sharpness and CA on 85mm f/1.2

[Jon Ruyle]

I'm posting a few crops of pics taken with my 85 f/1.2- I'm interested to hear from other users of this lens if the sharpness/CA are within the range of what should be expected from this lens (Keith- I'm particularly interested in your opinion)


The
sharpness looks pretty good to me (at least, for the tiny part of the
pic that is in focus )


[img]/resized-image.ashx/__size/550x0/__key/CommunityServer-Discussions-Components-Files/8/6825.IMG_5F00_2008.JPG[/img]


Note, though, the obvious CA:


[img]/resized-image.ashx/__size/550x0/__key/CommunityServer-Discussions-Components-Files/8/4571.IMG_5F00_1782_2D00_1.JPG[/img]


Of course, glints of sunlight are pretty tough, but is this worse than usual?


Light bulbs at night are tougher still; here is another 1-1 crop wide open:


[img]/resized-image.ashx/__size/550x0/__key/CommunityServer-Discussions-Components-Files/8/1057.IMG_5F00_1870_2D00_1.JPG[/img]


Now imagine trying to take pictures of stars at night with this thing without stopping down


This is a lot of CA, but is it more than expected? The
50 f/1.4 is almost as bad, but not quite. The CA on either lens goes
down significantly (almost dramatically) when stopped down to f/2. And even wide open, I see no CA in most situations. This is also shot @ f/1.2:


[img]/resized-image.ashx/__size/550x0/__key/CommunityServer-Discussions-Components-Files/8/0636.IMG_5F00_1562.JPG[/img]


And finally, the surest way to avoid CA (@ f//1.2 again):


[img]/resized-image.ashx/__size/550x0/__key/CommunityServer-Discussions-Components-Files/8/1323.ching-bw.JPG[/img]

[Keith B]

Hey Jon


Since the shot of the lights look straight on, I'm going to say that is a lot of CA. I took some shots of light to show. I was able to make some heavy CA but I really had to try.


This was trying to achieve CA. This is the worst I have seen with my 85L.


[img]/resized-image.ashx/__size/800x0/__key/CommunityServer-Components-UserFiles/00-00-00-25-81/_5F00_MG_5F00_0941crop.jpg[/img]


[img]/resized-image.ashx/__size/800x0/__key/CommunityServer-Components-UserFiles/00-00-00-25-81/_5F00_MG_5F00_0941sm.jpg[/img]


Initial shot without trying to get CA


[img]/resized-image.ashx/__size/800x0/__key/CommunityServer-Components-UserFiles/00-00-00-25-81/_5F00_MG_5F00_0944sm.jpg[/img]


(I need to dust. That cobb web isn't nearly as visable in person.)





The sharpness in your eye lash shot looks a little soft compared to my experience with the 85L, but it is tough to tell because of the focal plain.


What body are you shooting with? 5DII? Obviously if you have MA you've done them right? I know dialing mine in wasn't so cut and dry but I think I have it dead on. It took a few attempts though.


I think CA has a lot to do with particular focal length. I find my 50 1.2 has a lot less CA than the my 85 and my 35. So I think the fact the 50s don't manipulate perspective, the light doesn't get as bent.

[neuroanatomist]

Unfortunately, CA seems to be a 'feature' of wide aperture and/or wide angle lenses. But, CA means more than one thing (and I'm not talking about the state where I was born and raised!). There is lateral CA and also longitudinal CA (aka 'bokeh fringing'). Both show up at high contrast transitions and especially around specular highlights. Lateral CA affects wide angle lenses, is worse at the edges of the frame and absent from the center, and is not effectively reduced by stopping down the lens. Longitudinal CA affects wide aperture lenses, can occur anywhere in the frame but is manifest in out-of-focus areas, and is effectively reduced/eliminated by stopping down the lens.

I think CA has a lot to do with particular focal length. I find my 50 1.2 has a lot less CA than the my 85 and my 35. So I think the fact the 50s don't manipulate perspective, the light doesn't get as bent.

I think this is partially correct, Keith. Lateral CA is related to focal length, more specifically to angle of view. The amount that light is bent (refracted) is proportional to the angle of view, since light from the entire field has to be bent to fit in the image circle of the lens. So of those three lenses, your 35L bends the light most, and your 85L bends light the least, with your 50L somewhere in the middle. Longitudinal CA, on the other hand, doesn't depend on focal length - but it is related to spherical aberration, and it may be that the intentionally undercorrected SA that produces the superior bokeh of the 50L also reduces it's longitudinal CA, so coupled with the lack of lateral CA at a normal focal length, it's the best of the three.


So, I think the primary effect we're seeing with the 85L is longitudinal CA. That usually shows up as purple fringe in front of the focal plane and green fringe behind the focal plane. Below are composites of some test shots I took this morning, to show the effect. These were shot with the 5DII and 85L, focused manually with 10x LiveView. The 4 eneloop batteries are each about 3" apart, and the specular highlight is from the chrome ball of a SpyderCube with an LED Maglite shining on it from about 15" to the left (just out of the frame). The SpyderCube is positioned close to (but not exactly at) the same depth as the second battery. The only post-processing was a Kelvin white balance.


The first composite shows larger views of the f/1.2 images, focusing on each battery in order from front to back.


[img]/cfs-file.ashx/__key/CommunityServer-Components-UserFiles/00-00-00-35-15/CA-composite_2D00_wide-open.jpg[/img]


You can see that the CA at the chrome ball is green in the top left image, purple in the bottom two images, and not very evident when the ball is in/near the plane of focus (top right). Try to ignore the bluish cast to the reflection in the chrome ball - that's from the bluish light of the LED Maglite, which you can also see illuminating the first three batteries and the surface of the dresser.


The second composite are near-100% crops of the chrome ball atop the SpyderCube, alsofocusing on each battery in order from front to back, with the indicated aperture.


[img]/cfs-file.ashx/__key/CommunityServer-Components-UserFiles/00-00-00-35-15/CA-composite_2D00_crops.jpg[/img]


So, even at f/1.2 the CA for the in focus part of the image is not too bad (certainly much better than the OOF regions, especially when you consider that only the front of the chrome ball is in focus). Of course, at f/1.2 there can be a lot of an image that's not within the depth of field!!


It also seems to me that the front-of-focal-plane CA (green, top left image) is much less objectionable than theback-of-focal-plane CA (purple, top right image). So, you can reduce the impact of CA if you can compose the image with the specular highlights and sharp contrast transitions in the background rather than in the foreground.

Now imagine trying to take pictures of stars at night with this thing without stopping down

But, have you tried it? As much as I find it distasteful to agree with anything Ken Rockwell may write, stars are effectively at infinity focus, meaning minimal effects from longitudinal CA even wide open. It might be that the 85L would be fine wide open for astrophotography, and in some ways superior to a wide-aperture wide angle lens(since transverse CA often plagues wide angle lenses and is not reduced/eliminated in the in-focus part of the image).


Hope some of this is useful...


--John

[Daniel Browning]

Great thread! Thanks for the really informative posts, Jon, John, and Keith.

[Richard Lane]

Nice contribution guys!


Neuro, remember when I took a few shots of the Sailboat with the CA on the water and then the CPL filter removed the CA, and at that time you had suggested that the CPL, may be good at removing specular highlights. I was wondering if you have tried that at all since then and if you thought it was relevant to this conversation?


Rich

[neuroanatomist]

Neuro, remember when I took a few shots of the Sailboat with the CA on the water and then the CPL filter removed the CA, and at that time you had suggested that the CPL, may be good at removing specular highlights. I was wondering if you have tried that at all since then and if you thought it was relevant to this conversation?

Thanks for the reminder, Rich. I didn't try a CPL for this test, but I did put your suggestion to use when I was at the Head of the Charles Regatta rowing event, and it definitely improved the appearance of the river in the shots.

[Keith B]

John = genius


Really great info but I was happy bubbling and stumbling and making uneducated guesses as to why. Every time someone gives me the science behind things, it screws me up. My photos were so much better before I tried to figure the tech stuff out...MAN![8o|]


Thanks a lot John!


Photography for Keith = RUINED![]

[Jon Ruyle]

Thanks Keith and John, for your detailed responses.

What body are you shooting with? 5DII? Obviously if you have MA you've done them right?

5DII, sorry for not mentioning that. I haven't bothered with MA yet. But I believe- especially in the eyeglass shot- we can see the subject pass through the focal plane.

The sharpness in your eye lash shot looks a little soft compared to my experience with the 85L

I agree. I'm happy with the sharpness of my lens wide open, but I don't think it is as sharp as yours (though it is hard for me to tell for sure about these things).

but it is tough to tell because of the focal plain.

I'll use a fancier one next time []


The CA on your lens *seems* like a lot less, but I'm not totally convinced. In those light pics, the background is black, so the lights are *way* overexposed. Yours have a darkish background, true, but perhaps not as dark. In normal pictures, even when I shot at glints of sunlight against dark backgrounds during the day, I get a lot less.


Oh, I don't know.



<div>

But,haveyou tried it?

I tried with the 85 f/.8 and the results were horrendous. I haven't had a chance with the 85 f/1.2 yet (tonight?), but I doubt very much that I will get usable results. I'll get back to you on that. It would be *awesome* if I could use the 85 at f/1.2 for astrophotos (in fact my joy might just be enough to outweigh my angst at KR being right), but I'm still pretty skeptical. I agree that we're seeing mainly axial CA, not lateral. It's absolutely gone at f/4.


I wasn't aware that axial CA was minimal at infinity. Are you sure about that? I don't know anything, but... if the purple wavelength is just a bit off from the green, I'll get a purple disk, even when green is focused at infinity, won't I? (I'll believe you if you say otherwise)


Thanks for the cool info on CA, by the way.



</div>

[Alan]

John = genius
<div style="clear: both;"]</div>

Oh, sure, he's genius, but can he cook? []

[neuroanatomist]

I wasn't aware that axial CA was minimal at infinity. Are you sure about that?

I don't think it's minimal at infinity, but rather minimal within the depth of field, and stars are in focus at infinity. Put another way, the effects of CA are magnified by defocusing. They will still be there, and might be strong enough to affect the image in a demanding application like astrophotography. Also, I really don't want to contribute in any way to KR being right - I always prefer empirical data, so I hope you get a chance to test it! Technically, calling it all CA is a misnomer, since aberrations formally affect only the focused part of the image. But I think it's fair to count CA of OOF light as an aberration since the OOF parts of the field also contribute to the final image (in photography, at least - in microscopy we use tricks like confocal pinholes to optically remove OOF light, or deconvolution to algorithmically remove it).

I don't know anything

Liar!! []