Topic: About the sharpening actions

[Alex314]

I have studied the 3 actions Mike posted and I tried them out on a digital image of moderate sharpness.
Though the actions work fine, my impression is that they all increase the noise in the images noticeably in more or less smooth areas. A finding that I did earlier using the Photoshop techniques as provided by Adobe.

For that reason I have always hesitated to use sharpening. My (theoretical) idea is that you cannot get what isn't there. I also disabled the camera sharpening - I use raw images and convert them in PS.

Am I on the wrong path?

Alex

[curvemeister]

Hi Alex,

Yes, sharpening does generally increase noise, particularly in noisy areas like the shadows.  In some cases, I will use a shadow or luminance mask to control this problem, effectively sharpening only the less noisy areas.  I've been impressed by the 3rd party sharpening utilities out there, including FocusMagic, as well as those that remove noise, including Noise Ninja, and Neat Image (which has a very effective stand-alone demo version).

[Alex314]

Thanks for your reply.
I have used Neat Image without really knowing how to set up the program. I find it pretty confusing, but I have seen already some good results. I guess I have to study the manual better!
Alex

[lbenac]

I've been impressed by the 3rd party sharpening utilities out there, including FocusMagic.

Hello Mike,

I use Focus Fixer (similar to Focus Magic) routinely right after conversion to TIF. I do not use any capture sharpening in the RAW converter. To me this produces better results than anything else that I could try (that is not exhaustive). The way I understand deconvolution, it does not operate based on color contrast like other methods but more akin to a lens correction algorithm.
I tried using deconvolution on a copy of the background layer to compare blending mode of Normal and Luminosity with no difference which would make sense based on the above. What is your experience with deconvolution, any tricks :-)
Another question.
Based on the same assumption I apply deconvolution before a trip to CM simply because it is more convenient (actions). Do you have any technical opinion on that?


Cheers,

Luc

[curvemeister]

Hi Luc,

No opinion at all.  Since sharpening and color correction are fairly separate functions, the order is not important, but there are some considerations that may apply for certain images.

One is that it's generally better to make the biggest changes first.  Regarding noise, If I were recovering a severey underexposed image, I'd be concerned that any noise that appeared in the shadows would be best dealt with after the curves (and other) operations.

With sharpening, the considerations are more subtle.  I doubt that the order would matter, but I would not be surprised if you, or someone else, came up with an example to the contrary, where a severely out of focus original gave a better result depending on whether the sharpening step was done first or last.

[Greg Groess]

I always thought you did the sharpen last so that you did not multiply any problems by having to re-sharpen?

Greg

[lbenac]

I always thought you did the sharpen last so that you did not multiply any problems by having to re-sharpen?

Greg

I think that it holds true for any type of sharpening based on contrast but deconvolution the way I understand it relies on a different principle more akin to lens geometrical correction. Also if you follow Bruce Fraser theory on sharpening, he advocates doing the capture sharpening first then the output sharpening last.
So for me it is better to do the deconvolution first because I am lazy and it fits easily on a droplet at the beginning of my work flow.
What I want to compare is what Dan Margulis suggests - USM on the L* channel as an addition to light deconvolution.

Sorry if I derailed somewhat the forum from color to sharpening but things often affects one another :-)

Cheers,

Luc

[Greg Groess]

not at all...
for me all aspects are important...

Greg