A World of Image Sharpening, continued
Of course, while it is easy to discern between the two extremes in frequency, it is more difficult when the frequencies fall in the middle. Generally, radius settings for medium-frequency images will fall between 0.5 and 1.5. While it may not seem possible to automate capture sharpening for different frequency images, the authors provides presets that can be used in Lightroom as starting points. With experience these become more automatic to select. The ultimate challenge comes when your image requires sharpening of both extremes. This is when the masking parameter comes into play. This parameter automatically will create an edge (or surface) mask to apply sharpening to high-frequency edges while masking the lower frequency surfaces from the effect. In difficult cases, or where the artist wishes to emphasize certain parts of the image, creative sharpening comes into play where you make your own masks and paint in sharpening.
4. When Sharpening Halos Are Desired
As discussed earlier, capture sharpening addresses the image softening that results from turning photons into pixels. In a similar manner, output sharpening addresses the loss of detail that results from turning pixels into dots on a piece of paper (in the case of inkjet printing). The primary reason for this is something called dot gain. Ink droplets will tend to spread, or bleed, as they penetrate the paper and dry. The growth of the droplets smears the fine detail creating a less sharp appearance. Therefore, for output sharpening it is normal to compensate with ugly looking halos on the computer monitor (zoomed at 100%) knowing that dot gain will tend to smooth out the apparent over-sharpening. Output sharpening is the easiest to implement since it is applied globally to the image with no layer masking nor preference to shadows or highlights. Consequently, output sharpening has been automated and Bruce Fraser’s PhotoKit Sharpener’s routines are a part of Lightroom’s printing module. In practice, matt papers need slightly more sharpening than gloss papers. You will only see the choices for matt and gloss paper in Lightroom since in the authors’ experiences it is overkill to have more than the two paper choices.
Again, this item brings us back to why multipass sharpening is preferred over single-pass sharpening. If you sharpen an image to look good on a monitor, you will end up having to sharpen it again for output, in this case the print. Furthermore, this highlights the pitfall of using the monitor as an arbiter for output printing. That’s why a print is often called a proof, since it is the standard when gallery display is the final outcome for your image.
5. Sharpening for Computer Display
In the past, a photographer’s portfolio consisted of contact sheets or large prints carried in a portfolio box. Today, personal websites, archival photographic sites, and more recently, social networking sites like Facebook and MySpaceTM, have become the prevalent way for friends and prospective clients to see and sample our images. Therefore, knowing how to sharpen for a computer display is critical. The conventional wisdom for web images is 72 ppi. However, the authors’ point out that it is a fallacy that all displays are 72 ppi. In fact, they provide a table of various display resolutions. Unless you have a monitor running at 1024x768 resolution or less, most contemporary monitors are capable of more than 72 ppi. I think it is fair to say that most photo editors will probably have large, high-resolution LCD displays that are capable of 100 ppi or more. There are two points to glean from this. First, knowing your monitor’s resolution is important to understanding the relationship between the pixels on your display and the final print. If your display is 72 ppi and you are printing at 300 dpi, then a 25% zoom will give you a better idea of the final print sharpness than will another zoom level. Second, know your customer. Would you want your images to look great for the lowest common display, say 800x600 pixels running at 50 ppi resolution, and risk them looking bad to an editor at National Geographic whose 30” display is running at 101 ppi? The choice is yours.
Most of us, myself included until recently, probably downsize our high-resolution master images to 72 ppi, 800 pixel images or smaller, for web use in one fell swoop. We know from the Photoshop Image Size dialog box that the resampling algorithm Bicubic Sharper is “best for reduction” and Bicubic Smoother is “best for enlargement.” But did you know that if you downsample in one step you are using only one algorithm which is not optimal? The authors recommend downsampling in three steps. For the first two steps reduce the width and height (assuming you are constraining proportions) 50% using the Bicubic resampling algorithm. For the final step, size to the desired pixel dimensions using Bicubic Sharper. While you end up with the same number of pixel as in the one-step process, the authors show the three-step process holds more detail in the fine texture. Try it and convince yourself. Have you ever wondered why some people’s images just look sharper on a website? Perhaps it is not because of the camera, lens, or technique they used to make the image, but rather the downsampling/sharpening techniques they used.
The purpose of Real World Image Sharpening is to teach and provide many real world examples showing that a three-step sharpening workflow generates a more optimal image than one sharpening step. I think the authors succeed and I highly recommend the book for those serious about their image quality.
Finally, some parting thoughts about ignorance being an unwillingness to learn. Aspiring amateurs and professional photographers have all felt the impact the internet and digital photography has had on the business side of selling photos. A glut of images, some good, but a lot of it just junk, has flooded the market. Combine that with the business models of stock houses big and small catering to shareholders and the next quarter’s financial results, and you have the perfect storm where downward pricing pressure dominates the industry regardless of image quality/uniqueness. One just has to look at the quality of some of the iReporter images - or for that matter the regular images - CNN uses on their website to be convinced of this. Where does all this lead us? Will photography become so simple, automated, and banal that a monkey will be able to take acceptable photos? Photoshelter® recently offered a seminar titled “Everyone Can Be A Photographer.” Perhaps everyone can, but I don’t think everyone should, become a photographer. An unwillingness to learn by that editor did a disservice to himself and his audience by ignoring information that provides an advantage in image quality to those who learn and practice its techniques. Had such an attitude prevailed when Ansel Adams used dodging and burning to enhance the mood of his black and white masterpieces, we might not have access to the same techniques in Photoshop today. From a practical viewpoint, the more we insist on making photography less of a craft - one that requires skill, technique, study, and experience - the more quality images will lose their value, the less profitable the field of photography will be for professionals, and the less likely photography will be around as a profession in the future. In the end, we all lose as art loses its value and its creativity.
