Curve, Foster Road, West Rowan County, NC III

This version was cropped from the same file as the following photo (click on right arrow).

 

Except for the crop and resizing in Photoshop, this picture is totally unmanipulated.

 

That includes NO SHARPENING!

 

--Lannie

There is NO SHARPENING or other manipulation on this one! It's ashame that Kodak gave up on perfecting these old DSLRs. Commentswelcome.

--Lannie

Unfortunately that phase has precious little meaning with a digital camera.  They do a lot of processing to the sensor outputs.  It is a lovely picture, though.  best, j

Okay, Jamie, so "Manipulation by Kodak" might be the most honest description of all.

 

Kodak used to know something about color, etc. It's a shame they forgot it.

 

--Lannie

Here is a 100 percent crop with no sharpening or other manipulation in post processing.

 

--Lannie

Here is a 100 percent crop with a very shallow S-curve applied, along with unsharp mask at 85--that is, with rather conservative post-processing.

 

--Lannie

Here it is with a bit more aggressive processing--a deeper S-curve along with unsharp mask at 150.

 

--Lannie

Finally, here is a version of the same 100 percent crop with deeper curves, some brightness and contrast adjustments, and unsharp mask set at 200.

 

--Lannie

Here's an bigger challenge in post-processing.

 

This was a simple crop with no post processing.

 

(The purple fringing against the sky is pretty obvious.)

 

--Lannie

Here is the same 100 percent crop with curves applied, along with unsharp mask set at 100.

 

I am left wondering where the focal plane was on this shot, as well as how much better modern digital cameras can do with the purple fringing problem.

 

--Lannie

processing by engineers can be pretty extreme. nothing out of a digital camera is unmanipulated or unmolested.

 

the purple fringing is the result of extremely high contrast. it is always there, but it is tiny in relative terms. when the bright parts are many times the maximum recordable exposure and the dark bit is a silhouette, the purple is strong enough to be seen due to the big numbers involved. i don't think there could be a very good algorithm to eliminate it because, as i said above, the sky is well above the recordable maximum, so the camera has no idea how much light was there. that information would be necessary to correct the fringing. best, j

No, Jamie, in-camera processing is never insignificant, but I assumed that readers knew that I was talking about post-processing when I labeled a given treatment as "unprocessed."

 

As for purple fringing, which company's cameras are doing the best in eliminating or controlling it, in your opinion?  When I sold the Kodak 14n to a guy in 2006, he told me (upon using the 14n for a while) that his Nikon D70 had been much better at avoiding purple fringing than the Kodak.  I never had the D70, but I do have an under-used D90 which I suppose that I could take out for a side-by-side comparison with the Kodak DCS Pro SLR/n.  I wonder if the guy was correct in suggesting that Nikon did a better job on purple fringing than Kodak did.

 

I am sitting here wondering what variables would be involved in reducing (or exacerbating) purple fringing.  If one controlled it, would other color aberrations take its place?  Why, for that matter, is it always purple that hits us in the face in backlit situations?  I confess that I have never seriously made any comparisons.  To me it is simply a given now that limbs (especially small limbs) against the sky (especially when those limbs or similar objects are backlit) are going to not only have purple around the edges, but are sometimes going to be purple in their entirety in certain cases.  My "solution" has been to try to avoid such situations when I do the actual shoot in the first place. 

 

I am assuming that the "purple" in question here is actually magenta, the complementary color of green.  I wonder if how a camera manages green has any bearing on how much purple/magenta is displayed.  I am assuming, of course, that the old RGB (additive) and CMY (subtractive) colors are somehow being played off each other in the in-camera processing phase, but I do not know this. 

 

I do know that certain types of color aberration are exaggerated with certain post processing procedures (such as adding too much contrast, and using curves in certain ways), but I have never seen a systematic treatment of the subject. 

 

--Lannie

that word takes advantage of the gross ignorance of the population.  Leica JPGs look fantastic, but only because they are so vastly more aggressively processed than JPGs from other cameras.

