After the series of posts related to sharpening in Photoshop, one reader commented and asked whether split sharpening or edge-mask sharpening would be a better approach. I indicated in my reply that edge-mask is just that, a mask created from the edge information in the image applied to a sharpened layer and that there was no reason one could not combine both. This brief post will present a new action, essentially a modified Smart Object Split Sharpening which creates an edge mask before the sharpening and applies that mask to the smart filters.
When will an edge mask be really useful? If the image has large low-frequency areas, like a portrait of a child or a woman, or some landscape with a large sky area with clouds, edge mask may help protect those low-frequency areas from the effects of sharpening. In this post, you will see some sample sharpened images which are often pushed over the sensible levels to show the effect of an edge mask. When the sharpening parameters are applied with restraint the importance of the edge mask diminishes. I will note the images as carefully as I can so that you can see the comparative evaluation of various before-after images.
Before going any further you may want to download the new Ekin Sharpen action set which will include all the previous ones and one new one, Split USM SO With Edge Mask. For instructions on where to copy it, refer to the previous post on sharpening.
When you are ready to apply sharpening, open the Actions panel, highlight Sharpen: Split USM SO With Edge Mask, click on the play button at the bottom of the panel and wait for it to run its course. In the end, you will have a smart object with two Unsharp Mask filters applied, lower is for lightened edges and the upper is the darkened edges. You can double-click on each and adjust the parameters from their default levels of Amount=10, Radius=1, Threshold=0 to suit the needs of your image.
Let us take a look at some pairs of images. You can use the slider bar right and left to see before or after states of each pair. What these states represent are clearly indicated above each image. You should note the following:
- The image is a 100% crop from a digital capture
- The parts that need sharpening are the eye and eyelashes, the eyebrow, and the hair
- The part that should be protected is the skin
- When a reasonable amount of sharpening applied the difference between edge-masked or with no mask is negligible
- When aggressive sharpening is needed (which is not in this case but presented to show the results) edge masking has a beneficial effect
- I applied the same parameters to the lighten and darken components in the smart filter to make comparisons easier
To start with, let me present a pair A=300, R=1 T=0. The Before-image is without the edge mask, and the After-image is with the edge mask applied. Keep in mind that even with the edge mask applied, this is probably too much sharpening. That said, the sharpening on the eyelashes remains essentially the same with or without the edge mask while the skin sharpening is suppressed with the edge mask applied (after state).
Now let us take a look at another pair where A=200, R=0.5, T=0. Again, the before state represents without and the after state represents with the edge mask applied. You will note that the skin is not overly sharpened even without the edge mask applied. However, if the skin area showed some noise, sharpening would make it more visible and the edge mask would be quite beneficial.
The final pair will show the image with no sharpening applied and a reasonable level of sharpening applied without the edge mask, and with the parameters, A=200, R=0.5, T=0. Even though the radius is kept small so as to make sharpening focus on fine detail like the eyelashes, the effect on the skin is minimal.
So, you be the judge whether you would like to add an edge mask to your sharpening or not.