learned a new method of sharpening (PS)

volosong · 05-02-2007, 01:54 PM

"Discovered" podcasts put out by TPN :: The Digital Photography Show and on one of their episodes, conducted an interview with Scott Kelby. When it comes to Photoshop, Scott knows what he is talking about.

During the interview, he shared the methodology by which he applies sharpening to his images. The biggest problem with sharpening is the "ringing" around edges of high contrast. His method avoids this. (He also shared his unsharp mask parameters. 85,1,4)

What Scott does is convert his RGB images to LAB color. (I'm used to calling this color space "HSI" for "hue", "saturation", and "intensity". LAB and HSI are the same, aren't they?) Anyway, he applies his unsharp mask, or smart sharpen mask to only the "lightness" channel, not the "A" nor "B" channel. What he says you can do is apply the same filter twice in succession without any degradation of the image. Then, he converts back to RGB color space for final output or further processing, if desired.

(Also "discovered" a better way to apply NeatImage noise reduction plug-in. I usually don't use profiles, but instead let the plug-in select a training area. Results are mixed. This time, I processed each LAB channel separately, with pretty good results. The training area selected by the plug-in differed with each LAB channel.)

LAB sharpening method seems to work pretty good. Hard to tell on the image below due to having to size it down for web posting. Pixel peeping on the full-res images really shows how much better LAB filtering is over staying in RGB mode. Whenever I need to sharpen, I'll use LAB color from now on.

Image taken on April 3rd of this year, at just past midnight of a full-sized copy of Michelangelo's David standing in the Piazza della Signoria in front of the Palazzo Vecchio in Firenze, Italia (Florence, Italy).

K10D
DA 21/3.2
P exposure mode
f/3.2
ISO 1000
0.3s (yup, 1/3 of a second, hand held - gotta love that Shake Reduction)
center weighted average metering
no flash

Last edited by volosong; 05-02-2007 at 02:06 PM.

Leya216 · 05-02-2007, 02:27 PM

Copy background layer > apply the "high pass" filter on the new copy, usually use between 0.7 to 1.3 value > change the copy layer to "overlay" . If you have a sharpened file and then you want to resize it and post it on the web sharpen again after resizing.

deejjjaaaa · 05-03-2007, 02:08 PM

Originally posted by volosong

What Scott does is convert his RGB images to LAB color.

why not just "EDIT -> FADE" after filter to avoid RGB->LAB->filter(s)->RGB as
many people do for years ?

philmorley · 05-03-2007, 07:36 PM

Originally posted by volosong

"Discovered" podcasts put out by TPN :: The Digital Photography Show and on one of their episodes, conducted an interview with Scott Kelby. When it comes to Photoshop, Scott knows what he is talking about.

During the interview, he shared the methodology by which he applies sharpening to his images. The biggest problem with sharpening is the "ringing" around edges of high contrast. His method avoids this. (He also shared his unsharp mask parameters. 85,1,4)

What Scott does is convert his RGB images to LAB color. (I'm used to calling this color space "HSI" for "hue", "saturation", and "intensity". LAB and HSI are the same, aren't they?) Anyway, he applies his unsharp mask, or smart sharpen mask to only the "lightness" channel, not the "A" nor "B" channel. What he says you can do is apply the same filter twice in succession without any degradation of the image. Then, he converts back to RGB color space for final output or further processing, if desired.

(Also "discovered" a better way to apply NeatImage noise reduction plug-in. I usually don't use profiles, but instead let the plug-in select a training area. Results are mixed. This time, I processed each LAB channel separately, with pretty good results. The training area selected by the plug-in differed with each LAB channel.)

LAB sharpening method seems to work pretty good. Hard to tell on the image below due to having to size it down for web posting. Pixel peeping on the full-res images really shows how much better LAB filtering is over staying in RGB mode. Whenever I need to sharpen, I'll use LAB color from now on.

Image taken on April 3rd of this year, at just past midnight of a full-sized copy of Michelangelo's David standing in the Piazza della Signoria in front of the Palazzo Vecchio in Firenze, Italia (Florence, Italy).

K10D
DA 21/3.2
P exposure mode
f/3.2
ISO 1000
0.3s (yup, 1/3 of a second, hand held - gotta love that Shake Reduction)
center weighted average metering
no flash

Hi, HSI and LAB are two different things. rather technical explantion of lab on wikipedia

Lab color space - Wikipedia, the free encyclopedia

dan margulis is very keen on it and amongst his other books he has one or two on lab

Amazon.com: Photoshop LAB Color: The Canyon Conundrum and Other Adventures in the Most Powerful Colorspace: Books: Dan Margulis

its not too bad and you can learn a lot about color correction, there are some things much easier to do in lab (and some that are not

and dgrin have a reading group which if your interested you can pick up stuff without having the book.

of course you can just use it for sharpening with less artifacts

Regards
Phil

volosong · 05-04-2007, 12:07 PM

Originally posted by philmorley

HSI and LAB are two different things.

Lab color space - Wikipedia, the free encyclopedia

Thanks for the clarification, Phil. I knew what HSI is, and have used it for some remote sensing satellite imagery "enhancements". First time I used LAB is with this attempt at sharpening. Thanks for the link too. It will take me awhile to get through that one.

