New 70-200 2.8L IS II Chromatic abberation

[randallfarhy]

<p>Posted is an image at 100% magnification which has undergone PP in LR3 and DPP (DPP for Chromatic abberation control only). Conditions were high, strong sun and a very dusty environment. Lens had a B+W ($90US) UV filter on it for protection, given the environment it made sense.<br>

There remains purple/blue on the edges where white meets black (number plate is the worst, also uniform) and red around the edges of yellow, not as strong. This is after LR3 reductions, which couldn't be pushed further without introducing further, or even different color shifts in these areas.</p>

<p>I've experienced these abberations before, but have always attributed them to lesser grade glass. My question-is this to be considered "normal" under these conditions and this particular lens? If not, what type of test should I do sans filter and dusty environment to confirm it is the result of environment, or a possibly faulty lens?</p>

<p>I don't want to send something back if it's not defective, but by the same token, I don't want to pay top dollar for a piece of equipment that yields less than acceptable results. Thanks for your time and input.</p><div></div>

[Marcus Ian]

<p>That's pretty noticeable... I suspect the root of the problem is the UV filter. I would see if you can replicate the CA to the same extent under more controlled conditions and see if there is a diff. w/ / w/o the UV filter.<br>

Of course that glare IS pretty harsh anyway. direct sun reflecting off a semi-reflective white contrasted against black... bout the worst possible scene for CA. What was the aperture?</p>

[randallfarhy]

<p>Ian, agree that conditions were not optimal, think the sun may have even been slightly in front of me on that part of the track. Believe this one was shot at F2.8 (that or F4) as I was trying to maintain as much background separation as possible.</p>

[bob_sunley]

<p>I'm not sure moving targets are a good choice for determining if CA is acceptable or not. You've got subject motion, camera motion and an IS system trying to smooth things out plus it looks a little out of focus.Not to mention about the worst lighting and colour scheme thrown in for good measure.</p>

 

[mbkissel]

<p>I've shot hundreds (thousands?) of photos with the 70-200 f/2.8L II. I went back and looked through for any where white meets black. I couldn't find any showing anything like you have pictured. I don't typically use a filter on that lens.<br>

An example...</p>

<div></div>

[mbkissel]

<p>And crop...</p>

<div></div>

[g dan mitchell]

<p>You overexposed for the white areas, which ended up being slightly blown out. When they adjoin darker areas in this way the purple fringing is the normal result, even with excellent lenses.</p>

<p>Dan</p>

[randallfarhy]

<p>Thanks for the responses and thoughts guys, they really are appreciated. Marcus, my apologies for transposing your F/L name.<br /> Mark- thank you, while not the same conditions, this picture gives me a better idea of what the lens is capable of, I will try to find a similar subject/lighting situation this weekend.<br /> Bob, G Dan- what you guys have said makes a ton of sense.<br /> However-I've since visited the websites of two other photographers (one whom was also covering that event, using Canon L Glass) who shoot the same type of imagery- their work doesn't seem to exhibit this problem, even when the shots are blown out worse than mine. I'm not being defensive, whether it's something I'm doing wrong, or gear related, I need to find out what it is to correct it. I'm still leaning toward user error of some sort as 90%+ of issues are attributable to U/E or ignorance.</p>

[mbkissel]

<p>Pull the U/V filter, Randall, as Marcus suggests. I think that might be a big part of what you are seeing. Combine that with Dan and Bob's observations/suggestions and try again. Let us know how that turns out.</p>

[phule]

<p>[[However-I've since visited the websites of two other photographers (one whom was also covering that event, using Canon L Glass) who shoot the same type of imagery- their work doesn't seem to exhibit this problem]]</p>

<p>Most professionals would correct any aberrations in post, wouldn't they?</p>

[mark_allen10]

<p>I only have a 70-200 f/2.8L IS I but I've never seen anything remotely like that. After looking at your image in photoshop and zooming in, it looks to me like those areas really <em>are</em> purple.</p>

<p>Look at the small checkerboard pattern on his shoulder. The purple area above and to the left of the white boxes isn't a purple fringe. It doesn't follow the black white edge of the boxes. It's actually a purple area which ends on a line you can draw along the upper corner of the white boxes.</p>

