Chromatic Aberration or Fringing?

traderdrew · 03-09-2015, 05:09 PM

Hi Everyone,

I was at Daytona Bike Week recently and took my K-5 II with the 16-50mm lens. It's not a lens that I don't usually use but I liked using it there and I really like the lens. Something I noticed about my photos after I took them. I have had a problem with a sort of purple or blue fringing associated with metal parts of motorcycles. It was gone when I adjusted my settings and took another photo. Here are some heavy crops:

The one with the fringing was taken at 1/160, ISO 200 at F/2.8. The second one at 1/60, ISO 500 at F/8. I think using F/8 was the contributor to losing the fringing. Any thoughts?

Attached Images

View Picture EXIF

View Picture EXIF

dyoon153 · 03-09-2015, 05:31 PM

I think it's because photo is out of focus at f2.8. You probably shot at the close proximity of the bike (I see you used widest end zoom), so focus pane is very narrow, and out of focus metal reflection probably causes that fringe...

stevebrot · 03-09-2015, 05:37 PM

That looks to be PF. There is one line of thinking that holds PF to be longitudinal CA in the UV portion of the spectrum re-emitted as a longer wavelength bloom at the sensor surface*. Stopping down increases the DOF and minimizes the effect.

Steve

* I was unable to find the reference, but the study was done using narrow band LED lamps where the spectrum was UV deficient. PF was absent until a UV component was added to the spectral mix.

carpents · 03-09-2015, 05:41 PM

Originally posted by stevebrot

That looks to be PF. There is one line of thinking that holds PF to be longitudinal CA in the UV portion of the spectrum re-emitted as a longer wavelength bloom at the sensor surface*. Stopping down increases the DOF and minimizes the effect.

Steve

* I was unable to find the reference, but the study was done using narrow band LED lamps where the spectrum was UV deficient. PF was absent until a UV component was added to the spectral mix.

This is the reference I usually cite:
Digital Photography FAQ

---------- Post added 03-09-15 at 08:47 PM ----------

Originally posted by traderdrew

Hi Everyone,

I was at Daytona Bike Week recently and took my K-5 II with the 16-50mm lens. It's not a lens that I don't usually use but I liked using it there and I really like the lens. Something I noticed about my photos after I took them. I have had a problem with a sort of purple or blue fringing associated with metal parts of motorcycles. It was gone when I adjusted my settings and took another photo. Here are some heavy crops:

The one with the fringing was taken at 1/160, ISO 200 at F/2.8. The second one at 1/60, ISO 500 at F/8. I think using F/8 was the contributor to losing the fringing. Any thoughts?

I've used a lot of Pentax lenses and find them to be somewhat prone to purple fringing (which is a form of chromatic aberration) wide open. The newer lenses, those released in the last ~3 years or so, are better. Stopping down helps resolve it.

stevebrot · 03-09-2015, 06:46 PM

Originally posted by carpents

This is the reference I usually cite:
Digital Photography FAQ

Cool! Thanks!

Steve

**Digitalis** · 03-09-2015, 06:48 PM

Originally posted by stevebrot

That looks to be PF. There is one line of thinking that holds PF to be longitudinal CA in the UV portion of the spectrum re-emitted as a longer wavelength bloom at the sensor surface.

If that was the case the UV light would have to have an intensity nearly four times that of visible light for such a spectral shift to be visible on a digital sensor. I'm not sold on the Ultraviolet hypothesis, the reason being is that borosilicate glass is pretty damn good at attenuating those high frequencies, and also the existence of a UV/IR absorbtion layer in font of the sensor will also soak up residual UV light.

stevebrot · 03-09-2015, 07:05 PM

Originally posted by Digitalis

I'm not sold on the Ultraviolet hypothesis, the reason being is that borosilicate glass is pretty damn good at attenuating those high frequencies

I understand what you are saying, but I believe it is the near UV light that is in question, plenty of which makes its way to the sensor. As you know, it is that portion of the spectrum that is so problematic for landscape film photography.

Originally posted by Digitalis

and also the existence of a UV/IR absorbtion layer in font of the sensor will also soak up residual UV light.

...and may be the cause of the problem.

I would have to do the experiment myself to see whether there is any truth to it, but the absence when UV is absent followed by presence when UV is present would seem to implicate UV.

As for intensity, I guess it would depend on the light.