4. When Sharpening Halos Are Desired
As discussed earlier, capture sharpening addresses the image softening that results from turning photons into pixels. In a similar manner, output sharpening addresses the loss of detail that results from turning pixels into dots on a piece of paper (in the case of inkjet printing). The primary reason for this is something called dot gain. Ink droplets will tend to spread, or bleed, as they penetrate the paper and dry. The growth of the droplets smears the fine detail creating a less sharp appearance. Therefore, for output sharpening it is normal to compensate with ugly looking halos on the computer monitor (zoomed at 100%) knowing that dot gain will tend to smooth out the apparent over-sharpening. Output sharpening is the easiest to implement since it is applied globally to the image with no layer masking nor preference to shadows or highlights. Consequently, output sharpening has been automated and Bruce Fraser’s PhotoKit Sharpener’s routines are a part of Lightroom’s printing module. In practice, matt papers need slightly more sharpening than gloss papers. You will only see the choices for matt and gloss paper in Lightroom since in the authors’ experiences it is overkill to have more than the two paper choices.
Again, this item brings us back to why multipass sharpening is preferred over single-pass sharpening. If you sharpen an image to look good on a monitor, you will end up having to sharpen it again for output, in this case the print. Furthermore, this highlights the pitfall of using the monitor as an arbiter for output printing. That’s why a print is often called a proof, since it is the standard when gallery display is the final outcome for your image.
5. Sharpening for Computer Display
In the past, a photographer’s portfolio consisted of contact sheets or large prints carried in a portfolio box. Today, personal websites, archival photographic sites, and more recently, social networking sites like Facebook and MySpaceTM, have become the prevalent way for friends and prospective clients to see and sample our images. Therefore, knowing how to sharpen for a computer display is critical. The conventional wisdom for web images is 72 ppi. However, the authors’ point out that it is a fallacy that all displays are 72 ppi. In fact, they provide a table of various display resolutions. Unless you have a monitor running at 1024x768 resolution or less, most contemporary monitors are capable of more than 72 ppi. I think it is fair to say that most photo editors will probably have large, high-resolution LCD displays that are capable of 100 ppi or more. There are two points to glean from this. First, knowing your monitor’s resolution is important to understanding the relationship between the pixels on your display and the final print. If your display is 72 ppi and you are printing at 300 dpi, then a 25% zoom will give you a better idea of the final print sharpness than will another zoom level. Second, know your customer. Would you want your images to look great for the lowest common display, say 800x600 pixels running at 50 ppi resolution, and risk them looking bad to an editor at National Geographic whose 30” display is running at 101 ppi? The choice is yours.
Most of us, myself included until recently, probably downsize our high-resolution master images to 72 ppi, 800 pixel images or smaller, for web use in one fell swoop. We know from the Photoshop Image Size dialog box that the resampling algorithm Bicubic Sharper is “best for reduction” and Bicubic Smoother is “best for enlargement.” But did you know that if you downsample in one step you are using only one algorithm which is not optimal? The authors recommend downsampling in three steps. For the first two steps reduce the width and height (assuming you are constraining proportions) 50% using the Bicubic resampling algorithm. For the final step, size to the desired pixel dimensions using Bicubic Sharper. While you end up with the same number of pixel as in the one-step process, the authors show the three-step process holds more detail in the fine texture. Try it and convince yourself. Have you ever wondered why some people’s images just look sharper on a website? Perhaps it is not because of the camera, lens, or technique they used to make the image, but rather the downsampling/sharpening techniques they used.
The purpose of Real World Image Sharpening is to teach and provide many real world examples showing that a three-step sharpening workflow generates a more optimal image than one sharpening step. I think the authors succeed and I highly recommend the book for those serious about their image quality.
Finally, some parting thoughts about ignorance being an unwillingness to learn. Aspiring amateurs and professional photographers have all felt the impact the internet and digital photography has had on the business side of selling photos. A glut of images, some good, but a lot of it just junk, has flooded the market. Combine that with the business models of stock houses big and small catering to shareholders and the next quarter’s financial results, and you have the perfect storm where downward pricing pressure dominates the industry regardless of image quality/uniqueness. One just has to look at the quality of some of the iReporter images - or for that matter the regular images - CNN uses on their website to be convinced of this. Where does all this lead us? Will photography become so simple, automated, and banal that a monkey will be able to take acceptable photos? Photoshelter® recently offered a seminar titled “Everyone Can Be A Photographer.” Perhaps everyone can, but I don’t think everyone should, become a photographer. An unwillingness to learn by that editor did a disservice to himself and his audience by ignoring information that provides an advantage in image quality to those who learn and practice its techniques. Had such an attitude prevailed when Ansel Adams used dodging and burning to enhance the mood of his black and white masterpieces, we might not have access to the same techniques in Photoshop today. From a practical viewpoint, the more we insist on making photography less of a craft - one that requires skill, technique, study, and experience - the more quality images will lose their value, the less profitable the field of photography will be for professionals, and the less likely photography will be around as a profession in the future. In the end, we all lose as art loses its value and its creativity.