 

it reminds me of my time working in a stereo store.  lots of people honestly believe that if one amplifier plays louder than another, with both volume knobs at 9:00, it will also play louder in an absolute sense at other settings.  no amount of explaining will help them.  there is no substantive difference when you say a picture is unprocessed.  it's set at 9:00.  it goes to eleven.

 

the stuff you discuss - sharpening, contrast enhancement - cameras do that.  i think it's difficult to even distinguish between what is necessary to get a physically representative signal and what is enhancement.

 

purple finges come from the lens, not the camera.  as i said, any engineering solution in a camera would have to be based on clever tricks, meaning that it wouldn't be based on sound principles.  a better lens should reduce the fringes.

 

the fringes are purple because the lens is corrected for chromatic aberration in a first-order sense - the differences between red and blue have been eliminated.  that leaves second-order differences between green and purple (red plus blue).  best, j

But, Jamie, red and blue make purple only in subtractive colors, i.e., pigments.  In additive colors (light), red and blue do not make purple, any more than blue and yellow produce green with light (although they do with subtractive colors).  Purple qua violet can surely be made with light, but I don't know what to mix with what to get purple.  I can surely get magenta with light, but purple qua violet and magenta are not really the same color any more than cyan and blue are the same color.

 

So, something still does not compute to me.  In any case, even though the purple begins as optical abererration, what we do in camera surely affects how it is exaggerated or diminished on the screen.

 

I could be wrong on all the above, of course.  Playing around with color balance in Photoshop to produce and modify colors can be very frustrating.

 

As for PS, mixing yellow and cyan appear to give turquise, not RGB green--as I see it.  Again, I could be wrong.

 

Explain to me what I am not seeing, Jamie.

 

--Lannie

sorry dude.  my ace trumps.  i'm a colour scientist, remember.  in subtractive mixture, red and blue make black, although never a very good black when it comes to pigment mixture, which is not exactly subtractive or exactly anything for that matter.  

red light plus blue light appears purple.  that's why purple is considered "extra-spectral" - there is no monochromatic light that looks purple.  the shortest wavelengths appear violet.  best, j

Ignoring the discussion above, which is an area I know nothing about, I will deal with the photo itself.  It projects a summer scene that I am familiar with even though I'm about 800 miles north of this location.  I like this version best as, with it's greater foreground,  it emphasizes the length of road.  Very nicely composed for sure though the contrast range is approaching the limits of the sensor to record it. It comes to my screen a tad hazy.   It's my experience that, at least with my only digital camera, the P&S Canon A540, some post processing is nearly always required to get the best from the image.  This post processing must be done with a light hand or it starts to LOOK manipulated.  Not a desirable result! I've taken some liberties with your image to satisfy MY needs. Hope I haven't overdone it. Your mileage, as they say, may differ but it works for me.  All in the intended spirit of elevating the image's attraction.  Best, LM.

Thanks, Len, I think that you have hit that "happy medium."   A similar shot was probably overprocessed by me in the same folder (two photos to the right of this one).  This one was not processed at all to try to preserve that light green color of April in these parts on the trees and grass.  Even so, you are right.  It needs something more.

 

One can never remember exactly what one saw on the day in question, and so I suppose that the best that we can do in most cases is something like what you have done: a bit of post processing, but with a light hand.  A heavy hand might catch the eye, but it definitely does not give a realistic portrayal and in many cases is just too much to take.

 

I am pretty sure that as the files come out of the camera, they are also not realistic.  I am not saying that "realism" is the be-all and end-all of post-processing, but I like to be as realistic as I can with landscape work--although we will all tend to punch it up a bit more from time to time.

 

Thanks for visiting, Len.  As usual, you and Jamie both are right on target.

 

Now I am hoping that Jamie will give me some titles of books to read about color, in both theory and practice.  I could use some help.

 

--Lannie

Len, I just looked at your version again.

 

That is truly nice work.  Now I feel like it is your shot!

 

On the one two photos to the right in the folder (or on this image to the right by two right arrow clicks), Jamie had recommended "backing off" a bit on the post-processing, which would, I think, have given results akin to what you have gotten on this similar but different file.

 

Again, thanks to both of you.

 

--Lannie