Channel Selections

One of the most popular ways to get precise selections from your photographs uses one of the oldest features in Photoshop: channels. As you remember from “How Photoshop ‘Sees’ Your Images,” channels represent the individual color components (typically red, green, and blue components, when speaking of images displayed on a screen) of your photograph, which are displayed in grayscale by default. The level of gray is roughly equivalent to the luminance of the given color. Because each color channel is represented as gray, we can use color channel information directly to create a special kind of channel called an alpha channel.

Alpha channels are essentially masks. More correctly, they are how layer transparency information is stored—instead of showing the brightness of a color, they show you the density of a mask, with white being transparent, and black being opaque (white reveals, black conceals with respect to the masked layer). Rather than being tied to a specific layer, however, alpha channels can be used to store or develop more complex masks that can be loaded and used anywhere in your document as a selection.

For selections, alpha channels work a little differently. The white areas represent what is selected, and black areas are what is ignored. So, levels of gray between black and white are “sort of” selected. When you load a selection from an alpha channel that has a range of values, think of the selection as density of color where white or lighter grays will have greater density than darks or blacks. Creating a layer from copying an alpha-loaded selection with gray values results in transparency being applied to the copied pixels.

Notice the Blue channel of this portrait shows the model’s hair as relatively dark. Try this: make a selection from the Blue channel, then choose Edit > Copy and Edit > Paste. This copies the selected pixels from the RGB channel and fills a new layer with them. In that new layer you’ll find that the area of her hair is mostly transparent. Because that area of the Blue channel was dark, little data was copied so there was nothing to paste into the new layer.

Let’s try a simple experiment that illustrates some of the selection techniques. (For details on creating this file, see “Building the Workbench Files” under “Project Examples.”) Each of the three solid color bars represents a canonical color in RGB space, so when we look at the Channels panel, each of those colors will be solid white on one channel, and black on the other two. The gradient at the bottom represents a complete RGB spectrum at full saturation (that is, the gradient range passes through both primary and secondary colors, so the pattern is Red, Yellow, Green, Cyan, Blue, Magenta, and finishing again with Red).

Holding Command (macOS)/Ctrl (Windows) and clicking one of the channel thumbnails loads the luminosity of that channel as a selection. In this example, I clicked the Red channel to select it.

The selection area is noted by a pulsing, dashed line of so-called “marching ants,” but that’s not a hard boundary. The ants are simply marching along a line through the 50% point of the selected luminosity range. The entire channel is actually selected. Clicking the Save Selection As Channel button at the bottom of the panel (it looks like a white box with a dark circle) creates a new alpha channel.

The Channels panel display always represents what is currently visible on the canvas, regardless of layer transparency, masking, adjustments, and so on. So what you select in the above approach is based on what you actually see. We’ll come back to this later.

Notice that the non-primary colors are represented by gradients on each channel. That means if we want to load up something other than the primary RGB color channels, we have to do something beyond simply clicking one channel. It turns out that Photoshop selection techniques support three basic Boolean operations—addition, subtraction, and intersection. The first two are probably familiar to everyone, but intersection is not quite as common. It simply means the overlap between areas, but nothing else.

The intersection of these three color dots is white.

Once you have loaded a selection, you can apply one of the Boolean operations by clicking while holding down a combination of these modifier keys: Opt/Alt, Cmd/Ctrl, or Shift. Remember:

• Opt+Cmd/Alt+Ctrl Subtracts

• Shift+Cmd/Shift+Ctrl Adds

• Opt+Shift+Cmd/Alt+Shift+Ctrl Intersects

For example, in the Gradient Experiments file I selected the Red channel, then subtracted the Blue channel.

