpening.
- With the Radius slider we determine how far the effect may radiate into the vicinity of an edge.
- With the Threshold slider we separate the zones that should be sharpened from those that should remain untouched.
The settings that have to be made in the dialog differ depending on the state of the image. In any case, it is important to carefully observe the effects of the actions taken during the process.
For this reason we bring the view of the image itself to 200% and leave the preview field in 100% view. In this way, we not only keep an eye on the details, but also on the overall impression.
> Zoom the image to 200% via cmd-Plus.
> Preview field: Bring the right eye and the skin area below it into view.
We set all sliders to the minimum.
> Drag all sliders to the left.
This time we set the radius value first. The height of the radius value depends very much on the resolution of the image. The higher the resolution, the higher the radius needs to be. For a 300dpi image that is to be subjected to moderate sharpening, radius values between 1 and 2 are recommended.
> Set Radius to 1.0.
For sharpening we pull the Amount slider all the way up.
> Set Amount to 500%.
Of course, 500% sharpening is much too strong and not at all desirable. But the exaggeration illustrates quite well what over-sharpening can do to an image. The excessively high edge contrast can be seen particularly clearly on the hairs, which now have white fringes. But the skin texture has also clearly deteriorated. The pores become very visible. The skin takes on a dry, almost leathery appearance.
Under no circumstances do we raise the basic sharpness to such an extent that undesirable effects like these form. Careful handling is also required when sharpening.
For this purpose I set the Amount control to a low value, e.g. 10% or 20%, and then raise it successively in steps of 10.
> Set Amount to 10%.
> Raise Amount to 100% via Shift-Up arrow.
When moving the Amount slider, we pay attention to the edges in the image, i.e. the large contrast zones that carry the impression of sharpness. In our case, these are the eyelashes around the right eye and the eye itself.
Let’s now make a before/after comparison by temporarily deactivating the preview checkbox.
> Deactivate the preview checkbox and activate it again.
We have now set the radius and the degree of sharpness. Now we want to exclude the finest structures, such as the pores of the skin, from sharpening. Because sharpening so far also highlights the contrast of the smallest and smallest microstructures.
> Raise Threshold via Up arrow up to 10%.
We successively raise the threshold until we can no longer detect any enhancement of the pores. Then we stop, because we do not want to undo the sharpening effect we have gained in the essential parts of the image in and around the eye and in the hair.
We finish sharpening with an OK and make another before/after comparison by fading the USM layer out and in.
> Apply filter via OK.
> Fade the USM layer out and in again several times.
Our sharpening goal has been achieved and our careful approach has left no unwanted traces of work.
The subject still contains an important hint for us: sharpening is not generally about getting blurred images sharp or reducing depth of field. Parts such as the left eye and the left facial area in general cannot simply be “focused”. Focusing is the photographer’s job. The task of the image editor is to give the image a good basic sharpness or to sharpen individual parts. No more and no less.
At the end of our tour through Photoshop, we would like to devote ourselves to a genuine aspect of digital images, and thus of digital image processing, which causes uncertainty for many people, even though it should not. Once you have understood the importance of image size, or image resolution, with regard to the requirements of various output channels, the subject can no longer frighten you.
We open a new image for this purpose.
> Open exercise file 16.
Does an image have enough resolution? Will the image pixelate in print? Is the resolution sufficient for an adequate screen display? The image editor asks herself these and similar questions before taking the first editing steps.
Strictly speaking, the question of resolution must already be clarified when taking the picture. This is because resolution means image information. If not enough image information is generated during photography, the image editor is at a loss. The image editor will always demand images of the highest resolution from the photographer, i.e. images that contain the maximum of what the image sensor of the digital camera is capable of.
The reason for this is that the image resolution can be reduced without any problems, but it can be increased only poorly. When the resolution is “downsampled”, pixels are simply thrown overboard, whereas when “upsampled”, missing image information has to be invented by Photoshop.
