View Full Version : What is this and why is it happening
Dylfish
11-02-2009, 7:18pm
hey guys, going over some photos i took the other week and i zoomed in to 100% then 400% (see screenie) and there is a purple haze bordering the subject. this was taken @ 4pm with the sun slightly infront of me with no lens hood if that helps? what is it and how can i avoid it in the future, i realised alot of that set of photos had the same issue.
thank you
Looks like chromatic aberration to me (CA)
Usually the result of harsh lighting with cheaper lenses
ricktas
11-02-2009, 7:27pm
Its called Chromatic Aberation (CA), some lenses are more prone to it than others. It occurs on the boundaries between high contrast areas, so in your case, between the whites of the top, and the darker background/shadows.
It can be corrected in Photoshop, Filter, Distort, Lens Correction. However any correction can only do so much.
You might find with some testing you can determine what causes it most with your lens and then avoid that situation.
MrJorge
11-02-2009, 7:27pm
Unfortunately that is the dreaded CA or Chromatic Abberation.
http://en.wikipedia.org/wiki/Chromatic_aberration
Some lenses suffer from this at various aperture and focal lengths.
There are a few ways to deal with it but I have never done it so I'll leave it to others to make their suggestions.
Dylfish
11-02-2009, 9:07pm
well that was at the lenses furtherest focal length and the most open apeture. i havn't noticed it before on other photos so could it have been the position of the sun and the lack of a lens hood?
ricktas
11-02-2009, 9:12pm
well that was at the lenses furtherest focal length and the most open apeture. i havn't noticed it before on other photos so could it have been the position of the sun and the lack of a lens hood?
Combination of several factors. One, some lenses are known to be more prone to CA than others. focal length and aperture are less likely to have an impact than using a lens that is known to produce CA when there are sharp ranges in contrast -really bright subject up against a darker one. It can often be seen in landscape shots, around the leaves of a tree on a bright sunny day. Often the edge of the leaves have highlights on them from the sun's reflection, creating a great opportunity for CA to rear its ugly head.
So, yes the sun can effect the result, cause it can create more areas of the photo with bright areas, butted against dark ones. A lens hood probably would not have made any difference as CA is caused against the subject, not based on sun hitting the lens.
You also have the double whammy there as well, as there is red CA on the helmet.
The question you need to ask yourself though, what does the image look like when viewed normally?
MrJorge
11-02-2009, 9:16pm
This review seems to imply it's a known feature of this lens.
http://www.photozone.de/canon-eos/203-canon-ef-100-300mm-f45-56-usm-lab-test-report--review?start=1
Work with CS3 etc to figure out your best defense against it is my suggestion.
arthurking83
11-02-2009, 10:23pm
More specifically it's called Purple Fringing, and is the easier of some of the other CA's to minimise(there are other types like green and red and yellow)
With purple fringing you simply need to avoid harsh light, high contrast scenes and stop the lens down.
it's usually a sign of slightly cheaper lenses.... but just about all lenses do it to varying degrees more expensive lenses do it in far lower quantities.
best way to minimise it is to use a smaller aperture(larger number) and compensate with shutter or ISO to a lesser degree.
The PF usually happens in very bright sunlight which is hard/harsh.... very rarely in lower softer incandescent light.. well... none of my lenses do it!
I think flash can also produce PF in highlights but as they are blurred it's not an issue.
if you check your exif data and judging by the pixel width of that PF on the batsmans shirt, I'd guess that the lens was wide open or 1/3 of a stop down.
And probably a consumer grade lens as well, genuine or third party.... it doesn't matter! They all do it at that price level.. and some more so than others. The higher end pro $2k lenses with their multiple coatings and super exotic glass construction will do it mush less than you see here.
In this instance the PF is probably about 5 pixels in width at the shirt edge, and a great lens may only produce 1/2 pixel of PF.
If you stop the lens's aperture down by a couple of stops you will(should) notice an improvement, but not totally eliminated.
You shouldn't see it in lower light levels or softer light such as later in the afternoon.
One thing I've noticed is that it seems to be affected by the colour of the light.. bluer light as you get at midday seems to accentuate the hideous apparition more so than softer warmer yellow light as in later afternoon. Same camera, lens and settings, different light conditions seem to determine if it's either there or not for me.
It also happens to a higher degree at the long end of many super zoom lenses like 18-xxx types.
ps. while stopping the lens down is usually the best answer to reducing this PF phenomenon, it's not always the case with every single lens!
I know of one particular lens that I liked a lot(and had) that got worse as you stopped it down. Very annoying too but once you know the conditions at which it happens you quickly learn to working around it.