Here’s the cheat sheet for selecting colors. For each color, begin by Cmd/Ctrl-clicking the first channel to select it:

• Red, subtract Green, intersect with Blue = Magenta

• Red, subtract Blue, intersect with Green = Yellow

• Green, subtract Red, intersect with Blue = Cyan

• Green, subtract Blue, intersect with Red = Yellow

• Blue, subtract Red, intersect with Green = Cyan

• Blue, subtract Green, intersect with Red = Magenta

Selecting the Magenta segment with only channel operations requires noting that Magenta is made up of Blue and Red, but no Green. So load either the Red or Blue channel, intersect with the one you didn’t choose, and then subtract Green.

Using Calculations

Applying this to a photograph is exactly the same, but of course the selection result will be much more complicated. In this portrait, I want to isolate the blue jewels with a mask and change their color to green.

The Calculations dialog box is another option for making selections from channels, and has a lot more power than using the above Boolean operations, because it allows you to use blending modes as part of the process. Even better, it gives you a real-time preview of your results. The major limitation is that you can operate on only two channels at a time; as we’ll see in a moment, Boolean selections still have a place in your workflow.

I loaded the Blue channel as a selection using the previously described methods, then subtracted both the Green and Red channels. At this point, the marching ants that denote the selected area vanish because they are visible only when the selected region is more than 50% opaque. However, the selection is still active—trust me!

I then made the RGB composite channel active and in the Layers panel copied the selection to a new layer as above. In order to see it, you may need to put a black- or white-filled layer between the copied result and the background layer. I named the filled layer B − G − R (blue minus green, minus red).

That is the beginning of our selection, but we need it as an alpha channel or active selection. Be certain to hide the additional layers you created once you have your alpha channel. The first result could use some refinement, so this is a great excuse to demonstrate Calculations. With Calculations, there is no need to have an active selection or to have any particular layer active or visible. Choose Image > Calculations to see the dialog box.

Because the copied selection is already the result of the Blue channel minus both Red and Green, I want to further isolate the blue elements as much as possible. In the upper portion of the dialog box, under Source 1, I need to choose the copied layer (B-G-R) from the Layer menu, and then choose Blue from the Channel menu. In the Source 2 section of the dialog box, I choose the Background layer and its Red channel, then choose Divide from the Blending menu. Fortunately, Calculations gives us a preview of the results on the canvas.

Notice that the result is the opposite of what we want: The jewels are black and everything else is white. At the bottom of the Calculations dialog is a menu for selecting the Result (New Channel, New Document, or Selection). Choosing New Channel is the best option in this case because we need to invert the mask and saving the selection as a channel makes that easy. Click OK to close Calculations, and the new channel appears in the Channels panel. To invert it, make the new channel active and press Cmd/Ctrl+I. Cmd/Ctrl-click the channel to load it as a selection.

I added a Curves adjustment so the active selection was applied immediately as a mask. The final move is to choose the Blue channel from the menu next to the Auto button in the Curves Properties panel and drag the lights (right side) down to about 50 for Output. That removes most of the blue information, leaving the Green channel results visible.

Gradient Map Luminosity Selection

You can never have enough selection method options in Photoshop. This one came about from my desire to create a fast, arbitrary luminosity mask (that is, a mask created by choosing specific luminosity values or ranges) without going through all the usual convoluted operations of making tons of small masks and merging them together. Although I love using luminosity masks for detail work and fine control over richly colored images, sometimes I just need a little slice of value range without too much fuss.

A Gradient Map is used to isolate a range of values in your image based on luminosity, and you can set up virtually any region and width pretty easily. As a bonus, you get near real-time results while you’re tinkering, so you don’t have to guess at which values are being included.

Start with a simple Gradient Map adjustment layer, and choose the Black, White gradient in the Basics group. Click the gradient sample to open the Gradient Editor, drag the white color stop to about the 50% location, and add a new color stop at the end. Make the new one black. You should have a black-white-black gradient so far.

Bring both of the black stops towards the white stop, and watch the effect on your image. Adjusting the position of the three stops lets you isolate different ranges of luminosity values visually, rather than taking a more clinical or math-based approach, which is how many plug-ins operate. Although those methods are powerful, the ability to see the changes to the selection immediately is also incredibly useful.

