Photoshop and Lightroom/CameraRAW have a great way to control colours in your images: vibrance. What vibrance does is to enhance the saturation of desaturated colours in your image, leaving alone already saturated colours. What it doesn’t is to add colour variation.
In this tutorial I’ll explain a technique I came up about how to achieve a better colour variation in desaturated parts of the image, using the Lab colour method and saturation maps.
This is a fairly advanced technique, and I assume you have a basic understanding of Photoshop. I will try to be as detailed as I can, but you have to be able to move around Photoshop and perform basic tasks.
Just a note before we begin: I work on a Mac, so I’ll use Mac conventions. To convert them in Windows shortcuts, simply substitute ⌘ with CTRL and ⌥ with ALT.
To explain this technique I will use my Boy With a Green Umbrella image you probably already know. This image is already pretty colourful, but it is quite dull on the pavement. I want to add colour variation there.
We can start by duplicating the image using ‘Image → Duplicate…’. When prompted for a name, you can write “Lab”. Now we can convert the image in the Lab colour method: ‘Image → Mode → Lab Color’.
What is Lab colour? Lab is a colour method alternative to RGB. You should know RGB creates an image assigning every pixel a mix of red, green and blue. Lab works in a different and somehow twisted way.
The first channel is ‘L’, which contains the luminosity information of the image. The other two channels, ‘a’ and ‘b’, contain colour information. ‘a’ channel ranges from green to magenta, and ‘b’ channel ranges from blue to yellow, allowing us to create imaginary colours (colours that exist only mathematically, but not in reality).
I know, it sounds quite complicated, and it is! I will write about Lab in the future on this blog, but for now it suffice to say that Lab allows us to manipulate colour and luminosity separately.
If you really want to go deeper in understanding the Lab colour method and the magic you can do with it, you can read Photoshop Lab Color: The Canyon Conundrum and Other Adventures in the Most Powerful Colorspace by Dan Margulis. Beware this is a complicated book, quite scientific in its approach.
Now that we are in Lab Color, we can apply a technique that Dan Margulis calls Man from Mars. To do this step, we have to add a Curves Layer clicking on ‘Layer → New Adjustment Layer → Curves…’. We will set Lightness, a and b curves separately.
Before going ahead, I have to explain how I set up my curves panel. By default Photoshop shows the amount of light. I work a lot with prints, so I set it up to show the amount of ink. To change this behaviour you have to open the Curves Display Options and change the radio button from Light to Pigment/Ink %. A method isn’t better than the other, you have just to choose one and stick with it. For the sake of following this tutorial just stick with me or you’ll have to reverse all the curves.
Let’s go on. First of all, let’s find a representative spot on the curve using the colour picker on the pavement, and let’s place a dot on the Lightness curve. Then let’s move the mark on the horizontal axis halfway, just like in the figure.
Using the colour picker, let’s sample the same point for the ‘a’ and ‘b’ channels. Then, let’s move the mark on the horizontal axis until the top of the curve “hugs” the top of the frame. Look at the screenshots to see how it looks like.
At this point the image should look crazy (just as it came from Mars!), but try to look just at the pavement. You should notice a better colour variation. What the Man from Mars has done, is to enhance the simultaneous colour contrast of the image. We’ll tone it down soon.
Now we have to boost the colour variation even more. Let’s duplicate our merged layers. Just press ⇧⌘⌥+E and change the blend mode to ‘Multiply’. You should see your image become overly colourful, but at the same time you should see a greater variation and depth.
Play a little with the opacity of the last two layers to came up with something you like. Every image is different so there isn’t a universal value. Remember that the image should look a little over the top because we will be masking it. My values for this image are 100% for Man from Mars and 50% for the Multiply layer.
When we are satisfied, we can press ⇧⌘+E to collapse all the layers, then ‘Image → Mode → RGB Color’ to turn it back to RGB. Now we can go back to the original image. If you don’t have tabs visible, you can click on ‘Window’ and you’ll find open windows at the bottom of the menu.
