If there is one thing about processing I am fascinated with more than Dodge and Burn coupled with Luminosity Selections it is sharpening.
Ideally sharpening should be done during your Raw Processing instead of say in Gimp, however, I find often my particular editing approaches tend to either Ruin the Raw sharpening (Orton type effects) or Produce a halo (Dodge/Burn and or Blending). So if I am going to be doing certain things to an image I like to sharpen inside Gimp.
Sharpening this way can have some nasty effect like increasing noise as well as producing halos more so then if you are doing it in the raw stage. When I was using Photoshop there is a method of Sharpening called “Smart Sharpening”. This method of sharpening really is not smart just much more targeted in that it ignores smooth areas as well as noise.
I always knew it gave me the best results but I did not know why. As a new Gimp convert I missed this “Smart Sharpening” so I decided to learn how it works. Through my research I found that this method sharpens based off of Luminance combined with Edge Detection and Blur.
To implement the “Smart Sharpening” I did a few things slightly differently than what my research provided. Instead of Decomposing the image into Lab color space I chose to go a route of using specialized Luminosity mask to avoid the color conversion.
This is not meant to be a tutorial if one is wanted let me know I might be able to write one but the general steps are.
- Have your image processed, cropped/scaled etc… Save this sharpening for the end.
- Create a Lights 1 Luminosity Channel.
- Create a New Layer from Visible (if no layers a duplicate layer works too)
- Apply the Luminosity Channel to the image as a mask.
- Select the mask then apply Edge Detection and a Gaussian Blur.
(you could refine the edge detection before and after the blur with levels if needed)
- Select the Layer not the mask and apply a unsharp mask to the image.
I am not positive if this is the best way to approach this but it works quite nicely.
Here is a Comparison card I made with one of my images using 3 methods. No sharpening, a global unsharp mask, and the edge detect smart sharpen. The image contains a Orton effect and no base sharpening in the raw and my camera has a low pass filter so No sharpening might be slightly jarring
The Unsharp Global and the Edge Detect Smart Sharpen both have the same amount of sharpening setting so the comparison is equal in that respect.
Note how in the Unsharp the water has some noise being exposed as well as it being quite sharpened where in the Edge Detect the water is ignored.
Maybe we can have a little discussion about sharpening. I am no technical expert on the subject by any means but it would be interesting to hear about what methods people like to use and why they choose to apply that type and why some methods are better then others.