question about smart filters and sharpening

[DanK]

I generally do my final sharpening at the end of editing because other adjustments, such as texture, affect how much sharpening is needed. I have usually done this by creating a composite layer from everything and applying a sharpening filter to that new layer.

The drawback is obvious: every adjustment prior to that is baked into the layer with sharpening.

Is there any reason not to turn the base layer into a smart object and then add a sharpening filter to that at the end of the process? That would solve the problem I noted and would also make the file smaller.

Thanks.

[Manfred M]

Dan - I've always looked at final sharpening as a "throw-way" that can be replicated with a few keystrokes. If I wanted to, I could even write an action to do everything other than dialing in the amount of sharpening I need and masking out areas when I don't want the output sharpening to apply to. In landscapes this is often the sky and water and in portraiture, I often don't t the skin to be sharpened, especially if I have done some work to soften it.

My usual output workflow is:

1. Resize my image to the final output size / resolution. For printing this is 360ppi and for screen I tend to stick to around 110ppi (which matches my default screen resolution). I will also generally switch to the colour space I plan to output to; sRGB for screen and either AdobeRGB or ProPhoto RGB for printing. Changing the colour space collapses the layers in my setup.

2. I then either stamp (if I have multiple layers) or duplicate the layer if everything has collapsed to a single layer. I change the blending mode of the top layer to luminosity and change the opacity to 80%. I make sure that I am viewing at 100%

3. I then open up the Blending Options and use the Blend If for This Layer and set the lowest value, by sliding the small triangle to around 20. I then <Alt><Click> on the triangle and slide it to around 40. I then go to the highest value to around 240 and <Alt><Click> and move the resulting marker to around 220. The reason for this step is that neither the extremely dark areas or the extremely light areas look right if they are sharpened. We don't want "crunchy" in those areas.

4. I then sharpen the layer to taste. I tend to use USM and others like SmartSharpen, but in the end, either give good results. I will apply a layer mask to the adjustment layer and paint any areas I don't want sharpened with a black brush. No need to be super exact here.

5. I adjust layer opacity a bit (the reason I set it to 80% is that I can up the sharpening a tiny bit if I seem to have undersharpened. If the image looks a bit too crunchy, I can dial the opacity down a touch.

If I use screen output, I collapse the layers and write a JPEG and if I go to print, I just print what I have.

I don't save this working file as it generally takes only a minuted to reproduce all this. As it is size specific, I will generally redo it as I will likely not print the exact same size again.

[DanK]

Manfred,

Thanks. The idea of not sharpening extremes is a good one. I hadn't done this.

Most of this process, however, doesn't fit with my workflow. I print from Lightroom, and I have had very good results with its output sharpening. So my key issue is find a more efficient process for the creative sharpening before that stage.

I rarely use USM for sharpening (as opposed to local contrast). I sharpen primarily with the LR/ACR tools (which have become quite sophisticated), Photoshop smart sharpen, and the old-fashioned high-pass filter.

Dan

[Manfred M]

I had not quite realized that you were looking at in process sharpening, rather than output sharpening.

My workflow probably would not work for you, as I do it at the beginning of my process, not at the end. When I finish with the raw convertor, I open the image in Photoshop, duplicate it and do all my in-process sharpening as my next step, all based on the initial copy of the image. This is one of the few processes that I do that uses destructive workflow and use a layer mask to expose the areas that I sharpen and hide the areas that I don't. When I finish a particular sharpening step, I collapse the layers and move on to the next bit of local / in-process sharpening.

[DanK]

Manfred,

Thanks. Masking is essential for my workflow as well. I sharpen globally much less than I used to. Your influence, in part.

Dan