what is the deal with chromatic aberration?

back alley

IMAGES
Local time
3:22 PM
Joined
Jul 30, 2003
Messages
41,289
isn't it barely noticable unless you go looking for it?

does it really ruin an image for folks?

do non photographers notice it?
 
Although I can find it in many of my images if I look, I rarely find it actually ruins many images. My 85L used to crop up some nasty aberration in the worst circumstances, but with a lens that wide it's not really a surprise.

It's usually trees against a blown out sky type shots that are the worst, and I don't take many of them :)

And considering I convert to B&W roughly 75% of the time, chromatic aberration usually doesn't make or break a lens for me.
 
The way my brain is wired (which is the ultimate negative disclosure!) it makes images look really digital. I don't ever recall seeing CA in analog prints, slides, or really even scans...and I really hardly ever had good glass for my old film stuff.
 
It is the ability to zoom-in 100% with the click of a mouse that made this and other lens aberrations so popular to discuss.

I recall slides that I adored until I had them greatly enlarged and then started to notice all kinds of imperfections. Now days, I take several such 'poor' images and I re-size them for the web only. They look great!!
 
isn't it barely noticable unless you go looking for it?

does it really ruin an image for folks?

do non photographers notice it?

It can really hurt a photo, but it's not too common where the aberration is *that* noticeable. I had some photos recently that are badly affected: some clear pond ice, lots of bubbles inside, with the sun shining hard. Each bubble has a little magenta edge, and a little green edge, and collectively, it's horrible. A crop from one:

6935760915_4cb93e8124_b.jpg



Thin, backlit tree branches can produce something similar, where the aberration is enough to really hurt a shot.
 
The way my brain is wired (which is the ultimate negative disclosure!) it makes images look really digital. I don't ever recall seeing CA in analog prints, slides, or really even scans...and I really hardly ever had good glass for my old film stuff.

There's always been chromatic aberration to varying degrees, even in film, I the digital age however has got us peeping pixels a lot more than we used to. Here's a shot taken in '81 with Kodachrome... but I never noticed the chromatic aberration until I went looking for it because of this post, see what I mean :) the second pic is a close crop of the top left corner. I've got a rather large blow up of this pic and the aberration never bothered me... I really hope that it's not the only thing I see now when I look at the print! I really thing it's one of those things not many people notice or care about in most images until it's pointed out to them.




babycrop.jpg
 
Chromatic aberration will cause a color fringe on strong contrast details, especially in the periphery. Of course the color itself can be disturbing, but the main problem is that it greatly reduces your overall resolution. It is probably one of the main factors determining the crispness of your image. This holds true, of course, for both black and white as well as color film.

- p
 
@paul exactly... granted I have see a few cases where it's really bad... in most cases you have to look really hard and BE looking for it. Many people find it's a deal breaker for some lenses or images, but on the whole I think there's a lot more to worry about when creating a good image.

I'd rather have a well composed photo of an interesting subject that's got a bit of fringing and maybe isn't razor sharp... than a perfectly sharp image of the perfectly parallel lines of an undistorted brick wall ;)
 
The way my brain is wired (which is the ultimate negative disclosure!) it makes images look really digital. I don't ever recall seeing CA in analog prints, slides, or really even scans...and I really hardly ever had good glass for my old film stuff.

I agree with this...

The M8 and M9 do it with really fast glass. Generally with trees and sky or seemingly anything else high contrast outside its dynamic range. However, according to Wikipedia:

"Other attributed causes of purple fringing in digital photography include many hypothesised sensor effects:

Digital noise in dark areas
Image processing and interpolation artifacts (almost all CCDs and CMOS require considerable processing)
Stray ultraviolet light
Stray infrared light
Image bloom from overexposure of CCD sensor (not applied to CMOS)"
 
@ back alley the shorter the focal length, the more potential chromatic aberration you get, so by default, it will affect medium and large format less as with MF, 80mm is your standard "normal" lens as apposed to 35-50mm for 135.
 
harry, i don't have a clue of any color issues in your shot.

Do you see all the magenta and green dots in all the highlights on the water? I do, and they bug me a lot. Hard to get rid of, too, and in an image like this, almost impossible.

It's also fixable by just going black & white, of course.

EDIT: http://farm8.staticflickr.com/7064/6935760915_96f3e29877_o.jpg to see the full-sized crop, where it's much worse. Sorry, forgot I was linking to a step-down for the image. If I were to try to print this, I'd be in trouble. The magenta and green dots are overwhelming.
 
@back alley - usually more prevalent in zooms as modern lenses have elements specifically to counter that type of aberration, but it's much harder to build compensation elements for every focal length you can set a zoom to as apposed to a fixed focal lenght
 
harry, i am mildly red/green colorblind, so i don't know - really - the exactly correct color for anything. the tiny bubbles look blue/white to me.
is it possible the water might be slightly polluted with gas or oil? this can make tiny bubbles really colorful. or it could be my laptop screen just isn't up to the task of such fine points. the ice photo is a fine one in any case ...
 
There are three types of chromatic abberations that cause artifacts in digital photograohy. Other artifacts are often mis-attributed to CA. The term has morphed into a description of any sort of edge color distortion.

Lateral CA is caused by lens problems where the different frequencies of light do not focus in the same place. This effect is easy to eliminate in post-processing as ling as it's relatively small. Otherwise the affected areas will have a white outline. I have owned lenses where the artifacts were not uniform thoughout the image. This means it difficult to remove the artifacts. Many cameras automatically correct for lateral CA when the jpeg image is rendered in camera. Lightroom 3 lens correction plug-ins work well too as do other solutions in other image processing software.

Longitudinal CA artifacts are also caused by lens problems. These artifact are related to different frquencies ending up in different focal planes. This type of CA is often seen in out of focus objects. There is a transition from red to green fringing in the out of focus regions. Longitudinal CA is common in fast lenses when they are used at low F stop numbers. It is difficult to minimize longitudinal CA in post processing. I understand Nikon's NX RAW doftware has a tool to reduce longitudinal CA. I have no idea how well it works as I would rather stick a red-hot needle in my eye than use Nikon NX.

Then there's purple fringing and artifacts that turn non-blue objets blue/purple when they are in or adjacent to an overexposed region.*Some of these artifacts arise from chromatic abberations from the sensor microlenses in the Bayer filter array. Artifacts that have nothing to do with CA can also be seen when regions of the frame are overexposed. Sensor blooming is due to current contamination from overexposed sensor sites to properly exposed sensor sites. Then there are errors in the algorithms used to model the Bayer array. These errors can reveal themselves when a region is overexposed.

Does CA matter? I have seen lateral CA in film images. I believe the Bayer array somehow makes lateral CA more apparent. I use Nikon digital cameras and longitudinal CA has saved me thousands of dollars. I refuse to by Nikon's best primes because the longitudinal CA is strong and very unpleasant to my eye. Their slower primes have less longitudinal CA so I buy these. Longitudinal CA is obvious in prints.

I have noticed that purple fringing has decreased significantly in the digital cameras I've used in the past 5 years. The D700 is much more resistant to over-exposure artifacts than the D300 or D200. The X100 is better than the D700. The X100 lens has coma artifacts with bright point-source lights, but it is difficult to generate blue/purple fringing.
 
Back
Top Bottom