It’s fair to say that Adobe does not really give you much information about Apply Image.
Apply Image applies a Blend Mode to a Target, using a Source. That Target or Source can be an Image, Layer Mask, Channel, or Selection.
Apply Image is useful in these situations:
- Frequency Separation (used for skin retouching).
- Creating simple luminosity masks.
- Refining or creating Alpha Channels for masks.
I’ve seen Apply Image used for the general enhancement of an image. But that’s more interesting than useful in my opinion. Using Layer Blend Modes will achieve the same effect without resorting to a relatively complex command, like Apply Image.
YouTube Video on Apply Image
Apply Image Workflow
- Select either the: Layer, Layer Mask, Channel you want to apply (the clue is in the name) the effect to. This now becomes the Target. Selecting of the Target has to be done first, you cannot change the Target inside of the Apply Image window.
- Go to: Image/Apply Image… Make sure Preview is ticked. With Preview ticked you can see the results in the Document Window in real time.
- Set the Source Document, which defaults to the image you are working on. You can pick another Source image, only if the image is open, and of the same Pixel Dimensions.
- Then choose the Layer, and Channel, or Transparency. Then Invert if necessary. Please note Merged (merged layers) will be an option if you have more than one Layer – take care.
- Pick the Blending you want. There are two Blend Modes called Subtract, and Add which have additional controls of Scale, and Offset. All other Blend Modes behave in the same way as Layer Blend Modes.
- You can change the Opacity, Preserve Transparency, or output the result via a Mask.
- Source = Document, usually the same image. It can be another open image with same Pixel Dimensions.
- Layer = Any Layer within the source. Including all the layers Merged.
- Channel = Any channel including, Alpha Channels, Colour Channels, or the Layers Transparency
- Target = Channel, Layer or, Layer Mask, you have selected in advance.
- Blending = Any Blend Mode. Add, and Subtract give you access to Scale and, Offset.
- Opacity = You can make the effect of the blending less opaque if you wish.
- Scale = Only available with Add, and Subtract. Scale divides the result of the Add, and Subtract blends. A scale of One (1) has no effect, two (2) has the maximum effect. Any number above one, to three decimal places, will make the result darker, i.e. 1.578.
- Offset = Only available with Add, and Subtract. The result of the Scale is passed to Offset. Offset has a range of 255 to +255, to darken, or brighten.
- Preserve Transparency = Only the opaque pixels will be affected.
- Mask = Output the result of the blending via a mask.
The two Blend Modes Subtract, and Add is the most flexible because they give you access to Scale, and Offset.
In the scenario above we subtracting the luminosity of Source, Composite Channel (RGB), from the Target.
Simply put we are subtracting the overall luminosity of the “Target” using the Source Layer’s overall luminosity, or grey value. It helps to think of the Target’s pixel grid laying on top of the Source’s pixel Grid.
In the Layer blend modes, they use standard values, where zero is black, and one is white. I am not entirely sure what numbering is used in Apply Image, as Adobe does not divulge this type of information. For all we know it could be a mixture of integer methods.
But, using the 0-255 system works well to explain with this scenario, where zero (0) is black, and 255 is white, even if it’s not the Integer system Adobe actually uses.
Let’s say we have a number of 200 Pixel Value on the Target layer’s pixel grid. And, the Source pixel has the value of 100.
200 minus 100 = 100 (Target minus Source)
The result is now darker. Subtract will always Darken, and Add will always lighten.
The 100 is passed to Scale, which then divides the result from the Subtraction. A scale of one has no effect. A scale of two has the maximum effect.
Let’s assume a Scale of two:
100 divided 2 = 50 – making the image darker.
Note: Dividing any number by one, changes nothing. Any scale number above one will result in a darker result with softer transitions.
The final stage is the Offset, ranging from -255 to +255. The reason for such a wide range is that it’s entirely possible to have a number below zero. For instance minus (-200) luminance. We will only see black on our screen, but the 200 is stored and passed on to the Scale.
You are effectively using Offset to bring the result back to the visible range of 0 – 255.
Offset to brighten or darken
In this scenario, a positive Offset of 128 would make it brighter.
50 + 128 = 178
Opacity can be useful sometimes if you want to lessen the effect of the blending.
Mask… looks complex, it isn’t. Mask simply outputs the result of the blending via the luminosity of a channel, or area of a selection. By the way, a selection is seen as an Alpha Channel by Photoshop, but you will just see the marching ants.
Let’s say you use one of the colour channels. Anything that is white on that channel reveals. Anything that is black will conceal. And grey will allow results to show depending on how near to white it is.
A little-known method is to select an area on an image first. Then run Apply Image, and everything will be output via the selection. The confusion might come if you want a selection as a Mask, which as in the example above. which is really counterproductive.
Apply Image is abused by many who create YouTube Videos. A lot of techniques are more interesting than useful.
For Frequency Separation it’s very useful. For creating simple highlight and shadow luminosity masks, it’s ok. Beyond these two examples, don’t worry if you don’t understand it – just play around with it.