Slightly different sharpening

Bill Pierce

Well-known
Local time
7:14 PM
Joined
Sep 26, 2007
Messages
1,407
Here is a thread entry that will probably generate less response than any other RF Forum entry.

Recently I was asked how I sharpen those images taken with cameras that have no anti-aliasing filters, in my case the Ricoh with a Leica M mount and the Fuji X-Pro, X-E1 and X100s. Indeed, it does seem that to take full advantage of these images, slightly different sharpening techniques from the ones we have been using can be beneficial.

What I came up with is really for pixel peepers putting their noses against large prints or, in my case, someone avoiding desk work by playing with images on his computer. I looked at Photoshop, Photokit Sharpener 2 (within Photoshop), Lightroom, Capture One, SilkyPix, RPP, Iridient Developer and AccuRaw.

All of them, used conventionally, performed more than adequately for most work. But I got the best results when pixel peeping images from these relatively new cameras without anti aliasing filters by applying an initial “presharpening” at a much lower radius (.3) than normally used. The programs that allowed that were PK Sharpener, Capture One, Iridient Developer and AccuRaw. This “presharpened” file can then be moved into any image processing program. Often it will benefit from additional sharpening at more conventional radius settings. That’s your call. But, I will say, some of the big prints from these little cameras are scary sharp and really do allow them to compete with the big boys in terms of image quality.

Your thoughts?
 
I agree, much improvement has been made with most all digital cameras, including tablet computers and smart phones. I don't particularly like to sharpen the files I make. To me, I see many images where things are overdone with sharpening being one item that's used.


A lady that has worked with me, she is my age, says I sometimes do too much with PS and has me back the opacity off with the layer I'm working on.

That is one of the many reasons I like to use film for some projects, printing in my chemical darkroom.
 
Your approach is similar to that suggested by Capture One (Phase One) themselves:

"First, sharpen the fine detail in the image.
Use the capture or raw pre-sharpening setting.
(In the Manage Preset menu)
Secondly, a more aggressive sharpening affect can be used but this should be applied after the image has been sized according to desired printing dimensions "

http://help.phaseone.com/en/CO7/Editing-photos/Details/Sharpening.aspx#item4

Note that the Phase One Backs have no AA filter. Lazy or not I have been happy with the C1 default settings for my P20 back but then I have not explored mural size printing.
 
With the vast improvement in site architecture and better quality lenses (often inexpensive when compared to M glass) have made using smaller cameras a better way for many of us to work. The Fuji X camera sensors will only improve as long as glass is available to take advantage of the better resolution. Others are following the Fuji path.

With digital rot a real problem, printing is the best solution for keeping the best images around for the years to come.

I still print from LR but that may change soon.
 
Why wouldn't your post generate a ton of replies? I think even prehistoric cave painters got into raging arguments about whose brush had better definition and paint flow.
 
This topic is nicely managed by the sharpening theory offered by the late Bruce Fraser and codified in the book he co-wrote with Jeff Schewe prior to his death, Real World Image Sharpening with Adobe Photoshop, Camera Raw, and Lightroom. To summarize...

Image sharpening has three phases to it in image processing: input sharpening, creative sharpening, and output sharpening.

- The amount of input sharpening required depends upon the strength of the AA filter, which to some degree depends on that pixel density. Cameras with relatively high pixel counts (12Mpixel and up) require less AA strength to minimize moiré, and so require less input sharpening than cameras with lower pixel counts and stronger AA filters.

For example, I shoot with both an Olympus E-1 (5Mpixel) SLR and a Leica M9 (18Mpixel) RF. Both have very high performance lenses. The E-1, with its relatively low pixel count, has a strong AA filter and the M9, with its high pixel count, has no AA filter. The default input sharpening provided by Lightroom for the E-1 is a little weak ... I usually bump it up some. The M9, on the other hand, has such crisp rendering with a well focused lens that for the most part I find I need to reduce the input sharpening to near nothing. I found the same thing on the Olympus E-5, another camera with weak to non-existent AA filter and 12 Mpixels—very little input sharpening required.

- Creative sharpening are all the selective sharpening operations needed to optimize a particular image's perceptual sharpness. This is dependent upon the subject and your intent in rendering it.

- Output sharpening adjusts perceptual sharpening based on the needs of the output medium and sizing of the final image. This can be automated, and is in Lightroom.

So, your fundamental question, "How should I sharpen photos made with the M9 and other cameras with no AA filter?" comes down to "Less input sharpening than other cameras, the same creative and output sharpening based on the intent and output desired."

G
 
Bruce Fraser was one of the authors of Photokit Sharpener:

http://pixelgenius.com/sharpener2/

I use it for prints - I do the basic tonality in Lightroom, but no sharpening, then fine tuning in Photoshop including capture sharpening, B+W conversion, and output sharpening. Still the best I have found.

Cheers,
Kirk
 
Bill,

Your observations are completely in-line with what I have observed using LR 4.4/5. In addition I find the clarity slider and even the defringing slider can have a larger impact on IQ than one might expect from working with raw from Bayer CFA cameras.

Also, I find fine adjustment of sharpening on an image by image basis benefits image detail more than my NEF raw renderings... where I rarely have to change sharpening from one image to another.

My hypothesis is when using ACR, both the absence of an AA filter and the unique demosaicing algorithm used for the XTrans CFA array require a different approach compared to other cameras. In my opinion ACR is capable of properly rendering XTrans raw. I am aware other raw software also works well. At this point I feel the differences between raw platforms are not that different than they are for different Bayer CFA cameras. Years ago people claimed Nikon's NX2 or DxO or another platform worked better for NEF raw than ACR. No doubt there are differences and people will use the platform that best meets their needs.
 
Thanks, Bill. Everyone who has used Accuraw has praised it...I'm going to give it a try based on your test.

What version of Photoshop did you use? I ask because Adobe is making a lot of claims about the sharpening being much better in the new subscription version. Also a lot of folks claim the best adobe sharpener in ACR (CS6 includes a version of ACR that included sharpening ability).

Thanks for doing this

Tom
 
While I continue to use CS6, I find that using Topaz Detail as a plug in, and being able to differentially enhance large, medium or small detail has produced results that I was unable to achieve in PS. No halos, the ability to not enhance noise...Quite happy now that I've incorporated it into my workflow.
 
Back
Top Bottom