Dan Margulis Applied Color Theory - Sharpening Workflow

From: "Cavanagh, Ken", INTERNET:ken_cavanagh@mcgraw-hill.com
Date: Mon, Apr 23, 2001, 11:28 AM
RE: [colortheory] Sharpening workflow

I have several digital images I am preparing for reproduction...I am converting, color correcting and sharpening at 100% of their original size (12-20mb @300dpi).

I am not scaling to the final image size because either it has not yet been defined or they will be used as various sizes. Though the smallest size may be as much as a 30% reduction of the originals. The size will be determined later by an outside design firm and scaled by the prepress vendor.

My questions are:
Will sharpening first and scaling later, not ideal but necessary in this situation, cause an oversharpened "look" or other artifacts in the images...especially at the smaller sizes?

Could there be a difference in the quality of how the sharpening is affected when scaled if it was done in photoshop, quark or other program?

I've scaled a few of the larger images at 30% via photoshop and have not noticed anything too objectionable on the monitor. The subject is of kids...full body shots and portraits.

Thanks

Ken Cavanagh
Photographer
Macmillan/McGraw-Hill
New York, NY

From: Dan Margulis, INTERNET:76270.1033@compuserve.com
Date: Tue, Apr 24, 2001, 9:53 AM
RE: [colortheory] Sharpening workflow

Ken writes, >>I have several digital images I am preparing for reproduction...I am converting, color correcting and sharpening at 100% of their original size (12-20mb @300dpi). I am not scaling to the final image size because either it has not yet been defined or they will be used as various sizes. Though the smallest size may be as much as a 30% reduction of the originals. My questions are: Will sharpening first and scaling later, not ideal but necessary in this situation, cause an oversharpened "look" or other artifacts in the images...especially at the smaller sizes?>>

The opposite is true. Downsizing will tend to soften the impact of the sharpening. Upsizing will tend to harshen it and that is where one can get into trouble.

If a file is going to be downsized significantly, I actually prefer to sharpen it at the larger size--the results seem to be more accurate.

Obviously, it's a lot better if one knows exactly what the output size is going to be.

>>Could there be a difference in the quality of how the sharpening is
affected when scaled if it was done in photoshop, quark or other program?>>

This depends on the skill of the user and the program's capabilities. Photoshop's execution of the Radius and Threshold options is nice. It would be better if they had a lightening/darkening option. Other programs are better in some ways and worse in others.

Dan Margulis

From: Terry Wyse, INTERNET:terry@mail.allsystems.com
Date: Tue, Apr 24, 2001, 11:40 AM
RE: Re: [colortheory] Sharpening workflow

on 4/24/01 10:58 AM, Cavanagh, Ken wrote:

> Working with images that are already sharpened, but need to be
> resized, I was curious if I should expect a difference between the way quark
> and photoshop handle this.
> On a dye-sub proof, I compared the scaling of an image in quark as
> opposed to scaling in photoshop and dropping it into quark. The quark(4.1)
> version showed a moire pattern whereas the ps(5.5) version was clean.
> Should I expect that moire to show in print?

Make sure you check the "Full Resolution TIFF" box when outputting from Quark or it will resample during the creation of the PS which may be further resampled during RIPing.

Terry
_____________________________
Terence L. Wyse
PrePress Systems Specialist
All Systems Integration, Inc.
114 Cummings Park
Woburn, MA 01801
Main office:
781.935.3322 voice
781.935.6622 fax
http://www.allsystems.com
terry@allsystems.com
_____________________________

From: "Cavanagh, Ken", INTERNET:ken_cavanagh@mcgraw-hill.com
Date: Tue, Apr 24, 2001, 11:00 AM
RE: RE: [colortheory] Sharpening workflow

Dan wrote:

> This depends on the skill of the user and the program's capabilities.
> Photoshop's execution of the Radius and Threshold options is nice. It would
> be better if they had a lightening/darkening option. Other programs are
> better in some ways and worse in others.

Working with images that are already sharpened, but need to be resized, I was curious if I should expect a difference between the way quark and photoshop handle this. On a dye-sub proof, I compared the scaling of an image in quark as opposed to scaling in photoshop and dropping it into quark. The quark(4.1) version showed a moire pattern whereas the ps(5.5) version was clean. Should I expect that moire to show in print?

From: Bob Tyson, INTERNET:bobicho@earthlink.net
Date: Tue, Apr 24, 2001, 8:15 PM
RE: Re: [colortheory] Sharpening workflow

>This depends on the skill of the user and the program's capabilities.
>Photoshop's execution of the Radius and Threshold options is nice. It would
>be better if they had a lightening/darkening option. Other programs are
>better in some ways and worse in others.

Actually Photoshop DOES have a lightening/darkening option for sharpening--- sort of. If you use the "Sharpen/Blur" tool you can select Lighten or Darken in the Options palette. This is an incredibly powerful option for local control-- it never occurred to me until now but maybe you could make a "brush" big enough to cover the whole image--- really humungous--- and turn it loose with one or two clicks.

Might take the thing a week to do its work, though.

Bob Tyson

Bob Tyson -//- photographs
http://home.earthlink.net/~bobicho/bperspho.html

From: tflash, INTERNET:tflash@earthlink.net
Date: Wed, Apr 25, 2001, 1:22 PM
RE: [colortheory] Luminosity Sharpen

> For sharpening, a luminosity fade and or blend is ideal.

Just out of curiosity, is there any difference between:

A) Sharpen L channel in LAB.
B) Duplicate layer and sharpen with layer mode set to luminosity.
C) Sharpen, then do an Image>Fade Filter, with mode set to luminosity.

Todd Flashner

From: Dan Margulis, INTERNET:76270.1033@compuserve.com
Date: Wed, Apr 25, 2001, 5:13 PM
RE: [colortheory] Luminosity Sharpen

Todd writes,

>>Just out of curiosity, is there any difference between:
A) Sharpen L channel in LAB.
B) Duplicate layer and sharpen with layer mode set to luminosity.
C) Sharpen, then do an Image>Fade Filter, with mode set to luminosity.>>

B and C should be identical. Method A is technically superior but unlikely to make much difference.

Assuming for the sake of argument an RGB file, methods B and C first sharpen three channels and then average the result in a ratio of roughly 6 parts green, 3 parts red, one part blue. Method A in effect averages first and sharpens afterward.

Method A is theoretically better because

1) If a certain area of a certain channel in RGB is very light or very dark, it may not be possible to sharpen it fully because the sharpening halo can't get any lighter than 255 or darker than 0. But if the channels are averaged first, the area won't be so close to the edge and proper sharpening can occur.

2) If an RGB channel (especially the blue) is particularly noisy, the Threshold command won't be able to exclude the noise from the sharpening nearly as well as if the channels had been averaged first.

But again, it's not a big deal. I wouldn't convert from CMYK to LAB to sharpen, but I'd consider RGB to LAB.

Dan Margulis

Adobe Photoshop training classes are taught in the US by Sterling Ledet & Associates, Inc.