From here, you need to convert what you see into an alpha channel. Click OK in the Gradient Editor, open the Channels panel, and drag any of the channels down to the Create New Channel button. This automatically creates the alpha channel for you. Name it something relevant to your selection region, such as Mid Shadows.

Back in the Layers panel, turn off the Gradient Map visibility, then choose Select > Load Selection. Choose the channel you just created from the Channel menu.

Now, simply add an adjustment layer or a layer mask to an existing layer and the luminosity mask will be applied automatically. You can create any number of alpha channels the same way.

In this ship example, I wanted to both warm up the sky, and cool off the low shadow areas to create a duotone effect. Using the previous selection, I targeted first the shadow regions and applied a Curves adjustment to provide a misty blue cast. Using the gradient method allowed me to visually dial those areas in without making manual selections. I could also disregard the darkest shadows to allow them to retain just a little detail but otherwise achieve full black.

Warming up the sky required another pass with a different gradient for the highlight regions. Again, I did not want to include the extreme white areas because they would “blow out” to flat white, but keeping the texture in the clouds was a little tricky. To sell the effect, the clouds also needed to retain just a touch of the original blue-gray. The adjusted gradient is not quite centered, so there is more feathering of the mask in the higher value regions. You can change the rate of transition by dragging the small diamond-shaped midpoint controls that appear next to a selected color stop.

I used the final selection to create a mask for a Color Lookup adjustment layer (CLUT), an adjustment layer I describe later in the “Color & Value” chapter.

Most luminosity selection methods have fairly uniform boundaries and specific ranges. Using a Gradient Map gives you quite a bit more freedom to select exactly what you need without having to mix and match or blend different masks together. It also lets you create only the masks that you need, while lots of plug-ins and automated methods give you tons of masks that you may not use and then have to delete.

Once you have the basic gradient map created, save it as a preset for easy recall. I named my starting point Luminosity Selection, because I typically start from the same mid-range. However, you could easily create a series of presets for specific ranges (more on that in a bit).

The stops, of course, can swap positions, too. If you want to select just the shadows, for example, simply drag the left black stop towards the middle a little, and then drag the white stop over it to the left edge.

Combine this technique with Boolean selections (see the “Channel Selections” section of this chapter) to carve out very specific areas of your image quickly. Remember to store the results as alpha channels with a good naming system so you can easily get your selections back.

Hue/Saturation Selection

One of the secrets to making good selections is to understand that Photoshop can present you with alternative views of your image data. Some of the blending modes in Photoshop feel more utilitarian than others, and this technique involves the Difference blending mode—a tool most people associate with aligning image layers or detecting changes. When combined with a Hue/Saturation adjustment layer, however, Difference has the fantastic ability to isolate elements in your image.

The basic idea here is pretty easy to grasp: Make a slight change to some element of your picture, compare it with the original, and then extract the difference. In this case, we’re going to change the color of something and create a mask from the result of that comparison. Don’t lose sight of the fact that we’re introducing a change, which could be just about anything we can target; it doesn’t just have to be color, but the most useful techniques will involve features we can easily change with some level of automation.

To start with, add a Hue/Saturation (H/S) adjustment layer above your image layer. Check out the color bars (the lines at the bottom of the Properties panel that look like spectra—from here on, I’ll refer to them as color ramps): This is what we’ll use to make the change, but we have to make a selection first. Click the adjustment layer’s icon in the Layers panel so that you are not making a selection of the mask. This is a weird quirk in Photoshop: When you add the adjustment, the mask is automatically selected, so that if you try to use the Targeted Adjustment tool, you’re sampling from the mask instead of the image.

Select the Targeted Adjustment tool in the Properties panel. Find a region you want to select whose color is distinct from its surroundings, and click. Don’t scrub, drag, or anything else—just click in the area of interest to load the color that you want. Look at the color ramps back in the Properties panel and notice the indicators have changed.

