Problem Solving

[bisso7]

Hi All,

I was recently out shooting some hummingbirds. After reviewing the photos, I discovered a problem I have not ever seen. Please see below, cropped photo.

Notice on the outer edge of the white flowers, some sort of bluish glare. Could someone explain what this glare might be? None of the photos I shot were keepers, but I saved this one as a learning experience.

I used a Canon EF 85mm f/1.8 USM lens.

[Glenn NK]

Seems to be on all the white flowers around the petals.

For starters:

1) shooting RAW or JPEG?

2) is the image clipped in the histogram? (it looks like they are blown out a bit).

3) have you applied any sharpening to the image?

4) do you use a lens hood?

5) what software was used if any was used, and what adjustments were applied?

Glenn

[Mark von Kanel]

As glen says we need sooooo much more info!

[tclune]

Question: did you shoot this hand-held with image stabilization enabled?

[IzzieK]

Next time you do get to do another one of those glaring plastic flowers, take several shots, tripod enabled, of the same scene. Watch your histogram or the blinkies. Practice adjusting either your ISO, SS or Aperture till the blinkies disappears. If all else fails adjust your EV. Then merge your best shots together in Photoshop. You'll be right, mate!

[DanK]

looks to me like an extraordinarily bad instance of purple fringing, a kind of chromatic aberration that is generally most noticeable on high-contrast borders. However, I have never seen it this bad, which makes me wonder--if it is purple fringing, why is it this bad? I'd second Glenn's questions.

[tao2]

It is CA... fast aperture primes seem tae show this often, even the better ones.

Yer exif shows...

"Lens: Canon EF 85mm f/1.8 USM (shot wide open)
Exposure: Auto exposure, 1/512 sec, f/1.8, ISO 100"

Stop down...

[William W]

+1 to Dan's & Boab’s analyses.

I have this lens and I do use it at F/1.8.

The extreme Purple Fringing (CA) is as result of (as Boab describes) - intrinsic to the lens & the aperture that was chosen.
I’d add that the CA is exacerbated in this particular shot by the extreme reflective white of the Object and also the strong harsh light on it.

If a filter was on the lens, that might worsen.

WW

[Glenn NK]

I've had some purple fringing (CA) the odd time under severe conditions - shooting through the leaves of a tree with sky in the background, but I've never seen anything as bad as this. So I discounted CA as a possibility.

In my experience it usually occurrs when the contrast is very high, and in this case, a white flower on green foliage didn't seem to be an unusually contrasty situation.

I've photographed white flowers on a green foliage BG and never saw any. That being said, I use a 100 macro lens that isn't prone to the problem.

So the diagnosis of CA seemed a bit of a surprise to me.

Glenn

[NorthernFocus]

There are two possibilities worth consideration:

1) Chromatic aberration which that particular lens seems to be susceptible to as pointed out by others, or

2) The plastic flowers are radioactive. In which case I hope you have your affairs in order...

[William W]

That particular lens is reasonably tragic for it.

It’s winter down here but a sunny morning – I just pulled these two shots using an EOS 5D and the EF 85F/1.8 for your interest and as a rough example.

Note that the white area along the metal window frame base in both my samples is exposed OK; but in contrast, in the OP’s sample, the whites are blown out big time. (regarding Glenn's comment about contrast).

In both my examples, the top image was pulled at F/1.8 and the bottom image at F/5:

Here (JPEG’s both SOOC) view them large:


Enlarged bottom right quarter:


WW

[Glenn NK]

looks to me like an extraordinarily bad instance of purple fringing, a kind of chromatic aberration that is generally most noticeable on high-contrast borders. However, I have never seen it this bad, which makes me wonder--if it is purple fringing, why is it this bad? I'd second Glenn's questions.

So far we collectively seem to feel that CA is the problem; which surprised me as this is a very common portrait lens.

So I had a look at Photozone's review:

http://www.photozone.de/canon_eos_ff..._18_5d?start=1

To quote the section on Chromatic Aberrations:

"Lateral Chromatic Aberrations (colour shadows at harsh contrast transitions) are basically a non-issue - this is a very rare characteristic".

However under the heading BOKEH FRINGING (Longitudinal Chromatic Aberrations - LoCa), they state: "LoCA's are a primary problem with ultra-fast lenses, and the EF f/1.8 USM suffers quite a bit here".

Even so, the example posted by the OP seems quite extreme, so I'll admit I'm not convinced just yet. My 100 macro is only f/2.8 (an aperture I use a fair bit) and I've never seen anything near this bad (on any lens).

I'm almost ready to check the flowers out with a Geiger Counter (Dan Koretz). Those damn things are GLOWING.

Bill notes a fair bit of overexposure on the flowers - which ensures they have plenty of contrast with the BG.

G

[William W]

@ Glenn:
Don't misinterpret my comments. I think that it is very important for Jeff to answer all your questions. I still think that it is CA - but I think Jeff may have enhanced the CA in post-production or by another method.

Also there also could have been a filter on the lens, so I think that Jeff can certainly provide more information, even more than your questions ask.

*

Jeff:

It would be a good test to for you to make a reshoot as best as possible mimicking the lighting conditions and also the OVER EXPOSURE of those white flowers.

Perhaps shoot a couple of frames at F/1.8 and a couple more at F/5.6 and a couple more at F/11 - there need be no hummingbirds in the shot just the radioactive flowers.

WW

[bisso7]

For starters:

1) shooting RAW or JPEG?

2) is the image clipped in the histogram? (it looks like they are blown out a bit).

3) have you applied any sharpening to the image?

