Arg, should I get the 135mm f2 or 50mm 1.2L?

[TCampbell]

Never heard of the longitudinal chromatic aberration that the 50mm 1.2-L has??? Canon EF 50mm f/1.2 USM L - Review / Test Report - Analysis

I checked out the article. They inserted a sample image to show CA... which, btw, does not show CA. I looked at it and everything rushing through my head screamed "You imbeciles! You shot a photo of water in motion!" Absolutely NEVER EVER under ANY circumstance should you use water with waves on the surface to test for CA. Water is like a thousand prisms splitting light into it's constituent components." To test for CA you need something where the only thing splitting light is the lens itself. With water, the subject has optical properties of it's own so they haven't isolated anything to the lens.

BUT... then I found an image that does show legitimate CA on their next page. If you follow your link, go to page 2, and then pick the sample of image of the stone wall with windows (here's the link to the direct image: http://images.photozone.de/8Reviews/lenses/canon_50_12/samples/IMG_2841-01.jpg ) you can see the CA.

Check the upper left corner. There's a shadow of a stick and on the "inside edge" (the edge closest to the center of the image) you'll see a slight red fringe. On the outside edge there is a slight blue fringe. If you go down to the opposite corner you can find shadows showing the same thing (inside edge is red-fringed and outside edge is blue-fringed.)

They did this on a 350D body (a crop-frame) which means the CA would likely be stronger on a full-frame body where the distance from center to corner is even farther.

You can correct for this in software. I use Aperture, which has a few CA correction tools (it's possible to correct for it using any software which will let you split the RGB color channels -- you can resize and shift each color independently to re-align the image... but that's more cumbersome than just using a CA correction tool.) Aperture or Lightroom would be able to apply this correct en-masse (every image taken at the same f-stop would have identical CA). Every lens has some CA... the question is to what degree and is it enough to be worth fixing.

 

[thepilgrimsdream]

My 85mm 1.8 has some CA, but it doesn't bother me cause its gone with a click in Lightroom. Honestly, with the power of editing software I've never come across a single lens that I've had to worry about CA. If the colors are a little different per lens, than you simply make a preset.

 

[LShooter]

85 1.2 or 100 macro L

The 100 is close enough to the 135, takes amazing portraits and also give you a lens with multiple uses. The 85 1.2 because, well, it's the best portrait lens made.

 

[Derrel]

You can correct for this in software. I use Aperture, which has a few CA correction tools (it's possible to correct for it using any software which will let you split the RGB color channels -- you can resize and shift each color independently to re-align the image... but that's more cumbersome than just using a CA correction tool.) Aperture or Lightroom would be able to apply this correct en-masse (every image taken at the same f-stop would have identical CA). Every lens has some CA... the question is to what degree and is it enough to be worth fixing.

Partially correct; lateral chromatic aberration can be corrected in software. and that is what you are talking about. But longitudinal chromatic aberration can not be corrected in software. You seem to be missing the distinction between these two types of CA, as well as which one is correctable in software and which one is not correctable.

You mentioned red and blue...but longitudinal chromatic aberration is GREEN and MAGENTA fringing. Not red and blue. Longitudinal chromatic aberration shows up in front of and behind the plane of sharp focus. MANY fast lenses suffer from longitudinal chromatic aberration at wide f/stops, which is why I called the 50/1.2 a "fast f/2.8" lens...because like the 85/1.2-L, it has bad longitudinal CA when shot at the widest apertures.

The 85/1.2-L is even worse than the 50/1.2-L in terms of its strong green and magenta fringing (aka longitudinal chromatic aberration) on out-of-focus objects. A lot of non-apochromatic lenses suffer from pretty marked longitudinal chromatic aberration at their widest apertures. The idea that every lens has some CA?? Umm...not really an accurate statement.

 

[o hey tyler]

85 1.2 or 100 macro L

The 100 is close enough to the 135, takes amazing portraits and also give you a lens with multiple uses. The 85 1.2 because, well, it's the best portrait lens made.

Both of the lenses you mentioned are extremely slow focusing lenses. I would not say the 85/1.2 is the best portrait lens made either.

 

[TCampbell]

You can correct for this in software. I use Aperture, which has a few CA correction tools (it's possible to correct for it using any software which will let you split the RGB color channels -- you can resize and shift each color independently to re-align the image... but that's more cumbersome than just using a CA correction tool.) Aperture or Lightroom would be able to apply this correct en-masse (every image taken at the same f-stop would have identical CA). Every lens has some CA... the question is to what degree and is it enough to be worth fixing.

Partially correct; lateral chromatic aberration can be corrected in software. and that is what you are talking about. But longitudinal chromatic aberration can not be corrected in software. You seem to be missing the distinction between these two types of CA, as well as which one is correctable in software and which one is not correctable.

