The idea of 3-step sharpening has been around for a while but I meet a lot of photographers new to the hobby who aren’t familiar, so I am writing a brief summary.
This idea was brought to widespread attention by Bruce Fraser in the early 2000’s. He pointed out that there are 3 general reasons for sharpening a photograph.
The first is input/capture sharpening. This is needed to undo the effects of the digital capture process that turns the continuous gradations of the analog world into a discreet grid of pixels. The process will introduce a degree of softness into the image. He recommended an edge sharpening approach for this, such as with the high-pass filter.
Next is creative sharpening. We might want to add sharpness to help with the interpretation we want for an image. Often this will be in selected areas, not on the image as a whole.
Finally, we have output sharpening. Output is done at different sizes Resizing will add unsharpness to an image and needs to be corrected later. Also, the spread of ink on paper, no matter how small, will cause a loss of sharpness and can be compensated for with extra sharpening before printing.
Each of these types of sharpening serves different purposes and has different requirements. It doesn’t make sense to do them all at once. If you did, the file would only be suited to one type of output and you would have to create a whole new file if you needed a different type of output.
By splitting the sharpening into 3 stages, you gain some advantages. First, by compensating for the capture right away, you can more accurately judge the other adjustments you might need, especially for contrast. Early on, I followed the common “sharpen last” recommendations ad was always a bit frustrated because it would undo the contrast adjustments I had so carefully built in to the image.
Then, after you finish with any creative sharpening, you now have a master file that is optimized the way you want. If you need different output types, you can start from a copy of this file without having to repeat the previous work. This saves time and ensures consistency among the output versions.
This sounds great, you may say, but won’t all of this sharpening ruin the image? We’ve all been warned against over-sharpening and perhaps we have seen the unsightly results. Back in 2009, Glenn E. Mitchell posted a chapter from his ebook on sharpening that provides the answer. Looking at some of the sharpening algorithms, he noticed that in multiple applications, the desirable effects of sharpening accumulate faster than the artifacts. The secret is to keep the amount or strength low on each pass.
I have a demo image to show this. It is an image of backlit tree flowers. One has been sharpened twice at an amount of 75, the other was sharpened 15 times but the amount was only 10. So they both work out to an amount of 150. The effect is subtle. I suggest to open each image in a separate tab and then toggle the tabs. You will see a slightly better sharpness in the 15x image. I used a batch process to do this.
The main point here is not to suggest specific steps but to show why it is an advantage to do 3 steps and to show that it is possible if you keep the amounts low each time. I might look at some of these in depth in the future but you can look up these techniques and find many tutorials online already.
Sadly, we lost Bruce Fraser in 2006 and his articles seem to be disappearing. His associate Jeff Shewe has issued a new edition of the book Real World Image Sharpening with Adobe Photoshop, Camera Raw, and Lightroom, Second Edition, where you can find these ideas.
Two of his articles are still at CreativePro.com:
Out of Gamut: (Almost) Everything You Wanted to Know about Sharpening in Photoshop but Were Afraid to Ask
Here is the last known link for another article. Search for it at the Internet Wayback Machine
Glenn Mitchell’s site no longer seems to be active and I have no information about him. Here is a link to try in the wayback machine:
Unfortunately, the link to the pdf download was not preserved.