Also notice that the color ramps have updated to the closest primary base color from your selection. Now before we can make things work, change the H/S layer’s blending mode to Difference. The canvas should go black because there is no difference—yet. Drag the Saturation slider in the Properties panel all the way to the left to desaturate the image.

What you end up with is probably very difficult to see in most circumstances. For the moment, you can see things better by adding a temporary Levels adjustment layer to the top of your stack for more contrast. You’ll get rid of it in a moment, so think of it just as a helper layer. In the H/S color ramps, drag the right-side control handle back and forth a little to see how it affects the results. By manipulating both sliders, you can refine the selection region to display more or less color variation.

At this point, don’t worry about areas that you do not want to include; it’s easy to remove them by painting black in the layer mask in the upcoming steps. The outer points are basically thresholds that clamp down on the color being affected, and the center slider is the most important or strongest color being selected. Adjusting all of them changes what will end up in the selection process.

Now the question is how to get from this odd view to a reasonable selection. Use The Claw (Opt+Shift+Cmd+E/Alt+Shift+Ctrl+E) to create a stamped copy of what is currently visible on the canvas, and name the layer Mask Source. On the Mask Source layer, first desaturate it using Shift+Cmd+U/Shift+Ctrl+U to completely desaturate the layer without opening the H/S dialog box. Click OK, and now turn off the Levels adjustment layer if you were using it. Making sure the Mask Source layer is active, press Cmd/Ctrl+M to open the Curves dialog box. Making changes here will apply the adjustment directly to your layer, so be sure to take care! Adjust the curve for maximum contrast without making pixelated edges.

In many cases, you’ll want to preserve the textures in your selection, so don’t feel like you have to make everything completely white inside the area of interest. For more solid masking, use a slight Gaussian blur of around 1–3 pixels (this value will vary depending on the actual pixel size of your image), then use the Levels command to tighten up the edges. Check out the next section, “Tweaking Channels & Masks,” to get details for enhancing the mask.

Once you have the contrast and density you want, clean up any unwanted areas by either painting over them with black or using the Lasso tool to select large areas and fill with black. To check the accuracy of your mask, change the Mask Source’s blending mode to Multiply, which lets you see through the white regions of the layer (be sure to turn off any other adjustment layers you were using, especially the H/S). If you need to make refinements, this is the time to do it. Remember that Multiply makes white areas transparent in the current layer allowing color information from lower layers to show through.

When you have an acceptable selection area, set the blending mode back to Normal, and then open the Channels panel. Drag one of the channels to the New Channel icon to duplicate it. You have a shiny new alpha channel ready to go! Doing this will leave your new alpha channel selected in the Channels panel, so you will need to click back on the RGB composite channel before going back to work on your layers in the main document. Name your new alpha channel something appropriate for organization, such as Magenta Swatch.

To use your new alpha channel, simply Cmd/Ctrl-click its thumbnail to load it as a selection. You may optionally continue working on your channel before making a selection, or even use it with other techniques to create additional alpha channels. Remember that if you create a mask or add an adjustment layer with an active selection, that selection will be used as the layer’s mask.

Use this technique for situations where you need some delicate selections while also having the advantage of somewhat uniform colors and good boundaries. It’s more precise and flexible than the Magic Wand tool or other direct color selection methods, and really does not take much time. More to the point, this should give you some additional insight into creating so-called difference features that you can exploit for (slightly) automated selection tasks.

In this case, I used another Hue/Saturation adjustment layer (set to Normal blending mode) with my new selection to change the magenta areas to green.

Tweaking Channels & Masks

Sometimes getting an amazing mask takes more than great selection techniques. As the previous sections illustrate, you may have to get in there and do some serious tweaking to get exactly what you want. Select Subject does some mind-blowing work, but it rarely provides a one-click solution. Let’s take a look at a handful of ways to refine an already created mask by treating it like a regular Photoshop image layer.