Once we are on the original image, we can create a new empty layer (⇧⌘+N) and name it “Lab”, then we can open the ‘Image → Apply Image…’ command. When the window pops up, choose as ‘Source’ our ‘Lab’ image (the other image opened), as ‘Layer’ the ‘Background’, as ‘Channel’ the ‘RGB’, and ‘Normal’ as ‘Blending’.
After this step, we should have our crazy over the top Lab image on top of the original image layer. Now we have to mask it! As I said at the beginning of this tutorial, we want to add our colour variation to the less saturated parts of the image, so we need to map the saturation of the image.
Among colour methods, you probably have heard of the HSL colour method. In HSL, the three channels represent respectively Hue, Saturation and Luminosity. What if we could use the Saturation channel?
To do that you need a free plugin by Adobe. I don’t know why Adobe doesn’t install it by default in Photoshop, but you can easily download and install it. For free. The plugin is HSBHSL and is contained in the Adobe Photoshop Optional Plug-ins ( Windows Version).
Let’s duplicate our image again: ‘Image → Duplicate…’. This time we can name it “ HSL”. Then, let’s discard the “Lab” layer. Now, let’s convert the image in HSL: ‘Filter → Other → HSB/HSL’. In the option window that pops up, let’s choose as input “RGB”, and as output “HSB” or “ HSL”. It’s the same since we need just the ’S’ channel.
The result of this operation is a strangely coloured image. We can’t care the less, because we are interested just in the channels. Although the names are still “Red”, “Green” and “Blue”, the contents of the channels are “Hue”, “Saturation” and “Brightness/Luminosity”. Right, we have a saturation map in the “Green” channel. Let’s select it!
What this ugly black and white image represents? Easy! White represents saturated colours, and black represents desaturated colours. Since with our mask we want to target desaturated colours, we have to invert this channel. With the channel selected press ⌘+I. Now, to load the channel as selection, ⌘+Click on the just inverted “Green” channel. To obtain a greater separation of the saturated and desaturated parts of the image, let’s multiply the selection by itself: with the selection loaded, ⇧⌥⌘+Click on the “Green” channel. Now let’s create a new black channel and, with the selection still loaded, let’s fill it with white.
I repeated the steps above on the new “Alpha 1” channel to make an even stronger “Alpha 2”. In my image I obtained this “Alpha 2” channel:
This is our vibrance mask. Beautiful, isn’t it?
At this point all we have to do is to import the “Alpha 2” channel back in the original image, and apply it as a mask. Let’s do that!
Let’s come back to the original image where we have the background and the Lab generated layer on top of it. Let’s select the Lab layer and let’s create a new layer mask. Just press the little rectangle with a small circle in it at the bottom of the layers panel.
With the newly created layer mask selected, let’s hit ‘Image → Apply Image…’. In the option window let’s select as ‘Source’ the ‘HSL’ image, as ‘Layer’ the ‘Merged’, and as ‘Channel’ the ‘Alpha 2’ (our vibrance mask). Let’s select ‘Normal’ as ‘Blending’.
Now the image should look natural and colourful. Try to toggle the visibility of our “Lab” layer e let’s see the difference. Quite better, isn’t it? If it’s still over the top, you can tone it down setting the opacity of the layer, or darkening the mask using a ‘Curve Adjustment’.
Use the slider to look at the before and after of the image:
The image isn’t finished yet. This is just a step you can add in your workflow. We still have to take care of contrast, and we still have to stylise creatively the image. I want to add that this technique doesn’t work well with every image and you have to experiment a bit.
In the final image there’s still something you can enhance. I find there’s a little too much blue in the shadows. I’d like to leave you with this exercise: try to work a little with the mask, maybe mixing the vibrance mask with a luminosity mask, using various blend modes.
Is it worth doing all these steps? Isn’t it sufficient to boost the vibrance slider? Here is a version of the image with just a boost in vibrance. Quite dull in comparison, isn’t it?
The secret of post processing is to always experiment. You’ll do ugly things, but in your experimentations you always come up with something useful to study.
I hope this tutorial is useful. If you find it of any interest, why don’t you share it with your friends?