All Forum Comments
Comment by ChrisA posted on Horrible purple cast, not a new species of robin at 21/03/2007 - 23:30
Quote:
Yes, I know. I agree it's not ideal. I'll experiment with the other techniques when I get a chance; meanwhile I'll save up for some Pentax glass...ChrisA wrote:
Haven't tried the plug-ins yet, but a simple technique, described at:
http://www.great-landscape-photography.com/remove-purple-fringing.html
...works reasonably well in Photoshop Elements 5.
But it only works, if you do not have any REAL purple or similar colors in your image. In that case you are eliminating that, too!Haven't tried the plug-ins yet, but a simple technique, described at:
http://www.great-landscape-photography.com/remove-purple-fringing.html
...works reasonably well in Photoshop Elements 5.
Comment by ChrisA posted on Horrible purple cast, not a new species of robin at 21/03/2007 - 08:48
Haven't tried the plug-ins yet, but a simple technique, described at:
http://www.great-landscape-photography.com/remove-purple-fringing.html
...works reasonably well in Photoshop Elements 5.
I expect I could do better with a bit more care, but here's what just following the instructions in the link yields:
It's obviously not as good as not getting the fringing in the first place, but for salvaging disasters it's not too bad, it would seem.
http://www.great-landscape-photography.com/remove-purple-fringing.html
...works reasonably well in Photoshop Elements 5.
I expect I could do better with a bit more care, but here's what just following the instructions in the link yields:
It's obviously not as good as not getting the fringing in the first place, but for salvaging disasters it's not too bad, it would seem.
Comment by ChrisA posted on Horrible purple cast, not a new species of robin at 20/03/2007 - 23:39
Quote:
I'm not sure whether Chris' lens was a DSLR specific lens
It isn't. It's a manual-focus Sigma 400mm F5.6 lens, PK/A fit.I'm not sure whether Chris' lens was a DSLR specific lens
This one looks identical:
http://cgi.ebay.co.uk/Sigma-400mm-F5-6-Telephoto-Pentax-Fit_W0QQitemZ16009659908...
Comment by ChrisA posted on Horrible purple cast, not a new species of robin at 19/03/2007 - 16:39
Quote:
It's purple fringing for sure, caused by Chromatic (colour) abberation.
So does it happen only with the strongly backlit shots, in the very contrasty areas, simply because there's more light to end up where you don't want it?It's purple fringing for sure, caused by Chromatic (colour) abberation.
Comment by ChrisA posted on Horrible purple cast, not a new species of robin at 19/03/2007 - 10:12
Just to muddy the waters a bit, here's another extract... from one of the links at the Wikipedia CA page.
http://www.vanwalree.com/optics/chromatic.html
It's testable though, if, as he says, "axial colour is cured by stopping down the lens". Although the bad shot was taken at F8, though - so I wonder how far it needs to be stopped down.
Still don't understand the "fringing is only observed in places with a high contrast" though.
http://www.vanwalree.com/optics/chromatic.html
Quote:
Longitudinal and lateral chromatic aberration can both give rise to colored edges, but properties are different. Axial color causes fringes all around objects, whereas lateral color only affects tangential details. Axial color can occur at any position in the image, whereas lateral color is absent in the image center and progressively worsens toward the image corners. Axial color is cured by stopping down the lens, whereas lateral color is present at all apertures. Chromatic aberration can manifest itself as a mixture of axial and later color, in which case it is partly cured by the use of a small aperture. Axial color yields fringes of a single color for an "in-focus" object (Fig. 3), whereas lateral color delivers two differently colored fringes at either side of the tangential structure (Fig. 4). Fringes due to axial color are particularly clear in parts of the scenery that are (just) out of focus, e.g. purple in front of the subject in sharp focus and green behind it, or vice versa. Although color fringing is only observed in places with a high contrast, both the longitudinal and transverse chromatic aberration degrade the overall resolution of a lens.