Masks typically behave like any other pixel layer; they just operate solely in shades of gray. For the most part, working on a mask directly lets you see the results on the layer, but you cannot see the mask itself. There are situations that call for a different view, though, especially when the material being masked has a lot of visual clutter and confusion, or you want to carefully control the mask development. The first trick most people learn about, when moving to more advanced mask editing, is to hold Opt/Alt while clicking the mask thumbnail. That presents a view of the mask itself.

Being able to see the mask directly gives you quite a bit of power and flexibility, and also it sometimes keeps you from missing important details that may otherwise be lost. This same view is available in the Channels panel when the masked layer is active and selected—Photoshop creates a temporary alpha channel that responds to the same tools as the layer mask. All of the layer-based tools available in Photoshop can be used on a mask or alpha channel, but these are destructive changes. How about a non-destructive workflow?

That requires promoting the mask or alpha channel to a regular layer.

Promote the Mask

There are a couple of ways to promote a mask (that is, to convert a mask to a regular layer). View the mask as mentioned previously and use Cmd/Ctrl+A to select the entire mask, then copy with Cmd/Ctrl+C. Create a blank layer, and simply paste the material back in with Cmd/Ctrl+V. Boom, you have a mask on a regular layer (remember to turn off or remove the original mask so you can see your changes). This also works for copying an alpha channel to the layer stack. In this case, you will have an exact duplicate of your mask or channel. Set the new layer’s blending mode to Multiply, which will cause anything that is not black to allow lower layer content to show through.

Another common approach is to load a mask or alpha channel directly as a selection using either a keyboard shortcut or menu command. For the shortcut, hold Cmd/Ctrl and click the layer mask or alpha channel. Alternatively, use the Select menu. With your masked layer active, choose Select > Load Selection and in the Channel menu, choose the option that gives your layer’s name with “Mask” at the end. Both methods give identical results.

Once you have an active selection, create a blank layer and fill the selection with white: Press D to ensure your background color is white, then press Cmd+Delete/Ctrl+Backspace. If you feel the mask needs to be more dense, repeat the fill shortcut a few times. After filling, press Cmd/Ctrl+D to deselect. Below the newly filled layer, add another blank layer and fill that one with black by pressing Opt+Delete/Alt+Backspace.

Remember that loading a selection treats gray values as partially selected, so white areas are fully selected, and black areas are not selected, with gray in between. When you fill in such a selection, the fill respects the transparency amount in the selection so while the color of the fill is technically white, it will be semi-transparent.

Group the filled selection and plain black layers, and set the group blending mode to Multiply. Now you can see the effects of the mask while working on a regular pixel layer.

It’s an extra step, but I prefer this workflow of filling a selection because it provides greater flexibility as you work. It allows you to create additional layers, add multiple masks together, and to paint non-destructively. When the mask work is complete, you will need to set the group blending mode back to Normal, and then create an Alpha channel as described earlier. Instead of just making a direct flat copy of your mask, you may want to fill the selected region with a solid color or gradient, or modify the selection with the Expand/Contract commands on the Select > Modify menu prior to filling. Further, you can opt to fill the selection to a blank layer, leaving the flat black layer as a base. These are more difficult to accomplish starting from a flat copy.

Let’s now take a look at some nifty mask editing tricks that also demonstrate a little bit about how layers—and adjustment layers in particular—work inside of Photoshop.

Levels Tweaking

Note that for this part of the example, we’re looking at a single layer consisting of a white square surrounded by black. This will become important shortly. I have added guides to show the exact boundary of the original square. The square represents our starting mask, as if we had copied and pasted an alpha channel to a regular layer. I’ll refer to this as a “mask layer” during this exercise.

Start by adding some blur to the layer, then add a Levels adjustment above the layer. Using a Gaussian blur of about 10 pixels is just fine for this demonstration. Now drag the right (white point) Levels control toward the left side of the display. The white square should grow a bit and extend beyond the guides.

Right away, this should give you some ideas about manipulating your masks. Of course, dragging the left control to the right has the opposite result, causing the square to shrink. And if you bring control points together, you can start to close in on a fairly sharp edge. Notice what it does to corners, however; they end up rounded. This is something you’ll have to pay attention to in some circumstances, but for general work you’ll use very small blur amounts in the range of 0.5 to 3 pixels, so the effect won’t be a problem.