Curious if this was using the 100-300mm lens in your sig, and at what settings.. focal length, ISO shutter, aperture.. etc
pps. did the image actually look as bad at full view(that is being able to view the entire scene on your screen)? At 100% pixel view it will always look bad if it doesn't look so bad at full view then using PS's CA correction tools. If there's a specific one for Purple Fringing then use that.
The red fringing that someone noted is only red due to the colour of the helmet.
There are two other types of fringing that can occursimultaneously, and that's red and green fringing (LoCA's) and they are caused by focusing or lack of it. You see them fore/aft of a specific focal point.
That's not the same thing as this PF issue.
If you check the Photozone site, and look through the results for your lens you will notice a graph of CA's values. You may notice that at some point9maybe F8, maybe f/11 :confused:) you can get the CA's down to much more reasonable levels.. use that as a guide to getting the most out of your lens.
:)
Dylfish
11-02-2009, 11:47pm
yeah it was wide open, @ 100% its noticable but not the killer it seems @ 400% (obviosuly) are there any other methods to reduce it in PP other than lens correction in PS?
thanks for all the replies!
Dylfish
12-02-2009, 12:18am
full pic is the 5th post down here
http://www.ausphotography.net.au/forum/showthread.php?t=24992
its not majorly noticable unless you go looking for it
are there any other methods to reduce it in PP other than lens correction in PS?
You can always use the stand-alone programs PFree and CAfree (download here (http://wareseeker.com/publisher/tom-fiddaman/118172/), the original page is no longer available). Or remove the color fringe using selective color.
Edit: I just noticed the named software points to the original page (that has been removed). Some googling might turn them up somewhere else.
arthurking83
12-02-2009, 8:36am
I think I've actually tried to process it out once the old fashioned way(the Tammy 17-50/2.8 does it too in the same lighting conditions!).
What I did, was to 'paint a selection' using the brush tool all along the seriously affected areas using approximately a pixel diameter just a little bigger than the actual PF lines.
Paint as far as you can in small steps and then reduce the magenta colour balance to compensate.
Don't over do it.. just minimise it!
Damned tedious if there's lots of it, and it's easier to deal with if you stop the lens down, or avoid such high contrast scenes, or....... purchase more expensive lenses :p
Dylfish
12-02-2009, 9:28am
i plan all 3 :)
Nicholas N
12-02-2009, 11:38am
There is no point in ever viewing any image at 400%. Sometimes I can't understand why PS even lets you zoom in that far.
If you can't notice it at 100%, then it won't make much difference. If you feel the need for your own satisfaction to remove it just use the PS corrections.
Dylfish
12-02-2009, 4:49pm
its noticalble just not to that extent. zooming @ 400% makes it easier to select pixels for corrections i find
its noticalble just not to that extent. zooming @ 400% makes it easier to select pixels for corrections i find
Exactly :)
smorter
13-02-2009, 1:35am
A lot of Canon's most expensive Prime lenses (without CaF2 or Super UD or even UD elements - almost all super telephotos have CaF2 elements) are particularly susceptible.
As Arthur said, cheaper lenses do tend to exhibit it, but as he said it is usually but not always the case, there are numerous exceptions.
This is a AUD $3000 (at this time) lens at 100% :
http://dawei.zenfolio.com/img/v0/p433878869.jpg
http://dawei.zenfolio.com/img/v6/p209951947.jpg
Aspherical fast primes tend to display longitudinal CA at wider apertures, and these are impossible to correct for. Lateral CA is much easier to correct for.
Nicholas N
13-02-2009, 7:25am
Low Dispersion elements are basically a standard now aren't they?
Just about every lens nowadays has good CA countering attributes.
Is that the 50/1.2, Snorter? It seems to get a lot of criticism for this (and a few other) issues, though it has some very dedicated fans as well.
arthurking83
13-02-2009, 7:47am
Is that the 50/1.2, Snorter? It seems to get a lot of criticism for this (and a few other) issues, though it has some very dedicated fans as well.
99.9% most likely is, and Nikons fast glass does it too.
85/1.4, my 50/1.2 ... etc.
Someone mentioned the coatings, which are supposed to help(Nikon uses ED nomenclature) but they even those lenses still produce PF.
I suppose one reason to have the fastest possible glass is that you can stop it down by one stop or two to minimise the effect, eg a f/1.2 lens stopped down to f/2 probably won't be as affected as a f/1.8 lens will be at the same aperture.(eg. the Nikon 50/1.2!)
purephotos
26-02-2009, 8:13pm
:o I found this happening with my good Nikkor 80-200mm f2.8 the other day at a wedding. Not happy!
-R
Powered by vBulletin® Version 4.2.3 Copyright © 2024 vBulletin Solutions, Inc. All rights reserved.