You mentioned red and blue...but longitudinal chromatic aberration is GREEN and MAGENTA fringing. Not red and blue. Longitudinal chromatic aberration shows up in front of and behind the plane of sharp focus. MANY fast lenses suffer from longitudinal chromatic aberration at wide f/stops, which is why I called the 50/1.2 a "fast f/2.8" lens...because like the 85/1.2-L, it has bad longitudinal CA when shot at the widest apertures.

The 85/1.2-L is even worse than the 50/1.2-L in terms of its strong green and magenta fringing (aka longitudinal chromatic aberration) on out-of-focus objects. A lot of non-apochromatic lenses suffer from pretty marked longitudinal chromatic aberration at their widest apertures. The idea that every lens has some CA?? Umm...not really an accurate statement.

We still have software that corrects for it.

As you've pointed out, the "problem" is that the distance from the rear-most element to the focus plane (the back-focus distance) is different depending on the wavelength of the light. Since visible light on the blue/violet end of the spectrum (nearest to UV) is short and bends the most, it tends to have the shortest focus distance. The longest waves -- reds near the IR -- bend the least and tend to have the longest focusing distance. CA (regardless of orientation) is _always_ caused by the law of physics that causes light waves to bend by a different amount when they transition through any medium at an angle. Since there difference between red and blue (technically violet) is the strongest, that's where you notice the CA (even though it's really happening all throughout the spectrum.)

You can think of the single image as being three separate images... a "red", "green", and "blue" image. If you were to turn off the green channel and just compare the red to the blue channel, you'd discover that the red image (and this is a generalization since corrective elements are actually able to reverse this -- so I'm really just talking about a single element problem) is actually a tiny bit larger than the blue. If you were take the red image and reduce it in size ... just shaving it down very fractionally, you can get the red and blue images to be the same size and the images will converge. You can do the same with the green channel... it wont need to be resized as much as the red. Note that it's not shifting the layers... it's resizing them and keeping their center points aligned.

That's basically all the software does... but rather than make a person use photoshop to split the channels, resize them, align them, and re-merge them, the CA correction does it all automatically.

CA is a law of physics... pass light through the air to surface of your lensing medium and it'll split into it's constituent frequencies. Multiple elements try to correct for it -- the correction is never technically "perfect" - you can just get it so low that it's not noticeable, but it's technically not zero.

Lower focal ratios typically have more of a problem because the glass elements are physically larger -- so they're starting out with a bigger problem and have more work to do to correct for it.

 

[JohnTrav]

I would go with the 135 since you have the 50mm range covered with a few lenses.

Also I would look in to some reviews on the 50mm if you really plan to buy it. Like mentioned before it might not be worth it to spend the money on it over the f/1.4 model. I use the f/1.4 and it is really nice and sharp. Just as sharp at my 70-200 L glass IMO. I have also heard of people saying that the 50mm f/1.4 is sharper than the f/1.2. I am not sure how true that is. I read reviews on the f/1.2 saying that below f/2.8 its hard to get a sharp shot also.

 

[JALBERTSTUDIOS]

I used the 50mm 1.2 this weekend and loved it! I didn't have any issues with any purple fringing. I found the lens super sweet if you use it at 1.2 to 2.2. I am purchasing the 135 from BH tomorrow and the 50 1.2 in a couple weeks. I just purchased the 85mm 1.2 L II and the new 11-24mm 4.0 and love them!! I rented the 135 from lenspro to go and it is super sharp, wide open! I had no issues with longitudinal chromatic aberration as Derrel, mentioned. My next after these two purchases will be the 24-70 2.8 and 70-200 2.8

 

[dpt]

You'd buy 135 mm or 50 / 1.4. For FF 135 mm are just fine for portraits.

 

[davidrb1946]

I mostly shoot portraits. I also am a second shooter for a lot of weddings. I love shooting with primes wide open using natural light. I know the 135 f2 is a better focal length for portraits and is outside of the range of lenses that I have, but the 50mm 1.2 is on rebate right now and lets in a stop and a half more light.

What lens would you go for? I feel like either of these would really compliment my shooting style

I have a
Canon 5d mk2
17-40 f4L
28-70 f2.8L
85 1.8

Your mind is already telling you the 135 is the lens to get. Re-read your post and see that the 50mm focal length is covered by your fastest lens. The reasons you list for the 135 are much stronger than your preferences for the 50, they are actual applications. The 135 surprised me. I used it all winter to shoot birds perched on feeders outside my window. The lens stayed on the camera all winter, I found things inside that interested me as photo models. Remember it is f/2.0, that is faster than what you currently have. It is a great lens inside as well as outside. You already own one of Canon's "sleeper lens," the 17-40. Both lenses are "L" quality, very versatile, and very in-exspensive. Another feature that will add to the 135's versatility is it's compatibility with the extenders. Now you can shoot at 270mm @ f/4.0! Several more doors have opened! Good luck and have fun with whichever way you choose.