Canon 50mm 1.2 disappointment

[szelee]

A hilarious article that pokes fun on this 50mm 1.2 lens issue. Thought it might be interesting for you guys

EF 50mm f/1.2helL USM « Fake Chuck Westfall

[bmauter]

By this are you saying you compose with what you want in focus directly over a focus point, or that you compose and choose the autofocus point nearest the detail you want most in focus?

What I mean is that I compose with the thing I want in focus behind one of my AF focus points. Then I click the AF button and roll the cursor until that AF point lights. Then I push the shutter half way to make sure it gets a focus (red blink and beep!) and then I push the shutter all the way down to take the picture.

In looking at your image, to me it looks like the ear is more focused than any other part of the image. My guess is that the baby moved more so than you did. Have you tested on a still object? I mean, taking a breath can cause a person - let alone a baby - to move more than 5/8 of an inch.

I agree. She could have moved. I have also tested the focus on my pickup truck. You'll see it on one of the attached links earlier. Anyway, I focused on the seam between the driver door and the rear part of the cab--a nice mostly vertical high contrast thing. Miss.

Focus and compose implies you just use the center focusing point to lock focus (e.g., the eye) and compose.

Exactly, although sometimes I used one of the other AF points so I wouldn't have to move the camera as much. Most of the time though I'd use the center. The Xti has 9 AF points. The center one is more precise because it works on horizontal and vertical axises (I think). On the 40d and 50d (and maybe others) all of the AF points are horizontal and vertical except the center also adds diagonal to the mix. However, that diagonal axis can only be used with f/2.8 and faster lenses. The non-center AF points could very likely nail the focus, but it's more likely to get it right with the center point.

Chromatic aberration.

So what you're saying is that really fast lenses tend to exhibit this more frequently than slower lenses and that if you nail focus, the CA is minimized. Time for a new body or maybe one of those aftermarket focus screens.


That fake Chuck Westfall link was very funny. I did use the lens all day today at the Chattanooga Aquarium. I stayed away from 1.2 for the most part. This lens has the best colors I've seen from any lens I've used. I'll be looking through what we took today and maybe I'll post something tomorrow.

Cheers and thanks everyone!
-Brian

PS: Thanks, Darren, for the mention in the email newsletter.

[inkista]

Originally Posted by bmauter

...So what you're saying is that really fast lenses tend to exhibit this more frequently than slower lenses and that if you nail focus, the CA is minimized.

No. What I'm saying is that really fast lenses tend to exhibit this more frequently than slower lenses, and stopping down (using a smaller aperture) will minimize it. Also avoiding overexposure/blowing out any highlights/backlit subjects/hard specular highlights--but sometimes you can't control that. Purple fringing and sensor bloom are a different kind of chromatic aberration than is usually talked about (the cyan/red variety, where the refraction of the glass has caused light of different frequencies to separate to different locations). Purple fringe and red fringe are typically caused by some form of sensor bloom, where the charge on the sensor overflows to neighboring pixels. I think. It's been a while since I looked up the varieties of CA.

[djotaku]

The info in this article might prove useful

Focus Fallibility: Lens Test Fallacies - SLRgear.com!

Basically, camera AF systems are crap and if you're at f/1.2 you're only exacerbating the problem.

[bmauter]

Originally Posted by inkista

No. What I'm saying is that really fast lenses tend to exhibit this more frequently than slower lenses, and stopping down (using a smaller aperture) will minimize it. Also avoiding overexposure/blowing out any highlights/backlit subjects/hard specular highlights--but sometimes you can't control that. Purple fringing and sensor bloom are a different kind of chromatic aberration than is usually talked about (the cyan/red variety, where the refraction of the glass has caused light of different frequencies to separate to different locations). Purple fringe and red fringe are typically caused by some form of sensor bloom, where the charge on the sensor overflows to neighboring pixels. I think. It's been a while since I looked up the varieties of CA.

Sorry, didn't mean to imply that you said to nail focus. One of the links that LoveDSLR included was from Ken Rockwell. In there, he directly states that "There aren't any color fringes, so long as you're in perfect focus." I think I got lost in all of the reading and replying.

[bmauter]

Originally Posted by bmauter

I did use the lens all day today at the Chattanooga Aquarium. I stayed away from 1.2 for the most part. This lens has the best colors I've seen from any lens I've used. I'll be looking through what we took today and maybe I'll post something tomorrow.

http://www.mauter.com/images/gill.jpg
F/2.8, 1/125s ISO 1600 (cleaned up with Noise Ninja), scaled down to 600x400 with Gimp