05-02-2007, 01:54 PM	#1
volosong Veteran Member Join Date: Sep 2006 Location: North Idaho Photos: Gallery Posts: 696	learned a new method of sharpening (PS) "Discovered" podcasts put out by TPN :: The Digital Photography Show and on one of their episodes, conducted an interview with Scott Kelby. When it comes to Photoshop, Scott knows what he is talking about. During the interview, he shared the methodology by which he applies sharpening to his images. The biggest problem with sharpening is the "ringing" around edges of high contrast. His method avoids this. (He also shared his unsharp mask parameters. 85,1,4) What Scott does is convert his RGB images to LAB color. (I'm used to calling this color space "HSI" for "hue", "saturation", and "intensity". LAB and HSI are the same, aren't they?) Anyway, he applies his unsharp mask, or smart sharpen mask to only the "lightness" channel, not the "A" nor "B" channel. What he says you can do is apply the same filter twice in succession without any degradation of the image. Then, he converts back to RGB color space for final output or further processing, if desired. (Also "discovered" a better way to apply NeatImage noise reduction plug-in. I usually don't use profiles, but instead let the plug-in select a training area. Results are mixed. This time, I processed each LAB channel separately, with pretty good results. The training area selected by the plug-in differed with each LAB channel.) LAB sharpening method seems to work pretty good. Hard to tell on the image below due to having to size it down for web posting. Pixel peeping on the full-res images really shows how much better LAB filtering is over staying in RGB mode. Whenever I need to sharpen, I'll use LAB color from now on. Image taken on April 3rd of this year, at just past midnight of a full-sized copy of Michelangelo's David standing in the Piazza della Signoria in front of the Palazzo Vecchio in Firenze, Italia (Florence, Italy). K10D DA 21/3.2 P exposure mode f/3.2 ISO 1000 0.3s (yup, 1/3 of a second, hand held - gotta love that Shake Reduction) center weighted average metering no flash Last edited by volosong; 05-02-2007 at 02:06 PM.

05-02-2007, 02:27 PM	#2
Leya216 Senior Member Join Date: Mar 2007 Posts: 184	My favorite sharpening tech. Copy background layer > apply the "high pass" filter on the new copy, usually use between 0.7 to 1.3 value > change the copy layer to "overlay" . If you have a sharpened file and then you want to resize it and post it on the web sharpen again after resizing.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Your most basic PP method?	KierraElizabeth	Digital Processing, Software, and Printing	31	08-03-2009 11:21 PM
A New Method...	Buddha Jones	Photographic Technique	29	05-05-2009 06:49 PM
Archiving Your Photos: What's Your Method?	vinzer	Photographic Technique	17	01-09-2009 11:36 AM
Clean behind focusing screen - best method?	ksignorini	Pentax DSLR Discussion	18	05-14-2008 10:57 AM
Another Workflow Method...	benjikan	Digital Processing, Software, and Printing	0	03-16-2007 02:37 PM

05-03-2007, 02:08 PM	#3
deejjjaaaa Veteran Member Join Date: May 2007 Location: steel city / rust belt Posts: 2,043	Originally posted by volosong What Scott does is convert his RGB images to LAB color. why not just "EDIT -> FADE" after filter to avoid RGB->LAB->filter(s)->RGB as many people do for years ?

05-03-2007, 07:36 PM	#4
philmorley Veteran Member Join Date: Mar 2007 Location: in a house in Armidale, Australia Posts: 472	Originally posted by volosong "Discovered" podcasts put out by TPN :: The Digital Photography Show and on one of their episodes, conducted an interview with Scott Kelby. When it comes to Photoshop, Scott knows what he is talking about. During the interview, he shared the methodology by which he applies sharpening to his images. The biggest problem with sharpening is the "ringing" around edges of high contrast. His method avoids this. (He also shared his unsharp mask parameters. 85,1,4) What Scott does is convert his RGB images to LAB color. (I'm used to calling this color space "HSI" for "hue", "saturation", and "intensity". LAB and HSI are the same, aren't they?) Anyway, he applies his unsharp mask, or smart sharpen mask to only the "lightness" channel, not the "A" nor "B" channel. What he says you can do is apply the same filter twice in succession without any degradation of the image. Then, he converts back to RGB color space for final output or further processing, if desired. (Also "discovered" a better way to apply NeatImage noise reduction plug-in. I usually don't use profiles, but instead let the plug-in select a training area. Results are mixed. This time, I processed each LAB channel separately, with pretty good results. The training area selected by the plug-in differed with each LAB channel.) LAB sharpening method seems to work pretty good. Hard to tell on the image below due to having to size it down for web posting. Pixel peeping on the full-res images really shows how much better LAB filtering is over staying in RGB mode. Whenever I need to sharpen, I'll use LAB color from now on. Image taken on April 3rd of this year, at just past midnight of a full-sized copy of Michelangelo's David standing in the Piazza della Signoria in front of the Palazzo Vecchio in Firenze, Italia (Florence, Italy). K10D DA 21/3.2 P exposure mode f/3.2 ISO 1000 0.3s (yup, 1/3 of a second, hand held - gotta love that Shake Reduction) center weighted average metering no flash Hi, HSI and LAB are two different things. rather technical explantion of lab on wikipedia Lab color space - Wikipedia, the free encyclopedia dan margulis is very keen on it and amongst his other books he has one or two on lab Amazon.com: Photoshop LAB Color: The Canyon Conundrum and Other Adventures in the Most Powerful Colorspace: Books: Dan Margulis its not too bad and you can learn a lot about color correction, there are some things much easier to do in lab (and some that are not and dgrin have a reading group which if your interested you can pick up stuff without having the book. of course you can just use it for sharpening with less artifacts Regards Phil

05-04-2007, 12:07 PM	#5
volosong Veteran Member Join Date: Sep 2006 Location: North Idaho Photos: Gallery Posts: 696 Original Poster	Originally posted by philmorley HSI and LAB are two different things. Lab color space - Wikipedia, the free encyclopedia Thanks for the clarification, Phil. I knew what HSI is, and have used it for some remote sensing satellite imagery "enhancements". First time I used LAB is with this attempt at sharpening. Thanks for the link too. It will take me awhile to get through that one.