You can use this basic technique to smooth out rough, pixelated masks, especially those created by automated selection tools, such as the Magic Wand, Object Selection tool, or Select Subject. Although those features can achieve awesome results in a few clicks, the masks often need just a little cleaning. Take careful note that Select Subject or Object Selection tool will attempt to create a solid fill. Just tuck that little insight away for future reference.

Levels adjustment has the added advantage of letting you place your edge boundary almost anywhere within the blurred region. To recover the original edge placement, drag both end points toward the middle.

There are lots of ways to apply this method, too. Let’s say you’re building a complicated mask with several parts, and you have just one section that needs some smoothing because it has a little pixelization. In the next example, I used the Blur tool to manually smooth out the pixelated area, then used the Lasso tool to make a selection around it, being sure to include some of the already smooth, crisp areas. With the selection active and calling up the Levels adjustment (not adding an adjustment layer) with Cmd/Ctrl+L, I tweaked the edge into shape. Because the smooth areas already have high contrast, they aren’t affected and the blurred area blends right in.

Why was it important to note the figure is only a single layer earlier? It turns out that the way Photoshop handles layers has an impact here. Recall from “How Photoshop ‘Sees’ Your Images” that what you see on the canvas is a representation of all the visible pieces from each layer. There are situations where you can’t immediately see a visible change between different ways of ordering layers, such as placing an adjustment layer inside a group versus outside of the group. This can cause some frustration, but it also gives me an opportunity to highlight the importance of knowing how your tools behave under different circumstances.

If I had placed the white square on a layer by itself, surrounded by transparency with the black background layer below it, the above experiments would have worked exactly the same unless we made a small change. If the Levels adjustment were instead clipped to the white square layer, it would have appeared as though the adjustment had no effect.

This happens because Levels (like all other adjustment or layer groups) operates as if everything below it is a single layer—everything coming in from below is added to its equation. Clipping the adjustment layer limits the input to only what is on the layer that it’s clipped to. In this case, the fuzzy edges are blending with transparency, but Levels deals with actual pixels only and does not affect transparency at all. When we move the sliders, there is nothing to compare with, so the fuzzy edges don’t change.

For Levels to behave like an edge refinement tool, it needs something for each pixel to be compared with, and that means either working on a flattened layer, working on a group, or not clipping the Levels adjustment to the white square layer. Let’s illustrate this by adding a black circle to the middle of a white square, leaving a transparent region all around the square. The circle and square are on the same layer, and so the blur has been applied to both.

Applying a blur to the Square and Circle layer gives predictable results. The edges of the white square become soft and expand into the transparent regions, while the edge of the black circle blend with the white square. The blurred regions appear identical at this point.

Placing a Levels adjustment at the top of the stack and dragging the controls to the middle gives us what we expect—hard edges for both shapes (with rounded corners).

However, if we clip the Levels adjustment layer to the white square layer, check out the results.

The internally “captured” pixels of the fuzzy circle can be compared with their neighbors, so those get affected. The outside edges of the white square just sit there, happy and unchanged in isolation. Of course, if there were colors within the figure, or it were something other than flat black or white, Levels would have the usual effect, just as with the captured shape in the figure. In order for Levels to affect the edges of the white square, we either have to unclip it from the square layer, or group the black and white layers, and then apply Levels to the group.

Why should you care about this in context of masking? Because you can exploit blending modes to give you a view of your mask result in isolation while you are doing your refinements. Being able to see what does and doesn’t work with the pieces of a mask as you work can save a lot of frustration.

Multiply Blend for Group Masking

Let’s look at a relatively simple example. This gull is easily selected with the Select > Subject command, which does an amazing job. But it leaves a tiny bit of halo around the bird, making it difficult to use in a composite.