Of course, the photographer will not always and everywhere produce photos in maximum image quality. Snapshots, party photos, private shots that are not to be fed to a printing process do not have to meet any output requirements. And that is exactly why it is important that the photo editor communicates her wishes to the photographer in time.
Some of you will have already been confronted with this problem in an unpleasant way: You want to prepare a great image for a particular output medium and find out that the resolution is too low. Of course, you want to avoid a rude awakening like this.
The maximum demands of the image editor on the photographer are therefore:
- File format: Raw.
- Color depth: Maximum.
- Quality, resp. resolution: Maximum.
We have already discussed the first two image aspects, now let’s move on to explore the mystery of image resolution.
Image resolution, how could it be otherwise in a pixel image, is measured in pixels per inch. In contrast, printer resolution is measured in dpi, or dots per inch. In practice, the two terms are often used synonymously. Thus, when asked about the required resolution, the printing technician will speak of dpi, whereas the image editor will speak of ppi.
But we do not want to be confused by this. What is important is to know what resolution is required in each case, its designation is irrelevant.
Although attention must be paid to resolution at the time of photography, and although resolution is also of great importance during image processing, it is ultimately an output requirement.
When we speak of image resolution, we mean the resolution of the image in the output format.
Let’s take a concrete look at the two standard resolutions to better understand this statement.
If the image is to appear accurate on a display, monitor or similar output device, it must have a resolution of 72ppi in the target format, i.e. in the desired image size. If the resolution is too low, it will appear too small on the display; if the resolution is too high, it will appear too large, because the standard monitor resolution is just this 72ppi.
Resolution and format are therefore in a stringent relationship to each other.
Currently, we view our image in a display size of 100%. And that means nothing else than that 72 image pixels are placed on one monitor inch. The display size therefore currently corresponds to the monitor resolution.
The same relationship applies to the print output.
If the image is to appear accurate in print, it must have a resolution of 300ppi in the target format, i.e. in the desired image size. If the resolution is too low, the image will pixelate in the print result, because the standard print resolution is just this 300dpi.
To clarify: If we change the display size in Photoshop by zooming in or out, this has no effect on the actual image resolution. We only change the view.
> Zoom in strongly via multiple cmd-plus.
A higher zoom level simply means that there are less image pixels on a monitor inch. The pixels just appear bigger.
Let’s return to 100% view via cmd+1 and open the Image Size dialog.
> Fit on Screen via cmd+1.
> Open Image / Image Size: Fit to Original Size, unit of measurement millimeters, Resample checkbox disabled.
Since Image Size is one of the most frequently used dialogs in Photoshop, we immediately remember the key shortcut: Cmd-option+I.
In the Image Size dialog we have gathered all the parameters and setting options that can be used to determine and, if necessary, change the image resolution.
In the upper area, the connection between image format and file size becomes obvious. Our image measures exactly 1037 x 778 pixels and therefore has a file size of 2.31MB.
The Fit to menu provides a number of standard formats that can be used to recalculate the image. My recommendation here is always to work with the Original Size entry and to adjust the format and resolution specifically by entering the values for width, height and resolution.
> Width: Expand and collapse the Units of Measurement menu.
If the image is prepared for print output, the width and height are displayed in millimeters or, in the case of very large images, in centimeters. The web designer would of course use pixels here.
Since the image editor is used to specifying the resolution in pixels per inch (ppi), Pixels/Inch is selected in the Resolution Units menu.
> Open and close the Resolution Units menu.
With these settings of the Image Size dialog we can read what is going on, i.e. how big the image is or how much image information it contains.
Adjust image size
Currently, our image has a width of 365.83 mm and a height of 274.46 mm at a resolution of 72ppi.
However, if we want to print the image, we do not need 72ppi, but the famous 300ppi. 300ppi ensures that the pixel size on the paper does not exceed the size of the printer dots, so we don’t have to worry about “pixelation”.
By entering the appropriate value in the resolution input field, it is now easy to meet this printing requirement.
