Ad
Ad
Ad
Pages: [1]   Bottom of Page
Print
Author Topic: what does sharpening->detail really do?  (Read 1640 times)
hjulenissen
Sr. Member
****
Offline Offline

Posts: 1713


« on: February 18, 2013, 08:26:17 AM »
ReplyReply

I find all of the explanations really fuzzy. Does it distinguish between edges and structure, or is it merely some midfrequency/highfrequency tilt parameter of the "gaussian"?

-h
Logged
BartvanderWolf
Sr. Member
****
Offline Offline

Posts: 3909


« Reply #1 on: February 18, 2013, 09:01:08 AM »
ReplyReply

I find all of the explanations really fuzzy. Does it distinguish between edges and structure, or is it merely some midfrequency/highfrequency tilt parameter of the "gaussian"?

Hi,

It's fuzzy, a bit of a blur, indeed. Since it's proprietary, we'll never know exactly, but at 0 it's supposedly a form of USM sharpening, and at 100 it's a form of deconvolution sharpening (I'm not sure if it's based on a Gaussian PSF shape, or something more disc shaped). In between 0 and 100 it's a sort of mix between USM and Deconvolution.

Cheers,
Bart
Logged
Philmar
Sr. Member
****
Offline Offline

Posts: 355


WWW
« Reply #2 on: November 06, 2014, 10:43:03 AM »
ReplyReply

Are there certain types of photos that work best with the sharpening/detail slider set at 100% ?
Logged

An office drone pension administrator by day and a photo-enthusiast by night, week-end and on vacation who carries his camera when traveling the world:
Please have a chew on my photos:
http://www.fluidr.com/photos/phil_marion/sets
Eyeball
Full Member
***
Offline Offline

Posts: 150


« Reply #3 on: November 06, 2014, 11:10:31 AM »
ReplyReply

Are there certain types of photos that work best with the sharpening/detail slider set at 100% ?

For me, these are the characteristics that would encourage me to try the detail at 100:
- Low noise image (in LR, I find high detail brings out the noise and if you use noise reduction that will undo the high detail)
- Lots of fine detail in the image
- I am interested in mild "capture sharpening" only.  In other words, high detail but low amount.
- The image is just slightly soft at 100%, perhaps due to lens characteristics, diffraction at small apertures, or the anti-alias filter on the sensor (in other words, typical "capture sharpening" territory.)

Really though, I would normally not do this in LR unless I was in a hurry or didn't have PS or Focus Magic available.
Logged
Hans Kruse
Sr. Member
****
Offline Offline

Posts: 822



WWW
« Reply #4 on: November 06, 2014, 11:25:07 AM »
ReplyReply

Are there certain types of photos that work best with the sharpening/detail slider set at 100% ?

I use detail=100 for f/16+ pictures and a radius slightly higher at 1.0 than my default at 0.8. It seems to make f/16 pictures (on 35mm FF) slightly more crisp. It is my understanding that it is deconvolution algorithm at 100.
Logged

deejjjaaaa
Sr. Member
****
Offline Offline

Posts: 1133



« Reply #5 on: November 06, 2014, 11:31:54 AM »
ReplyReply

I find all of the explanations really fuzzy. Does it distinguish between edges and structure, or is it merely some midfrequency/highfrequency tilt parameter of the "gaussian"?

-h

try the posting from Eric Chan may be ? him being the proverbial whoever :

Eric Chan : ( http://www.luminous-landscape.com/forum/index.php?topic=45038.msg378027#msg378027 ) @ July 24, 2010

"...the sharpening in CR 6 / LR 3 is a continuous blend of methods (with Detail slider being the one used to "tween" between the methods, and the Amount, Radius, & Masking used to control the parameters fed into the methods). As you ramp up the Detail slider to higher values, the deconvolution-based method gets more weight. If you're interested in only the deconv method then just set Detail to 100..."
Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 8028


WWW
« Reply #6 on: November 06, 2014, 11:50:00 AM »
ReplyReply

Hi,

Yes, quite clear from Eric Chan's writing that detail at 100% is deconvolution sharpening, but I don't know what PSF that would use. A good PSF (Point Spread Function) is essential for good deconvolution.

Detail at 0% supresses halos.

Personally I found that radius 0.7 may be optimal with medium apertures and I often find that 1.3 is optimal at f/16 - f/22.

Best regards
Erik

I use detail=100 for f/16+ pictures and a radius slightly higher at 1.0 than my default at 0.8. It seems to make f/16 pictures (on 35mm FF) slightly more crisp. It is my understanding that it is deconvolution algorithm at 100.
« Last Edit: November 06, 2014, 11:51:34 AM by ErikKaffehr » Logged

Eyeball
Full Member
***
Offline Offline

Posts: 150


« Reply #7 on: November 06, 2014, 11:57:14 AM »
ReplyReply

I and others have mentioned it before in other threads but LR's deconvolution never seems quite up to par compared to specialized deconvolution tools like Focus Magic.
For one thing, it always seems to me to be more difficult to "focus" the settings visually.
With other tools like FM, it is easy to find the "sweet spot" for the deconvolution.
Logged
deejjjaaaa
Sr. Member
****
Offline Offline

Posts: 1133



« Reply #8 on: November 06, 2014, 12:17:52 PM »
ReplyReply

I and others have mentioned it before in other threads but LR's deconvolution never seems quite up to par compared to specialized deconvolution tools like Focus Magic.

note that ACR/LR strive to produce a real time preview experience for a variety of computers (and users) out there... that somewhat limits what they can do computation wise

Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 8028


WWW
« Reply #9 on: November 06, 2014, 01:05:06 PM »
ReplyReply

Hi,

FocusMagic is quite impressive.

I would think that LR sharpening is quite good, but I have some doubts about the demosaic parts: http://www.luminous-landscape.com/forum/index.php?topic=94812.0

Best regards
Erik


I and others have mentioned it before in other threads but LR's deconvolution never seems quite up to par compared to specialized deconvolution tools like Focus Magic.
For one thing, it always seems to me to be more difficult to "focus" the settings visually.
With other tools like FM, it is easy to find the "sweet spot" for the deconvolution.
« Last Edit: November 06, 2014, 01:07:19 PM by ErikKaffehr » Logged

Pages: [1]   Top of Page
Print
Jump to:  

Ad
Ad
Ad