Chromatic Aberration - 17-40 vs Primes

I've got a 17-40 f/4 L and I've been unhappy with Chromatic Aberration I've seen with this lens. I'm wondering if anyone else has converted from the zoom to a prime and seen a change in CA - slight reduction, huge reduction...?

Here's an example, @17mm f/8:

From what I've seen, tokina (or was it tamron? ) 17mm doesn't have that problem.

Was that shot wide open?

As stated, both shots were the widest the lenses could do, and both were at f/8.

Scott, could it be possible that you have a bad copy of that lens. I find it hard to believe that you could get such bad CA with an L lens. I have a Sigma 17-35 EX and CA isn't as bad on it as it looks on yours. AAMOF, CA isn't even really noticeble until I zoom in 300 or 400%. Was your sample a 100% crop or was it maginified more than that?

You know I haven't seen very much CA in the pictures with my 17-40.

I'll take it out for a spin today take some test shots for you and post them, I am tempted to assume you have a bad copy.

But then again, you are bound to have SOME in a high contrast scene as above.

Mmmm.... I prefer to use my 17-40L over the 24mm f/1.4L where light allows because the 17-40L shows LESS CA. Based on that, I'm surprised your 17-40L is showing so much.... in fact I can't recall CA ever being an issue with my 17-40L.

Likewise,. I have not felt at all aware of too much CA on my 17-40mm either.. (granted I've never done any testing)

It might be time for a trip to Canon?

Ooops. I forgot to mention that both shots are crops zoomed to 200%.

But it sounds like I should be seeing less CA with this lens. Anyone have an example of their 17-40 with a scene that should show CA?

I may have to do a real test run and compare it to my Tamron and nifty 50 (only short prime I have). CDS, maybe next time we meet could you bring your 17-40 and I'll compare both with an identical shot?

Okay, I took a couple test shots to see how it acts.

17mm f/4.0 , backlit tree:

HTTP response: 404 | MIME changed to 'text/html' | Byte size: ZERO

HTTP response: 404 | MIME changed to 'text/html' | Byte size: ZERO

I'm a little confused.
1. How does one tell the difference between CA and the "dreaded digital sensor purple fringing bloom on high contrast over exposed edges" phenomenon? Is there are difference?
2. Don't all wide angle lenses have some CA if asked to perform taking pictures of high contrast over exposed details?

Jack

1. I don't know. But given that the sample images look to have blown out skys,. it may very well be purple fringing... ? Right?

2. Yes..

But we've certainly seen cases where someone has a "bad" copy of a lens that shows much more CA than others...

If it's genuine CA it will tend to have red-ish fringing on one side the blue-ish fringing on the other as seen in Scottes posts...

I can't for the life of me find the post now, but WAY back someone here did a comparison between the 10-22 and 17-40L, and I remember him saying the CA and lens flare were problems on his 17-40L. Not saying the 10-22 is a better lens, but just making reference to another person claiming CA problems with the 17-40L. If I remember correctly, he had posted some sample landscape shots up on pbase.

May just have been out of whack. When my 100-400L came back the first time after having its IS repaired it displayed a lot of CA. After being returned to Canon twice more for "recalibration" the lens is fine. I guess that can be a peril of drawing conclusions about a lens from a single sample. When a lens is produced in large quantities like the 17-40L there are bound to be some not-so-good copies out there.

So 'purple' fringing means it's to do with the sensor and 'red on one side, blue on the other' means a not so good lens...

Hmmm, I just purchased a 17-40, will have to do a similar test when I get home to check this out.