However, please make sure that the resample checkbox remains unchecked for the time being.
> Enter Resolution: 300ppi.
As soon as we increase the resolution, the values for width and height of the image will decrease. This relationship between format and resolution is illustrated by the small link symbol.
If 300 pixels instead of 72 pixels are to be packed onto one inch, the format of the image must inevitably shrink. The values shown mean that our image can be printed without any problems, since it has the proper resolution, if it is not larger than 87.8 x 65.87 mm.
You have surely noticed: The Image Size, i.e. the value for the file size, has remained the same at 2.31MB despite the change in resolution. This is due to the fact that we have not recalculated the image size so far, but simply changed the scale.
This fact can be seen even after applying the new settings. Please pay attention to the ruler.
> Close Image Size via OK.
> Press cmd-option+Z several times for visualization.
The ruler changes, the scale is flipped. The image itself shows no change. We have modified the output size and resolution without changing the total number of pixels.
Let’s return to the initial state in the History palette and open the Image Size dialog again.
> Return to the initial state via Snapshot History palette.
> Open Image / Image Size: Fit to Original Size, unit of measurement millimeters, Resample checkbox disabled.
Resample image size
But what if the task is to print the image in the original size of 365.83 x 274.46 mm? 72ppi are not enough for this, we would need 300ppi as we know. But where to get the missing image information if it was not provided by the photographer?
In this case, we need to upsample the image, i.e. get Photoshop to interpolate the missing image information from the existing pixel material. And this is exactly what we can do if we activate the resample checkbox.
> Activate the resample checkbox.
Please note that now the link of width, height and resolution is no longer completely given. The resolution can be chosen freely.
> Open the Resample menu briefly and set it to Automatic if necessary.
I will talk about the content of the Resample menu later. Currently we select the default setting Automatic.
> Close the Resample menu again.
> Enter Resolution: 300ppi.
If we enter the desired 300ppi in the Resolution field, the effects on the image are already apparent.
The preview shows a greatly enlarged image and the development of the values in the upper part of the dialog box also give an idea of what will happen: Photoshop calculates the missing pixels and inserts them into the image. The number of pixels and the file size will grow enormously.
What effect this process will have on the image quality, which we have to keep an eye on in spite of everything, will become apparent in all clarity after the application.
> Close Image Size via OK.
Our image has been blown up in the truest sense of the word. The result is anything but satisfactory. We are confronted with the typical shortcomings that upsizing usually brings: essentially overdriven contrasts and artifacts in the edge area and blurring wherever you look.
The example illustrates quite well why the image editor wants to avoid artificially increasing the resolution of an image at all costs. She therefore always requires the photographer to produce images in the best quality, i.e. in the highest resolution.
Unfortunately, this requirement cannot always be met in practice. Whenever the photographer is unknown or unavailable, when the high-resolution version of the image has been lost or was never taken at all, the image editor must try to make the best of the available data material.
In such cases, the image editor is forced to upsample images as best she can.
> Return to the original state via cmd+Z.
> Open Image / Image Size: Fit to Original Size, Resample checkbox checked.
We enter the target resolution as usual, but choose a special algorithm from the Resample menu for the interpolation.
> Enter Resolution: 300ppi.
With the setting Automatic I leave it to Photoshop to find an adequate algorithm for the task. This works quite well one time, but other times the automatic application doesn’t give usable results.
To make sure that the most suitable algorithm is used in the process of moderate resampling, I therefore choose Preserve Details 2.0 with a high Reduce Noise value in many cases.
> Resample menu: Preserve Details 2.0, Reduce Noise 80%.
> Close Image Size via OK.
In fact, this AI-based algorithm usually delivers the most acceptable results. If the conversion seems too smooth, cancel it with Cmd+Z and try the whole thing again with a smaller Reduce Noise value.
If this approach does not produce an acceptable result, you must accept that the material is not suitable for printing in the desired size.