Steve

stevebrot · 03-09-2015, 07:13 PM

Originally posted by stevebrot

As for intensity, I guess it would depend on the light.

To expand a little, I have two lenses that have the reputation for PF, the FA 35/2 and FA 77/1.8. In general use, I seldom see PF with either lens. A few months ago I set off to compound a recipe to reliably create PF. This is how I do it:

Crumpled aluminum foil
Electronic flash as main light source
About two stops over-exposed
Aperture wider than f/2.8
Focus on near point of foil wad
Subject distance about 24 inches (~60cm)

The PF shows as malignant bloom on the highlight surfaces of the OOF areas. For this scenario the light is quite intense with many of the sensels fully clipped.

Steve

**Digitalis** · 03-09-2015, 07:32 PM

Then you should measure the number of photons emitted by the flash with frequencies of 350-200nm with energies greater than say 3.10ev up to 6.2ev - As far as I'm aware, no one has ever quantified this. There are emission curves for Xenon arc flash tubes but we have no idea of the longevity of emission - it is possible the Ultraviolet pulse from flash is longer than the visible pulse, which would explain the increase of purple fringing with flash.

desertscape · 03-09-2015, 09:39 PM - **1 Like**

In most cases when this fringing is seen at wider apertures but not when stopped down, it is the result of dispersion and spherochromatism. Colors are going to separate when refracted. The designer must prevent this by using glasses that bring the colors back to focus at the focal plane. However, not all colors can be be corrected and there will be a residual called the secondary spectrum ( in achromatic lenses) and tertiary spectrum ( in APO lenses). Even a highly corrected lens will have a sweet spot in the area of the optic where it is better corrected than in other areas. This is typically in the paraxial (central) and zonal areas (out 20- 60% from the optical axis)of the lens. The marginal zone is usually poorly corrected because the designer knows that the shooter can stop down the lens to truncate the poorly corrected marginal rays. With color fringing, what you are seeing is the uncorrected colors coming from the marginal zone of the lens. One color from different zones of the lens will not focus at the same point. This is the combination of Longitudinal Chromatic and spherical aberration (Spherochromatism).

DeadJohn · 03-10-2015, 07:52 AM

@traderdrew, processing software often has tools for fixing fringing. Lightroom lens correction is one way to do it.

My copy of the DA* 16-50 is underwhelming at f2.8. I only shoot f2.8 if it's very dark or I want minimal depth of field. Stopping down to f4.0 makes it a spectacular lens.

Sagitta · 03-10-2015, 08:02 AM

If UV was the culprit, wouldn't adding a UV filter reduce or eliminate it? From everything I've read UV filters are basically only useful as damage protection for lenses on digital cameras.

carpents · 03-10-2015, 09:44 AM

Originally posted by Sagitta

If UV was the culprit, wouldn't adding a UV filter reduce or eliminate it? From everything I've read UV filters are basically only useful as damage protection for lenses on digital cameras.

UV filters have different cut-offs, they don't perfectly start/stop at the visible wavelength.

I understand the same as you for utility of UV filters on digital - they are not sensitive to UV like film is, which is why the filters are less useful. But they're obviously not 100% insensitive to it because people mod their cameras to shoot exclusively UV.

kh1234567890 · 03-10-2015, 10:33 AM

The K-5ii does not seem to have any great problem shooting a UV LED - I could probably get a better shot if I could be ars*d to set it up properly

photosbyspeed · 03-10-2015, 10:36 AM

Man this post is getting so scientific!! To sum it up, it sucks that the lens has such bad fringing in the OOF areas. Also, the bike isn't in good focus, causing the fringing to be all over your subject! The second image also was not in good focus, eliminating the fringing, but preventing you from creating a quality photo.

All this talk about science experiments and adding filters is pretty much a waste of time. Try focus peaking or zooming in on live view to get that sharp focus and minimize that fringing!

