[Brightcolours]

The purple would occur in a not focussed area, and here the purple is VERY extreme in the IN focus area. Also, we don't see a purple edge, but purple bleeding into the dark area from the very bright area.



LoCa will occur just because of being OOF. This kind of PF occurs ONLY with a lot of light. While to the casual viewer it may look all the same (purple artifacts), the phenomenons are just very different.



PF is being combatted by different lens coatings (specifically the back element), which for instance Sigma introduced with the DG and DC series, and Tamron when they started to introduce Di and Di II lenses.



As far as I know LoCA can only be prevented by making the lens design really apochromatic.

[Guest]

Hi Chaps,

All this technical stuff is incredibly fascinating ... but does anyone actually know how to "easily correct" it?



I would guess not in case like this, since there is a major loss of original colour/contrast information on the affected area. When the fringing area is this large, reducing purple saturation will leave an unnatural black/grey halo, and transforming the image to B&W likewise results in a strange grey halo on the nose.





Ian, it looks like you are using a very small part of an image - do you need to crop it so much for real-life usage, or did you only do that to demonstrate the problem?

I ask because this problem often occurs with stage performances, and the best way around it is ultimately to get closer to your subject or use a longer lens - this won't 'cure' it, but it will make the affected area less significant in relation to the size of your image. Also, try experimenting with this type of high contrast/sillhouette shot until you find the optimum settings/situations for you equipment (for example, dark background light subject or subject completely 'blacked out').

[Brightcolours]

[quote name='Pinhole' timestamp='1308139898' post='9257']

Hi Chaps,

All this technical stuff is incredibly fascinating ... but does anyone actually know how to "easily correct" it?



I would guess not in case like this, since there is a major loss of original colour/contrast information on the affected area. When the fringing area is this large, reducing purple saturation will leave an unnatural black/grey halo, and transforming the image to B&W likewise results in a strange grey halo on the nose.





Ian, it looks like you are using a very small part of an image - do you need to crop it so much for real-life usage, or did you only do that to demonstrate the problem?

I ask because this problem often occurs with stage performances, and the best way around it is ultimately to get closer to your subject or use a longer lens - this won't 'cure' it, but it will make the affected area less significant in relation to the size of your image. Also, try experimenting with this type of high contrast/sillhouette shot until you find the optimum settings/situations for you equipment (for example, dark background light subject or subject completely 'blacked out').

[/quote]

You can correct (La)CA, if the lens has a gradual CA character and one of the offending CA colour is a primary colour.

You can not correct PF, but you can mask it. select part of the image that shows PF, and use the replace colour option to change it to whatever suits the area. Pretty simple.

[Dick England]

When we check the bokeh fringing shots in the more recent Photozone tests of fast lenses, we find plenty of purple bleeding into black areas we would consider to be in focus. The purple only disappears toward the back of the DOF.

[Brightcolours]

[quote name='Dick England' timestamp='1308142855' post='9260']

When we check the bokeh fringing shots in the more recent Photozone tests of fast lenses, we find plenty of purple bleeding into black areas we would consider to be in focus. The purple only disappears toward the back of the DOF.

[/quote]

When we make a photo of a page of paper with black numbers/letters totally in the plane of focus, we do not see such purple "bleeding" even when the lens is wide open. You simply have little understanding of what you are looking at.

[wim]

There are actually 3 types of aberrations in this picture.



1) PF. Those are the violet, bright purple haloes around the overexposed highlights on the metal parts of the instrument. In principle, this is not really true Chromatic Aberration, it is a mix of things, very likely, although it normally only happens at large apertures, and is much more distinct with digital than with film, so there likely is a sensor or sensor circuitry element involved as well. Generally, these can't be undone, because information is lost. This seems, effectively anyway, to be a weird overexposure effect. PF can occur in and out of the focus zone, but is most clear in and close to the focus zone.



2) Spherochromatism, also called longitudinal chromatic aberrations, or LoCA. These only occour in out of focus area, generally reddish in front of the focused area, and greenish behind the focused area, wherever there are very hard contrast transitions. In the picture you can see this in the background, at the transition of the grey and white parts, and on the instrument itself, but those are only tiny patches, and they also overlap partly with the PF. It's best seen almost at the bottom of the photograph, in the centre almost, where just above a darkish part you can see a reddish edge overlayed on the gold colour of the instrument, which towards the left and top mixes with PF. LoCA can be corrected, but it is a lot of work to do manually, as you need to do a lot of masking, layering, and colour corrections. There is a piece of software nowadays, which can handle LoCa, but I can't remember which it is now. LoCa is caused by residual spherical aberrations, hence the official term of Spherochromatism, which term these days almost nobody seems to use anymore. As indicated, this can be corrected by very good (super)apochromatic designs and to a large degree by aspherical elements (although those are normally really best at or around a specific aperture). However, aporchromatic lenses normally have relatively small apertures, and therefore suffer less from spherical aberrations in the first place. These spherical aberrations at large apertures also provide good bokeh, generally speaking, so from that POV it is a good and a bad thing at the same time.