<p>And the humungous purple area to the left of the white stripes on the shoulder are way too big to be purple fringing. They also just look like part of the same color pattern: white then dark purple then black. I think it's just a shading pattern.</p>

<p>And the "MARSALIS" thing in the lower right hand corner looks to me like a white letter with that same purple shadow behind it.</p>

<p>The purple "areas" are only appearing on some of the white/black edges and they're super wide. There are some white/black edges where there is no purple fringing. I think most of this image is just a purple shading pattern. Maybe you can find a closer picture of the guy to check for sure.</p>

[ken schwarz]

<p>Do you see color fringing like this for areas of the picture that are sharply in focus? I don't think that you can expect perfect correction in out of focus areas.</p>

 

[randallfarhy]

<p>Thanks again, everyone. I've gone back through the images from this shoot (1500+) and this is the only shot I have with this rider. He wasn't in the final rounds.<br>

After considering everyone's thoughts, and reviewing several other similarly lit images, it seems that this issue is a montage of factors which came together all at once for this shot.<br>

The lighting is harsh and angled toward the camera.<br>

The area in question is soft or out of focus.<br>

Re: the Purple fringing:<br>

The uniform does appear to be reflective in nature, and the nameplate may in fact have a purple shadow field/line. While I couldn't find a similar nameplate in other images, I did find a few number plates that appear to have a thin purple shadow, making the numbers appear to float on the black field. (there are also white on black plates that don't appear to float, with no aberrations present ) </p>

<p>There are other images with fringe aberrations of several varieties, but they're either on oof areas OR they disappear with normal sharpening/NR. What isn't shown well here is the red fringe on the yellow fender which blended away in post. It's also oof, so I'm willing to attribute it to that. The filter may well be a contributing factor. I will do some controlled testing today/tomorrow and post some of the results.</p>

<p>Ken, here is a 225% mag crop of the rear fender from the motorcycle positioned next to the one shown earlier- thanks for bringing the point home.</p>

<p>Again, I can't thank you guys enough for taking the time, and making an effort to help me resolve this.</p>

<p>Randall</p>

<p> </p><div></div>

[Robin Smith]

 

<blockquote>

<p>Do you see color fringing like this for areas of the picture that are sharply in focus? I don't think that you can expect perfect correction in out of focus areas.</p>

</blockquote>

<p>Agree completely. Well, you can expect it, but you probably will not get it for any complex zoom or fast aperture prime. Stop worrying about it and take pictures is my advice!</p>

 

[g dan mitchell]

<p><em>[[However-I've since visited the websites of two other photographers (one whom was also covering that event, using Canon L Glass) who shoot the same type of imagery- their work doesn't seem to exhibit this problem]]</em><br>

<em>Most professionals would correct any aberrations in post, wouldn't they?</em></p>

<p>Good point. The reason that I recognize this problem is that... I've had to deal with it on occasion in my own photographs in post. </p>

<p>It is one of the more difficult ones to solve. I use ACR to do my raw conversions. Sometimes you can deal with things like this, or at least reduce them a little bit, with the usual CA correction tools. There is also a "defringing" option on the same page in ACR that will either work only on bright edges or work everywhere. Give it a try, since it can also help.</p>

<p>There are some other tricks you can apply in post, though since your question wasn't how to deal with it in post I won't go down that road right now.</p>

<p>Dan</p>

[randallfarhy]

<p>This is a test shot (chosen for it's range of contrasts, textures and element depth) taken in similar high contrast light.</p>

<p>Full frame, 200mm, ISO200 F2.8 @1/5000 (slightly overexposed to pop highlights as in the motorcycle shot). UV filter in place. Focus is on a wndow frame edge as to render some OOF areas on the a/c units. Minimal post work, sharpening and N/R only, no contrast, tonal or midrange work. NO C/A corrections made. Note in the slightly oof grill section to the bottom left, as in the motocross shot CA is noticeable on the edges @100%, and the overall image seems to exhibit a very slight red shift. Thoughts?</p><div></div>