03-09-2015, 07:13 PM	#8
stevebrot Otis Memorial Pentaxian Join Date: Mar 2007 Location: Vancouver (USA) Photos: Gallery \| Albums Posts: 42,007	Originally posted by stevebrot As for intensity, I guess it would depend on the light. To expand a little, I have two lenses that have the reputation for PF, the FA 35/2 and FA 77/1.8. In general use, I seldom see PF with either lens. A few months ago I set off to compound a recipe to reliably create PF. This is how I do it: Crumpled aluminum foil Electronic flash as main light source About two stops over-exposed Aperture wider than f/2.8 Focus on near point of foil wad Subject distance about 24 inches (~60cm) The PF shows as malignant bloom on the highlight surfaces of the OOF areas. For this scenario the light is quite intense with many of the sensels fully clipped. Steve

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Any problems with Chromatic Aberration?	spbouvier	Pentax K-3 & K-3 II	20	01-09-2015 08:28 PM
Kr to K500 - Chromatic aberration issue	juio	Pentax K-30 & K-50	24	09-02-2014 10:33 PM
removing chromatic aberation/purple fringing in Lightroom	toukan	Digital Processing, Software, and Printing	5	10-12-2013 10:22 PM
Chromatic aberration on a 77mm Limited?	magic.pill	Pentax SLR Lens Discussion	10	06-25-2013 01:04 PM
Purple Fringing vs. Chromatic Aberration	dasuhu	Troubleshooting and Beginner Help	9	08-07-2010 02:05 PM

03-09-2015, 05:31 PM	#2
dyoon153 Forum Member Join Date: Oct 2013 Location: Vancouver, BC Posts: 84	I think it's because photo is out of focus at f2.8. You probably shot at the close proximity of the bike (I see you used widest end zoom), so focus pane is very narrow, and out of focus metal reflection probably causes that fringe...

03-09-2015, 05:37 PM	#3
stevebrot Otis Memorial Pentaxian Join Date: Mar 2007 Location: Vancouver (USA) Photos: Gallery \| Albums Posts: 42,007	That looks to be PF. There is one line of thinking that holds PF to be longitudinal CA in the UV portion of the spectrum re-emitted as a longer wavelength bloom at the sensor surface. Stopping down increases the DOF and minimizes the effect. Steve I was unable to find the reference, but the study was done using narrow band LED lamps where the spectrum was UV deficient. PF was absent until a UV component was added to the spectral mix.

03-09-2015, 05:41 PM	#4
carpents Veteran Member Join Date: Sep 2006 Location: West Chester, PA Photos: Gallery \| Albums Posts: 1,420	Originally posted by stevebrot That looks to be PF. There is one line of thinking that holds PF to be longitudinal CA in the UV portion of the spectrum re-emitted as a longer wavelength bloom at the sensor surface. Stopping down increases the DOF and minimizes the effect. Steve I was unable to find the reference, but the study was done using narrow band LED lamps where the spectrum was UV deficient. PF was absent until a UV component was added to the spectral mix. This is the reference I usually cite: Digital Photography FAQ ---------- Post added 03-09-15 at 08:47 PM ---------- Originally posted by traderdrew Hi Everyone, I was at Daytona Bike Week recently and took my K-5 II with the 16-50mm lens. It's not a lens that I don't usually use but I liked using it there and I really like the lens. Something I noticed about my photos after I took them. I have had a problem with a sort of purple or blue fringing associated with metal parts of motorcycles. It was gone when I adjusted my settings and took another photo. Here are some heavy crops: The one with the fringing was taken at 1/160, ISO 200 at F/2.8. The second one at 1/60, ISO 500 at F/8. I think using F/8 was the contributor to losing the fringing. Any thoughts? I've used a lot of Pentax lenses and find them to be somewhat prone to purple fringing (which is a form of chromatic aberration) wide open. The newer lenses, those released in the last ~3 years or so, are better. Stopping down helps resolve it.

03-09-2015, 06:46 PM	#5
stevebrot Otis Memorial Pentaxian Join Date: Mar 2007 Location: Vancouver (USA) Photos: Gallery \| Albums Posts: 42,007	Originally posted by carpents This is the reference I usually cite: Digital Photography FAQ Cool! Thanks! Steve

03-09-2015, 06:48 PM	#6
Digitalis Site Supporter Join Date: Mar 2009 Location: Melbourne, Victoria Photos: Gallery Posts: 11,694	Originally posted by stevebrot That looks to be PF. There is one line of thinking that holds PF to be longitudinal CA in the UV portion of the spectrum re-emitted as a longer wavelength bloom at the sensor surface. If that was the case the UV light would have to have an intensity nearly four times that of visible light for such a spectral shift to be visible on a digital sensor. I'm not sold on the Ultraviolet hypothesis, the reason being is that borosilicate glass is pretty damn good at attenuating those high frequencies, and also the existence of a UV/IR absorbtion layer in font of the sensor will also soak up residual UV light.

03-09-2015, 07:05 PM	#7
stevebrot Otis Memorial Pentaxian Join Date: Mar 2007 Location: Vancouver (USA) Photos: Gallery \| Albums Posts: 42,007	Originally posted by Digitalis I'm not sold on the Ultraviolet hypothesis, the reason being is that borosilicate glass is pretty damn good at attenuating those high frequencies I understand what you are saying, but I believe it is the near UV light that is in question, plenty of which makes its way to the sensor. As you know, it is that portion of the spectrum that is so problematic for landscape film photography. Originally posted by Digitalis and also the existence of a UV/IR absorbtion layer in font of the sensor will also soak up residual UV light. ...and may be the cause of the problem. I would have to do the experiment myself to see whether there is any truth to it, but the absence when UV is absent followed by presence when UV is present would seem to implicate UV. As for intensity, I guess it would depend on the light. Steve

03-09-2015, 07:32 PM	#9
Digitalis Site Supporter Join Date: Mar 2009 Location: Melbourne, Victoria Photos: Gallery Posts: 11,694	Then you should measure the number of photons emitted by the flash with frequencies of 350-200nm with energies greater than say 3.10ev up to 6.2ev - As far as I'm aware, no one has ever quantified this. There are emission curves for Xenon arc flash tubes but we have no idea of the longevity of emission - it is possible the Ultraviolet pulse from flash is longer than the visible pulse, which would explain the increase of purple fringing with flash.

03-09-2015, 09:39 PM - 1 Like	#10
desertscape Pentaxian Join Date: Feb 2009 Location: Arizona Posts: 1,272	In most cases when this fringing is seen at wider apertures but not when stopped down, it is the result of dispersion and spherochromatism. Colors are going to separate when refracted. The designer must prevent this by using glasses that bring the colors back to focus at the focal plane. However, not all colors can be be corrected and there will be a residual called the secondary spectrum ( in achromatic lenses) and tertiary spectrum ( in APO lenses). Even a highly corrected lens will have a sweet spot in the area of the optic where it is better corrected than in other areas. This is typically in the paraxial (central) and zonal areas (out 20- 60% from the optical axis)of the lens. The marginal zone is usually poorly corrected because the designer knows that the shooter can stop down the lens to truncate the poorly corrected marginal rays. With color fringing, what you are seeing is the uncorrected colors coming from the marginal zone of the lens. One color from different zones of the lens will not focus at the same point. This is the combination of Longitudinal Chromatic and spherical aberration (Spherochromatism).

03-10-2015, 07:52 AM	#11
DeadJohn Pentaxian Join Date: Jan 2011 Location: New York Posts: 4,833	@traderdrew, processing software often has tools for fixing fringing. Lightroom lens correction is one way to do it. My copy of the DA* 16-50 is underwhelming at f2.8. I only shoot f2.8 if it's very dark or I want minimal depth of field. Stopping down to f4.0 makes it a spectacular lens.

03-10-2015, 08:02 AM	#12
Sagitta Veteran Member Join Date: Feb 2011 Location: Maine Photos: Gallery Posts: 3,081	If UV was the culprit, wouldn't adding a UV filter reduce or eliminate it? From everything I've read UV filters are basically only useful as damage protection for lenses on digital cameras.

03-10-2015, 09:44 AM	#13
carpents Veteran Member Join Date: Sep 2006 Location: West Chester, PA Photos: Gallery \| Albums Posts: 1,420	Originally posted by Sagitta If UV was the culprit, wouldn't adding a UV filter reduce or eliminate it? From everything I've read UV filters are basically only useful as damage protection for lenses on digital cameras. UV filters have different cut-offs, they don't perfectly start/stop at the visible wavelength. I understand the same as you for utility of UV filters on digital - they are not sensitive to UV like film is, which is why the filters are less useful. But they're obviously not 100% insensitive to it because people mod their cameras to shoot exclusively UV.

03-10-2015, 10:36 AM	#15
photosbyspeed New Member Join Date: Oct 2013 Posts: 24	Man this post is getting so scientific!! To sum it up, it sucks that the lens has such bad fringing in the OOF areas. Also, the bike isn't in good focus, causing the fringing to be all over your subject! The second image also was not in good focus, eliminating the fringing, but preventing you from creating a quality photo. All this talk about science experiments and adding filters is pretty much a waste of time. Try focus peaking or zooming in on live view to get that sharp focus and minimize that fringing!