Tip--PhotoShop Sharpening Technique

Reno gave me a great tip for sharpening in Photoshop that I thought I'd pass on. It is very useful for Web images, as well as for printing) because it virtually eliminates halos. The reason being that you are increasing the contrast between pixels, rather than colors. The procedure follows...

1. Open your sized RGB image.
2. Choose Layer>Flatten
3. Choose Image>Mode>Lab Color
4. In the Channels Pallet, choose the "Lightness" channel.

When you choose the 'Lightness' channel your image will revert to a very light B&W image. Don't worry.

5. Choose View>Actual Pixels
6. Choose Filter>Unsharp Mask

Settings for USM: Ammount=85%, Radius=1, Threshold=4. Preview--Checked

If you need to apply more USM, press Ctrl-F to apply USM again. Repeat this step as necessary.

7. Choose Image>Mode>RGB to switch back to RGB color.

The 'flattening' of the image in step 2 is important; if you switch from RGB to Lab without doing so, some layers will be discarded.

You can use Ctrl-F in RGB mode for further sharpening if you need to, but this may cause halos.

This is a great technique, and it really takes less time to accomplish than to explain. Give it a try!

Happy New Year everyone!

-PB

Preston,

Thanks for posting that. I've been using basically the same method. Do you know what "view actual pixels" buys you? I'm not sure what that step does. Also, thru trial and error, I use values of 2-300%, radius .2-.3 and threshold 0. I really don't have a good understanding of what is happening with USM regarding the threshold. My guess is that if you have a threshold of say, 4, that pixels of size 4 or less won't get touched.

Also, does anyone have an opinion about using the 50% resizing method when working with large files (200mb 4x5 scanned files for example.) In other words, I read somewhere that you should incrementally downsize a large file, applying some or small amounts of sharpening after resizing.

And for all the cases above, I'm just talking about sharpening and resizing for the web.

Thanks, Happy New Year everyone,
Lon

Lon,

The settings I got from Reno work very well for printing, but they are not 'cast in stone'. I've always found that a standard setting for USM is not always optimal: I might have to experiment with other settings to get the effect I want because there are so many variables.

When you view an image at any 'zoom factor', the on-screen image is, of course, resized; which can introduce (virtual) noise. By setting the view to 'actual pixels' you are looking at the image as it truly exists, making it easier to see the effect of sharpening. I have always done my sharpening at 100%. I then switch back to 'Fit on Screen' to see the overall effect.

Regarding incremental resizing: I have heard of this as well. I would like to know how well it works. Is there less noise? What exactly does incremental resizing do for the image?

And while we're at it: Which algorithm is best: Nearest Neighbor; Bilinear; Bicubic, Bicubic Smoother; Bicubic Sharper? Does the method used depend upon the specific image? I've always found this to be confusing. I usually use Bicubic, but if another method works better, I'm all over it.

I think I will post this question over at Digital Imaging and see if one of gurus can help us out.

Happy New Year to you!

-PB

All my readings say bicubic sharper for downsizing and bicubic smoother for upsizing, which is what I do. In my way of thinking, bicubic sharper would be used for upsizing and smoother for downsizing. I guess that's why I'm at home on the computer instead of in Seattle working for Adobe

Use bicubic smoother for upsizing, it is a modified version on the bicubic algorithm designed to behave as if we upsized in steps.
Use bicubic sharper for downsizing.

The theory behind the incremental resizing is that by taking smaller steps, the algorithms have an easier job generating the new pixeles and therefore generate less jagged edges and smoother transitions.
The downsizing version also included a small sharpening step between each reduction in size, this was designed to help preserve detail that might otherwise be lost as the calculations discarded information.

regarding the view actual pixels... this allows you to see exactly what changes are being made to the image one pixel on the image to one pixel on the screen.

Regarding the threshold... this determines the color/luminosity values required between pixels for the edge between then to be contrasted (sharpened). If you use a low threshold you increment noise in otherwise flat areas, having a higher thresholds will only contras areas where a bit of contras already exists.
The radius setting determines how wide the contrast are is, a low setting might not sharpen as much as we wish, a high setting will cause halos. By experience I can say that this setting depends on the resolution and display size of the image AND in the type of detail found in an image.

USM and other sharpening tools just up the contrast between areas, creating the impresion of detail but the actualy destroy real image data so go easy on the originals and sharpen your images at the end of your workflow and acording to the final output media, and keep an unsharpened version as it will hold more image information.

Curiously, using what might seem as absurd amount/radius values can enhance not sharpens but overall image contrast (try amount 15%, radius 250 pix, threshold 0 pix).

incremental uprezing makes no difference nowadays from my own comparisons. Reno's technique sounds similar to the sharpen only edges action that I received from my printing instructor. Most of the time I find that it does a great job.

All good info. Thanks Preston and Enrique.
Enrique I think I might have to make it to club to meet you... it'd be a shame not to meet in person. Are you going to be there next Thursday?

This is a great post and I have read it a few times now. I have worked with most of the ideas and also use other methods with various degree of success and outcomes. I was instructed a long time ago to use the 500% and .2 setting with 0 threshold setting. I did this blindly with a certain degree of success. This was always done at RGB without changing anything else. I did find that I liked the ability to toggle back and forth between the sharpened image and the previous un-sharpened with one stroke. Although sticking with a similar setting I further enchanced my ability to fine tune by selecting the image at 100% pixel ratio, pulling up the unsharp mask setting, and then fine tuning the image with the use of the radius setting with the use of the finger wheel on the mouse. I always use the mouse finger wheel for making adjustments as it helps me see the adjustments easier. I guess where I'm stuck overall, is what the best settings in the mask should be, and the need for going to the non-color aspect and then back to the RGB mode. By using that method I do not find away to instant toggle back and forth between the finished image and the original starting point. Any ideas on what is best overall mask setting per se? I do agree with Enrigue about the final touch is the sharpening, that was also drilled into me a long time ago. Thanks for any input here.

I have worked with the method as Preston has posted and it works fine for me. Thanks for this additional idea on overall sharpening technique. I have started using it over my previous approach and like this better.

I'm glad it's working out for you Paul! I really like the technique and the results are excellent. I hate halos, so having a way to prevent them is a big boon, indeed!

-PB