Reading this (my emphasis) strongly suggests axial CA - I have the "purple in front and green behind" that he mentions.Longitudinal and lateral chromatic aberration can both give rise to colored edges, but properties are different. Axial color causes fringes all around objects, whereas lateral color only affects tangential details. Axial color can occur at any position in the image, whereas lateral color is absent in the image center and progressively worsens toward the image corners. Axial color is cured by stopping down the lens, whereas lateral color is present at all apertures. Chromatic aberration can manifest itself as a mixture of axial and later color, in which case it is partly cured by the use of a small aperture. Axial color yields fringes of a single color for an "in-focus" object (Fig. 3), whereas lateral color delivers two differently colored fringes at either side of the tangential structure (Fig. 4). Fringes due to axial color are particularly clear in parts of the scenery that are (just) out of focus, e.g. purple in front of the subject in sharp focus and green behind it, or vice versa. Although color fringing is only observed in places with a high contrast, both the longitudinal and transverse chromatic aberration degrade the overall resolution of a lens.
It's testable though, if, as he says, "axial colour is cured by stopping down the lens". Although the bad shot was taken at F8, though - so I wonder how far it needs to be stopped down.
Still don't understand the "fringing is only observed in places with a high contrast" though.
Comment by ChrisA posted on Horrible purple cast, not a new species of robin at 19/03/2007 - 07:32
Ok, guys, I have a couple of supplementary questions about this problem.
I've cropped the top robin from my earlier shot; here it is:
I'm particularly interested in a couple of bits from Liam's link here:
Firstly this:
I've also split out the RGB channels of this crop, with Paint Shop Pro 7, and superimposed the three. Cycling through the 3 channels reveals no movement of the image. If it's CA causing the different colours to be simply focused in different places, wouldn't you expect the RGB channels not to be in the same place?
Towards the bottom of that article, he discusses some purple fringes which he says are not CA, but are very similar to mine, and, like mine, are in areas of relative darkness next to overexposed sky. On my other shots, where the robin is not against a strongly lit background, I don't get the fringes.
So I'm wondering if this phenomenon isn't simply a lens fault after all.
Thoughts?
PS - note to Mannesty. Did you ever notice anything like this with, ahem, your old 400mm Sigma lens?
I've cropped the top robin from my earlier shot; here it is:
I'm particularly interested in a couple of bits from Liam's link here:
Firstly this:
Quote:
Digital photography introduced some new types of color fringes, that are not lens chromatic aberration. They are frequently confused with TCA and unfortunately they occur often mixed with TCA. These effects might be visible as purple or blue fringes and are visible around overexposed areas in most cases.
They might have different causes:
* By sensor overflow visible as blooming. This blooming often has a purple color because of the bayer pattern of the sensor: If a sensor cell overflows to the neighboring cells all cells are affected the same. There are 50% green but only 25% red and 25% blue sensor cells. Hence blue and red are weighted more resulting in purple fringes around overexposed areas."
(... and other causes, none of which lens-related)
Also this (of transverse CA):Digital photography introduced some new types of color fringes, that are not lens chromatic aberration. They are frequently confused with TCA and unfortunately they occur often mixed with TCA. These effects might be visible as purple or blue fringes and are visible around overexposed areas in most cases.
They might have different causes:
* By sensor overflow visible as blooming. This blooming often has a purple color because of the bayer pattern of the sensor: If a sensor cell overflows to the neighboring cells all cells are affected the same. There are 50% green but only 25% red and 25% blue sensor cells. Hence blue and red are weighted more resulting in purple fringes around overexposed areas."
(... and other causes, none of which lens-related)
Quote:
Color fringes should be of complementary color (red-cyan, blue-yellow or purple-green) on opposite sides of a dark or bright area.
As you can see in my shot, the fringe is purple, and all the way round the robin.Color fringes should be of complementary color (red-cyan, blue-yellow or purple-green) on opposite sides of a dark or bright area.
I've also split out the RGB channels of this crop, with Paint Shop Pro 7, and superimposed the three. Cycling through the 3 channels reveals no movement of the image. If it's CA causing the different colours to be simply focused in different places, wouldn't you expect the RGB channels not to be in the same place?