3) Latitudinal Chromatic Aberrations (blue and yellow most often), which also occur in hard contrast transitions, essentially the blue parts in nose and the upper part of the instrument. Some PF is mixed in here as well. LCAs normally occur in in-focus areas. These can also be corrected to a large degree - most software can handle this automatically these days, although this is a rather severe case of these aberrations (plus some PF mixed in). There is quite a large transitional zone here with PF too.



I'd suggest that if this bothers you really a lot, you better convert this to B&W if you want to get a reasonable picture. That's aso the fastest way of dealing with these problems. See here, just a simple conversion (10 seconds of work):



[ATTACHMENT NOT FOUND]



HTH, kind regards, Wim

[popo]

[quote name='Brightcolours' timestamp='1308154027' post='9261']

When we make a photo of a page of paper with black numbers/letters totally in the plane of focus, we do not see such purple "bleeding" even when the lens is wide open. You simply have little understanding of what you are looking at.

[/quote]

It simply isn't bright enough to make the fringing obvious on the focal plane. I think I can come up with a quick and simple way to demonstrate that...







Both shots taken using the Canon 85mm f/1.8 wide open. Right one is exposed about +3 stops compared to the left one by varying shutter time. Focus was manually set to make the line beginning "Earth" sharpest in the left shot, and was not touched again. Note in the higher exposed shot the onset of purple appearing on what was the best focus, and is even apparent on the line behind it.

[Brightcolours]

[quote name='popo' timestamp='1308162350' post='9264']

It simply isn't bright enough to make the fringing obvious on the focal plane. I think I can come up with a quick and simple way to demonstrate that...







Both shots taken using the Canon 85mm f/1.8 wide open. Right one is exposed about +3 stops compared to the left one by varying shutter time. Focus was manually set to make the line beginning "Earth" sharpest in the left shot, and was not touched again. Note in the higher exposed shot the onset of purple appearing on what was the best focus, and is even apparent on the line behind it.

[/quote]

I said "such bleeding", not that one can not detect still some purple from OOF areas on "in focus" areas.



Try shooting a sheet IN THE FOCAL PLANE, not at such an angle. Do it is bright as you want. It will NOT give similar purple bleeding into the dark from the bright as the original poster's image shows. It is not LoCA.



I am sure that you can actually see the difference of what occurs in the OP's image and the LoCA test page, even though both are kinds of purple artifacts.

[Brightcolours]

[quote name='wim' timestamp='1308161822' post='9263']

3) Latitudinal Chromatic Aberrations (blue and yellow most often), which also occur in hard contrast transitions, essentially the blue parts in nose and the upper part of the instrument. Some PF is mixed in here as well. LCAs normally occur in in-focus areas. These can also be corrected to a large degree - most software can handle this automatically these days, although this is a rather severe case of these aberrations (plus some PF mixed in). There is quite a large transitional zone here with PF too.

[/quote]

Wim, La CA usually is green and magenta, or red and green-blue, the yellow and blue variation is relatively rare. LaCA affects the whole image, except the center of the image. The further you go to the edge, the stronger the LaCA will get. It messes up the image's contrast and sharpness, and in contrasty transitions you will see on one side of a darker subject magenta, on the other side green. Or, yellow and blue. Or red and blue-green. The colour depends on which part of the spectrum gets "bent" more or less than the rest of the spectrum, so which colour image has a larger or smaller projection than the rest of the spectrum. If red is the offending part of the spectrum, you will see red and blue-green edges. If blue is the "problematic" part, you will see blue and yellow, And if green is the offending colour, you will see magenta and green.



CA, usually, is correctable, if the optics show a nice and gradual CA shift. It also depends on the "purity" of the offending part of the spectrum, because the software I know only corrects for the different colour channels (R, G or B ). The correcting works by making the image or either the red, blue or green channel larger or smaller, till the size corresponds to the other 2 channels.



There is also software that does not correct, but rather masks, CA by desaturating just the coloured edges. Most in-camera "CA-correction" works that way. This is a less desirable method.

[Brightcolours]

This is the result of a minute or 5 making selections (simple lasso selections of whole areas that need a color spectrum to be replaced with all the same colour) and doing "Replace color..." in Photoshop: