This week’s Friday (not quite Friday) Fundamentals features a video from Deke’s latest Photoshop course, Photoshop CC One-on-One: Intermediate. “Gauging the Best Sharpening Settings” is an unlocked video I’ve placed here for your Labor Day Weekend enjoyment. (Disclaimer: I had to hit the road yesterday to beat the traffic to my own holiday enjoyment, thus the “Friday Fundamentals” posting on Saturday. At least you know I thought about you dekeLlies once I secured my own vacation spot.)
Sharpening is not about focus, but rather about editing an image to make its details appear more sharply defined. As Deke points out earlier in the course:
“It’s important to note that we’re not talking about focus. Photoshop can not reach back into your camera and adjust the optical focus of your lens element. Nor can it invent detail that does not exist. In other words, if an image was shot blurry, it will remain blurry. What Photoshop can do is take a well-focused image and make every detail appear crystal clear.”
Read on to see how it works:
Your sharpening needs will depend on your photo’s ultimate destination, namely screen or print. For instance, the top image in the graphic below shows the nicely focused butterfly prior to any sharpening applied. The middle image is sharpened for a standard computer screen, and the final image (which may look oversharpened on screen) has a bit of extra sharpening applied so that it will look best in print.
Sharpening can be applied in a number of ways. In the video, Deke uses the go-to sharpening tool Smart Sharpen, which is applied via the Filter menu. Deke’s strategy for determining where the sharpening sweet spot lies starts with cranking the Amount to 5 percent and starting with the Radius value at 1.
Then he lowers the Radius value while observing the preview window in the (newly enlarged) Smart Sharpen dialog box. Once you set the Radius where it needs to be, you can lower the Amount to an appropriate setting.
From that ideal screen setting, Deke’s got a formula for increasing the sharpening from there for print. He’s also got some advice within for adjusting this recipe if you’re using a high definition screen.
Stay sharp dekeOgraphers!
Smarty sharpie
artjobb
Deke
I feel like maybe I am being dense but I’ll ask anyway. Hopefully you feel the same as me that there are no dumb questions only dumb answers.
You apply Smart Sharpen as if you are doing it for screen. 1pixel on radius, 500% Amount to start. Adjust to the setup that looks good for screen. Then for print;
Multiply Radius Value x 3,
ADD 50% to your Amount setting . (Not multiply by 50% then add to the starting Amount.)
Correct?
You are a great instructor.
Thanks
Values for print
I think some of your math is incorrect there. You describe how to change the values from screen sharpening to settings suitable for print. But “adding 50%” makes no sense algebraically. If I determine the optimal value is 150% and I change it to 200%, that is a considerably higher increase in sharpening than if I start at 450% and go to 500.
Also I am skeptical about the tripling of the radius. That’s not how screening works in color separation. If you send a 72DPI image to an imagesetter, and the line screen is set to a default of say 150 line screen (a common setting in a high quality color sep) then your image DPI should be a minimum of 2x the LPI, so you’d use 300DPI. But if you imageset the 72 DPI image (suitable for ~30 line screen, less quality than a good Postscript laser printer) then you’re going to end up with a huge “halo” (as you put it, I’d prefer to call it an “enhanced edge contrast effect”) that is rendered with way more detail than is necessary. In fact, it would be 5x the sharpening required.
If you are sharpening for print, you have to work from the imagesetting LPI down, not from the image rez up. I have a formula for that around here somewhere, I haven’t worked on imagesetters in a while. But my recollection is that the smallest noticeable difference in an image is between 0.5 and 1.5 lines, so you’d try to set the sharpening just below what would make an obvious halo, somewhere in that range. Since the optimal image rez is 2x the linescreen, 2 pixels render as 1 line, so you would need to keep your radius between 1 to 3 pixels at 300DPI image rez. If you’re setting a radius of 1 in a 72DPI image, your radius is about 4 at 300DPI. You’re basically enhancing the edges of pixels instead of the image itself. At my old prepress studio, we used to call this effect “pixels as big as your fist.”
Edit: oops, replying to Deke here, not the commenter. The upthread comment is pointing out the same thing as me.