Towards the bottom of that article, he discusses some purple fringes which he says are not CA, but are very similar to mine, and, like mine, are in areas of relative darkness next to overexposed sky. On my other shots, where the robin is not against a strongly lit background, I don't get the fringes.
So I'm wondering if this phenomenon isn't simply a lens fault after all.
Thoughts?
PS - note to Mannesty. Did you ever notice anything like this with, ahem, your old 400mm Sigma lens?
Comment by ChrisA posted on Horrible purple cast, not a new species of robin at 18/03/2007 - 23:29
Quote:
It is called chromatic aberration.
...
Seriously, this is a lens fault, and better lenses will suffer less from it.
Thanks for this, guys.It is called chromatic aberration.
...
Seriously, this is a lens fault, and better lenses will suffer less from it.
If I understand it, then:
- in this case, it's the light at the blue end of the spectrum in the bright white background that is ending up where it shouldn't.
- it's worse with a very bright background because there's more of the blue to stray into areas where it'll be noticeable.
Is this right?
Comment by ChrisA posted on Horrible purple cast, not a new species of robin at 18/03/2007 - 13:18
Got a pair of robins nesting in my garden, so had a go at capturing them yesterday with the K10D
Some of the shots came out ok (albeit cheesy and totally unoriginal) - like this one:
Sigma 400mm, 1/180s at F5.6, spot metered.
But others had a horrible purple cast around the edges, like this one:
It was also spot-metered, 1/350s at F8.
I'm guessing that it's got something to do with the very overexposed sky as background, but I'd be grateful if anyone can explain what's going on - and more important, how to avoid this sort of thing in future.
Thanks!
Some of the shots came out ok (albeit cheesy and totally unoriginal) - like this one:
Sigma 400mm, 1/180s at F5.6, spot metered.
But others had a horrible purple cast around the edges, like this one:
It was also spot-metered, 1/350s at F8.
I'm guessing that it's got something to do with the very overexposed sky as background, but I'd be grateful if anyone can explain what's going on - and more important, how to avoid this sort of thing in future.
Thanks!
Comment by ChrisA posted on Horrible purple cast, not a new species of robin at 18/03/2007 - 12:43
Quote:
Someone suggested that Pete take more shots and select the best for posting, I would disagree to a point. I think trying to get it right with only a few shots is very commendable, filling a memory card and getting the shot of a lifetime is all well and good, but if you cant recal taking it or what was going through your mind at the time then you aren't going to learn much. Only taking a few shots and being able to remember the way you shot them makes it easier to evaluate your work later and remember what NOT to do next time.
Let me clarify...Someone suggested that Pete take more shots and select the best for posting, I would disagree to a point. I think trying to get it right with only a few shots is very commendable, filling a memory card and getting the shot of a lifetime is all well and good, but if you cant recal taking it or what was going through your mind at the time then you aren't going to learn much. Only taking a few shots and being able to remember the way you shot them makes it easier to evaluate your work later and remember what NOT to do next time.
I'm not advocating taking a hundred shots at random, and picking the best few in the unlikely event of there being a few decent ones in the pile.
However, endlessly fiddling with a couple of duds (sorry, I just hate the top, Ok?) in Photoshop or whatever he's using is equally unlikely to result in any improvement in photo taking ability.
FWIW, and as I've said, this is just my tuppence worth as a total non-expert, my feeling is that Pete needs to get more practice seeing when he's looking.
In my experience, seeing is hard to start with. I had to take a lot of rubbish shots before what I got on the photo corresponded to what I saw in the viewfinder.
Now that I get more or less what I see, I can set about getting what I intend... and that's even harder.
Comment by ChrisA posted on 2 photos Of my friend Jenna Using natural light with the K10 at 20/03/2007 - 22:18
Has anyone experienced this purple fringing thing with film, rather than digital? Everything I read seems to focus on digital.
Just curious. But since taking up SLR photography in 1984 with the Super A and ME Super, and quite a few not-at-all-Pentax-SMC-lenses, I never saw it with film.
Not even once.