Don Lacy
02-04-2014, 01:10 PM
In the maintaining image detail thread the discussion turned its focus to sharpening specifically when to sharpen and what methods produced the best results. Since this is a subject that can cause some confusion especially among members learning digital photography I was hopeful the thread would be a good learning opportunity but unfortunately that did not happen and which happens to often the participants started talking past each other and the thread was closed with many question left unanswered or unresolved. So I decided to start a thread outlining my workflow and what I did learn from the previous thread to hopefully give the members a starting point on developing their on sharpening workflow. Now I want to state upfront that I do not claim my workflow is the best or only way and since it uses layers and mask it is PS dependent. My sharpening workflow is broken down into three steps capture sharpening, creative sharpening and output sharpening which is pretty standard and the workflow advocated and most often taught by most people who make they're living teaching this stuff.
The first step in my workflow is capture sharpening which is usually done in the Raw converter or my first step in PS after converting the file, the goal of capture sharpening is to reverse the softness caused by the AA filter in front of the sensor and the process of turning the image into digital form . I find the default settings in ACR work really well for most of my images and on most of my images I will use the masking slider to selectively apply the effect to the areas I want sharpen. If I am using another Raw converter that does not allow me to mask out areas I will turn the sharpening off and apply my capture sharpening in PS on a Layer mask using either PhotoKit sharpening plugin or Niks Sharpening Pro plugin again I will only apply it to the subject or areas on the subject I want sharpen. I avoid sharpening the BG, OOF areas, and areas of solid color. When the image is adjusted to my liking I will size it for its intended purpose for the web it will be downsized and for printing it will most likely be uprez to it's printing size after the image is sized for its intended purpose I will sharpen it for its finale output then apply any creative sharpening if needed. For images sized for the web I will duplicate the BG layer and using a selection and mask apply smart sharpening again to the subject only with the following settings in basic mode, remove lens blur, and the more accurate box checked, the amount and radius depends on the subject a good starting point for avian subject is amount 50-100 and a radius between .03-.05 for landscapes I will usually have several layers and mask with different amounts and radius. When I am happy with the overall sharpness of the image I might add a touch more sharpening to certain parts of the image to make them standout to the viewer, like the eyes of birds and mammals the wings of dragonflies a specific rock in a landscape. For images that are to be printed I do the creative sharpening step first and to be honest I cheat and use a sharpening plugin since there are so many variables and I do not do a lot of printing I will use PhotoKit which allows me to choose printing method, image resolution, and paper choice which all impacts the amount of sharpening needed.
Some helpful hints:
Do not sharpen the BG, out of focus areas, and areas of solid color since there is no detail in these areas you will only be sharpening noise and enhancing it ( really important to any 7D users or other high MP cameras)
Try to keep as much of you're sharpening on layers so you can control both areas effected and strength with the opacity slider
When downsizing I feel you get a better image from using the Bicubic method in PS
Any image that is resized will need additional sharpening to offset the effects of resizing
To help avoid halos change the blending mode to luminosity on you're sharpening layer
Also to help avoid halos use a selection of the subject and feather it slightly
You can also use a mask and a soft edge brush to paint in the sharpening effect quick and dirty and works well on landscapes
While I do not use USM to sharpen it applies a nice tonal contrast at the following settings amount 10, radius 60, threshold 0. change the radius setting to change the effect ( which begs the question is it really a sharpening filter or a contrast filter :S3:)
Use creative sharpening to control where your viewers eyes travel within the frame, a viewers eye will always travel to the brightest and sharpest areas of a frame, one of the reasons I never sharpen the perch or ground surrounding my subjects.
