LAB Saturation

I stumbled upon a new (new to me) way for adjusting colors for my photographs in Photoshop using the LAB color mode. I'm not going to go into the technical aspects of the LAB color space because I honestly don't really understand much of it. There's lots of good information out there on the LAB color space on the Web and Dan Margulis has a book Photoshop LAB Color that I've read some if you really want to know more, especially the why part. There is good evidence that the LAB color space is good for certain types of images and certain types of edits. I figured it would be good to simply check this out and see for myself.


What I explored was adjusting color saturation using the LAB color mode. This is an alternative to other Photoshop techniques such as using Hue/Saturation or Vibrance adjustment layers. You may also use this in place of the same sliders found in Camera Raw. Putting aside theory and the why, I decided to let my eyes be the judge.

I decided to use one of my favorite images from our last trip to Scotland, Slioch upon Loch Maree. I created a new copy of my original RAW file and in Camera Raw I setup my image in my typical workflow which sets up the image with a good histogram and a rather flat or dull appearance that will later be fine-tuned and brought to life from within Photoshop. The idea I use is to start with a very workable image with the proper exposure and both highlight and shadow detail. Camera Raw helps to set things up, Photoshop is where my creativity is exercised.


Once in Photoshop I start with a Levels adjustment layer to set new black and white points. Next I'll use a Curves adjustment layer to add some initial contrast with a very basic S-Curve.








Now I'll select these new layers including the background and create a new, merged layer, using CTRL+ALT+SHIFT+E (yes, I use Windows). This will create a new layer called Layer 1 that will be a combination of the two adjustment layers and the Background.


Next I'll duplicate this layer and select a New document as the destination. Right-click on the layer in the Layers panel and then select Duplicate Layer... to get the dialog to also select the destination.












Now I have a new document with the new merged layer I just created.



I set the color mode to LAB. From the menu, Image -> Mode -> LAB Color.


Next I created a Curves adjustment layer. Notice now that there are different values in the channel drop-down of the Adjustments panel. They are Lightness, a, and b.













I adjusted the three channels, adding a standard S-Curve to the Lightness channel for contrast, then I performed adjustments to the a and b channels by pulling in the curves equally on both sides. Photoshop by default places a 4x4 grid in the Curves panel. To get the 10x10 grid just hold down the ALT key and click anywhere in the grid and you'll get a 10x10 grid. Do it again to revert back to a 4x4 grid if you want.


Variances in the amount of contrast and the adjustments in the a and b channels will alter the amount of the effect. Small adjustments actually have a profound effect on the image. The steeper the curve, the more dramatic the effect. If you don't balance the sides out you'll end up with various color casts. You can certainly get very creative and wild here, however, that's note my point in this exercise. I want to simply increase the saturation and vibrancy of the colors here. My adjustments looked as follows:

From here, I selected both my Curves adjustment layer and the Background layer and then created another new merged layer using CTRL+ALT+SHIFT+E. Then I duplicated this layer back to the original document I started with.


I finished up with a few more adjustments and then brought up the previous version of my image and compared them side-by-side in Photoshop.

So, there are not broad or obvious differences between these two images. The color variations are subtle. Is one better than the other? Perhaps at this time I do find the new image using the LAB saturation method more pleasing than the previous edit. Perhaps I'm just being biased to the new technique. Art's a very subjective thing and your likes and dislikes do change over time. LAB color saturation adjustments will be a new tool in my toolbox. It's certainly not the only method available and it shouldn't be. Art is subjective, remember.

Sharpening for the Web

Here is something that I've known about for some time, put it on my list of things to learn and do, but haven't taken the time to just do it until now...sharpen my images for the Web.

Some of us, me included, don't post SOTC (straight out of the camera), full-size images on the Web. That means you have to resize them using Photoshop, Lightroom, or whatever your image editing software of choice is. Resizing an image also means resampling the image from one dimension to another, say 2832x4256 (12MP) to 683x1024. When you resample an image this way, that is downsampling, you are reinterpreting the pixels and you are losing data! There's all sorts of names for the phenomenon that goes on when you do this, ringing, aliasing, blurring, artifacts, etc. There are also all sorts of algorithms that can be used to resample your images and each has a pretty significant impact on the quality of your image.

I won't go into a a long technical discussion about resampling algorithms and sharpening algorithms. You can read up all about them on the Web if you are interested. Instead i'll just show you a workflow that I'm using now that creates great looking images from your full-sized images for the Web. Perfect for Facebook, Flickr, Tumblr, and the like.

1. After you prepare you image and have completed all your post processing, export or create a full-sized JPEG of your photograph. Then open the JPEG in Photoshop. Here is the original, full-sized JPEG of the image I'll be working with here (Click on the image to see it full-size).

2. Resize your image to your new dimensions. Here I am sizing this to 683x1024 for the Web. Select Bicubic Sharper (Best for reduction) for the resampling algorithm.

3. Now that your image has been resized and resampled, it's time to apply some more sharpening to try to reel in some of the detail that has been lost. For this I use the Unsharpen Mask... Filter. I typically try amounts from 200% to 400% and a radius of 0.3 pixels and leave threshold at 0.













4. Now your image is ready for the Web!


Here's a comparison between the unsharpened image and the sharpened version after resizing the images. Roll-over the image with your mouse to see the sharpened version.

These images are zoomed at 300%. Notice the difference in texture of the football and the sharpening and detail in the face. While there are some visible artifacts in the sharpened version, remember we are viewing this image enlarged at 300%. At 100% you won't be able to see these (see the next image below).

roll-over with your mouse to see the sharpened version

Here are the two final images, again for comparison, now at just 100%. Roll-over with your mouse to see the sharpened version.

Can you spot the artifacts now? I don't think so.

Which one looks better? The sharpened version if you ask me.

This is my nephew btw. I can see much better detail in the football, his face, his hair, his tattoes, his shirt, and his pants. The overall effect isn't obviously noticeable but it results in a better looking image. Isn't that the point? So now you can create great images for the Web and have everyone envious asking "how do you take such great photos?"

roll-over with your mouse to see the sharpened version