It’s not too difficult to paint around, but that does take time and precision. A less labor-intensive approach is to set up the layers as shown.

Select Subject loaded the gull’s outline pretty well, so I created a blank layer and filled the active selection with white. I then put a solid black-filled layer below the white selection. I grouped these two layers, and then set the group to Multiply blending mode. Black areas remain solid and opaque, but anything white allows lower layers to show through.

Now I can add a little Gaussian blur to the white layer, and then add a Levels adjustment layer above that. The amount of blur should be just sufficient to allow the boundary to be moved inward a little, typically 3 to 6 pixels. Notice that the Levels adjustment does not affect the gull itself, only the interaction between the white and black layer in the group. If I had placed the adjustment layer above the group, that would not be true: Anything showing through the mask would be altered.

With this set up, I can freely adjust Levels to manipulate the boundary edge however I like. Although some bits under his wing and along his back need cleaning up, those are a lot smaller to deal with. After refining the mask, I set the group blending mode back to Normal, which shows me solid white and black. A quick trip to the Channels panel and I can duplicate any of the visible channels to save my mask. From there, I just load it as usual for masking.

Of course, you can load any selection you like, and apply the same tricks to only active selections.

Painting with Blending Modes

Sometimes there is no getting around manual methods for masking. That means painting. But again, Photoshop lets you do some amazing things with just the basics. Channels are often a great starting point for masks, because they rely on the image data directly. With a little tweak of Curves and Levels, you can get some amazing details and subtlety with little effort.

While adjustments affect the entire layer at once, painting relies on you as the artist to make choices. The red channel from this portrait has a stunning amount of detail around the model’s hair, so it’s a great candidate to start with for lifting her out of the background. Duplicating it gives us an alpha channel to work with.

All by itself, it’s nearly perfect, but does need some refinement. Grab a large, soft round brush and set the brush blend mode to Overlay. Lower the Opacity and Flow to around 25% each. When you’re painting with black, solid white areas will be untouched, but areas darker than about 50% black will be burned down to total black. The low opacity and fill values let you sneak up on tricky areas a little more slowly.

Note

When you are done with this technique, be sure to set the Brush blending mode back to Normal. Failing to do so may result in extreme frustration, and I do not need you throwing my own book at my head when your brush appears “broken” because you forgot about tool blending modes.

I simply brushed over the curls area to increase contrast and smooth out edges. I also spent a little more time around the edges of the mask area to further darken the background before tackling the corners. To speed things up, I switched to Normal blending mode for the brush and painted directly in those areas, being careful to avoid the model’s hair.

From here, it’s easy enough to continue painting with white and normal blending mode in the interior of the mask to remove facial features. Using a gentle approach just takes a couple of minutes, but yields some amazing results.

Photoshop being what it is, there are several ways to approach this challenge. After duplicating the channel, load it as a selection before painting. Try adjusting with Curves or Levels first, then brush with the selection still active. This can introduce some fringing, so this is a technique that’s ideally suited for the group Multiply blending mode trick shown in the previous section. That way you can see the results on the image as you paint and adjust. You can hide the selection’s marching ants by turning off Extras under the View menu. Photoshop being what it is, there are several ways to approach this challenge. After duplicating the channel, load it as a selection before painting.

Remember that the goal of using masks is to make your life easier during the creative process. Look for ways to accomplish that goal using the data and tools available to you. This entire chapter is meant to illustrate various ways to do just that, not present a “best” workflow that is guaranteed to get results in every situation! In addition to the specific techniques listed previously, you should spend some time investigating other selection methods and perhaps even automating them.

If you do a lot of portrait retouching, for example, it may be worthwhile to dial in some methods that are specific to skin tone selection and create actions that set up layers to get you at least part of the way there. A quick web search turns up a few third-party plug-ins you can purchase that create some exquisitely fine luminosity masks with utterly absurd levels of control. Personally, I think it’s money well spent if such tools help you express your vision more easily; but the true value can only be felt once you understand exactly what they’re doing for you!