Here is a link to a great series of articles on sharpening that go into much greater detail then what I have outline here covering everything from high pass sharpening to sharpening in the LAB color space . http://www.ronbigelow.com/articles/sharpen1/sharpen1.htm
Also a nice explanation of USM here if you read the bottom section it goes into deconvolution also http://en.wikipedia.org/wiki/Unsharp_masking
What I learned form the other thread:
That sharpening your image before resizing will result in a shaper image at its finale size but it will still need additional sharpening ( don't take my word for it do what I did and try it your self you do not have to be an expert to see with your own eyes)
That Richardson-Lucy deconvolution has a lot of potential but does not readily fit into my workflow so i will keep researching it. (all my research so far shows it being used on star images)
There is no miracle fixed to an unsharp image so still your best bet is to get it right in camera
Those still interested in Richardson-Lucy can download these images sharpen with that method and make your own conclusions
https://www.dropbox.com/s/s29oscxf8ge9m0w/bird_52B7483-results2-1200s%2B12.tif.tiff
https://www.dropbox.com/s/p6tvybw8uu640or/bird_52B7483-results2.tif.tiff
The first step in my workflow is capture sharpening which is usually done in the Raw converter or my first step in PS after converting the file, the goal of capture sharpening is to reverse the softness caused by the AA filter in front of the sensor and the process of turning the image into digital form . I find the default settings in ACR work really well for most of my images and on most of my images I will use the masking slider to selectively apply the effect to the areas I want sharpen. If I am using another Raw converter that does not allow me to mask out areas I will turn the sharpening off and apply my capture sharpening in PS on a Layer mask using either PhotoKit sharpening plugin or Niks Sharpening Pro plugin again I will only apply it to the subject or areas on the subject I want sharpen. I avoid sharpening the BG, OOF areas, and areas of solid color. When the image is adjusted to my liking I will size it for its intended purpose for the web it will be downsized and for printing it will most likely be uprez to it's printing size after the image is sized for its intended purpose I will sharpen it for its finale output then apply any creative sharpening if needed. For images sized for the web I will duplicate the BG layer and using a selection and mask apply smart sharpening again to the subject only with the following settings in basic mode, remove lens blur, and the more accurate box checked, the amount and radius depends on the subject a good starting point for avian subject is amount 50-100 and a radius between .03-.05 for landscapes I will usually have several layers and mask with different amounts and radius. When I am happy with the overall sharpness of the image I might add a touch more sharpening to certain parts of the image to make them standout to the viewer, like the eyes of birds and mammals the wings of dragonflies a specific rock in a landscape. For images that are to be printed I do the creative sharpening step first and to be honest I cheat and use a sharpening plugin since there are so many variables and I do not do a lot of printing I will use PhotoKit which allows me to choose printing method, image resolution, and paper choice which all impacts the amount of sharpening needed.
Some helpful hints:
Do not sharpen the BG, out of focus areas, and areas of solid color since there is no detail in these areas you will only be sharpening noise and enhancing it ( really important to any 7D users or other high MP cameras)
Try to keep as much of you're sharpening on layers so you can control both areas effected and strength with the opacity slider
When downsizing I feel you get a better image from using the Bicubic method in PS
Any image that is resized will need additional sharpening to offset the effects of resizing
To help avoid halos change the blending mode to luminosity on you're sharpening layer
Also to help avoid halos use a selection of the subject and feather it slightly
You can also use a mask and a soft edge brush to paint in the sharpening effect quick and dirty and works well on landscapes
While I do not use USM to sharpen it applies a nice tonal contrast at the following settings amount 10, radius 60, threshold 0. change the radius setting to change the effect ( which begs the question is it really a sharpening filter or a contrast filter :S3:)
Use creative sharpening to control where your viewers eyes travel within the frame, a viewers eye will always travel to the brightest and sharpest areas of a frame, one of the reasons I never sharpen the perch or ground surrounding my subjects.
Here is a link to a great series of articles on sharpening that go into much greater detail then what I have outline here covering everything from high pass sharpening to sharpening in the LAB color space . http://www.ronbigelow.com/articles/sharpen1/sharpen1.htm
Also a nice explanation of USM here if you read the bottom section it goes into deconvolution also http://en.wikipedia.org/wiki/Unsharp_masking
What I learned form the other thread:
That sharpening your image before resizing will result in a shaper image at its finale size but it will still need additional sharpening ( don't take my word for it do what I did and try it your self you do not have to be an expert to see with your own eyes)
That Richardson-Lucy deconvolution has a lot of potential but does not readily fit into my workflow so i will keep researching it. (all my research so far shows it being used on star images)
There is no miracle fixed to an unsharp image so still your best bet is to get it right in camera
Those still interested in Richardson-Lucy can download these images sharpen with that method and make your own conclusions
https://www.dropbox.com/s/s29oscxf8ge9m0w/bird_52B7483-results2-1200s%2B12.tif.tiff
https://www.dropbox.com/s/p6tvybw8uu640or/bird_52B7483-results2.tif.tiff