4) do you use a lens hood?

5) what software was used if any was used, and what adjustments were applied?

Thanks to ALL of you for your great comments, suggestions, and even photo examples (William) in an effort to help me understand what is going on this in this photo.

In response to the questions, Glen:

1. I always shoot RAW

2. Yes, the image is clipped in the histogram. The highlight is raised quite a bit, and then declines significantly towards shadow side. The radioactive plastic flowers are "blinkies" in the histogram, IzzieK

3. I've applied no sharpening to the image. What you see is "as shot." I was not satisfied with any of the shots, so I didn't spend any time in post-production.

4. No lens hood. I did wonder, however, if I had used one is this would have minimized this issue.

5. Again, no adjustments were made to the photos.

The settings in this photos are: SS 1/500, f/1.8, ISO 100, WB Daylight, and I was on a tripod , roughly, about 50ft from my subject, shooting in AV.


About this lens, and maybe some of you can give me more insight into this particular lens, it's a new lens. I haven't used it very much since I purchased it. I actually purchased this lens for portrait purposes. I wanted to see, however, how it did outside shooting wildlife shots, as I wanted the effects of the f/1.8 aperture in my photos.

What I have noticed about this lens thus far, is that it seems to be much more sensitive in light than my other 2 lenses. It seems to do very well during the golden hours of the day, however. It takes great indoor portraits, but outside it's much more sensitive and I have to be more aware of lighting conditions.

Great suggestion, IzzieK! Next time I will definitely take the same shot using different settings.

From what little I have studied CA, that diagnosis certainly seems appropriate. NorthernFocus, I'll have to go with your #1 possibility, as I don't yet have any of my affairs in order.

Thanks, again, everyone. . . I'll be looking for any further replies based on the information I have provided. . .

Regards,

Jeff

[Glenn NK]

CA - probably, but which one? Lateral or longitudinal? We should determine which.

The Photozone reviews suggests that lateral CAs are a non-issue with this lens, but that LoCAs may be an issue, particularly wide open.

But if it's really blown (which can only add to contrast), then this could be a major part of the problem.

I'd like to see Jeff re-shoot the image, but this time with an exposure that has no blown highlights.

I'm betting that the problem will disappear.


@Bill:

Hadn't meant to misinterpret your comments and didn't realize I had; was referring to your statement ". . whites are blown out big time". They sure are.

Glenn

[bisso7]

CA - probably, but which one? Lateral or longitudinal? We should determine which . . . . I'd like to see Jeff re-shoot the image, but this time with an exposure that has no blown highlights.

I'm rather ignorant on the subject of CA and its differences, unfortunately, but I'm glad to have a starting point. I will re-shoot this scene and try to mimic the lighting conditions as much as possible. When I do I will post my results and include several different settings as well.

Thanks to all for your good input! I appreciate the learning experience, very much..

Jeff

[Glenn NK]

I'm rather ignorant on the subject of CA and its differences, unfortunately, but I'm glad to have a starting point. I will re-shoot this scene and try to mimic the lighting conditions as much as possible. When I do I will post my results and include several different settings as well.

Thanks to all for your good input! I appreciate the learning experience, very much..

Jeff

Jeff:

We are all ignorant until we learn about something.

This might help:

https://nikoneurope-en.custhelp.com/...-aberration%3F

http://toothwalker.org/optics/chromatic.html

I rather like the second reference (although it's a bit longer and more technical).

I think that Longitudinal and Axial CAs are the same thing - in any event Lateral CAs seem to be more common.

Glenn

[William W]

. . . I'd like to see Jeff re-shoot the image, but this time with an exposure that has no blown highlights.

I'm betting that the problem will disappear.

Yes I think that there will be an huge diminishing - but not disappear. That's why I took my samples, at an "OK" exposure for the white so we could see the DIFFERENCE in the purple fringing when using F/1.8 or F/5.

So I think that the CA will disappear all the more, if the exposure is appropriate for the white area - AND - the lens is stopped down.

But the practical issue Is that: in that original shot – IF the white flowers are exposed correctly, then the Humming Bird will be quite underexposed, I expect.

*

@Bill: Hadn't meant to misinterpret your comments and didn't realize I had; was referring to your statement ". . whites are blown out big time". They sure are.

OK. I understand. (I understood previously that's what you meant)
I was only ensuring you (and Jeff) understood my posts were quite clear.

WW

[bisso7]

I'd like to see Jeff re-shoot the image, but this time with an exposure that has no blown highlights.

I'm betting that the problem will disappear.

I re-shot this scene, as precisely as possible, from the same distance. The lighting conditions were very similar to that of the original post (above), only this time, I first used the identical settings as in the original, and then I stopped the lens down on several shots to see if this would eliminate the purple fringing.

As I stopped down the lens, it seems that the purple fringing significantly decreases, yet, though more subtle, it still exists. Seems that it wasn't until I applied -1 exposure compensation that the problem disappeared. I also applied bracketing during the re-shoot.

Settings are indicated below each photo.



AV, 1/500, f/1.8, ISO 100



AV, 1/60, f/5.6, ISO 100



AV, 1/15, f/11, ISO 100



AV, 1/60, f/8.0, ISO 100, EC -1



AV, 1/30, f/8.0, ISO 100, AEB 0



AV, 1/30, f/8.0, ISO 100, AEB -1



AV, 1/30, f/8.0, ISO 100, AEB +1


This lens (i.e. Canon EF85mm f/1.8 USM) seems to be extremely sensitive to light. Though I haven't received it yet, I bought a lens hood for this lens and will continue experimentation using the hood to see if this helps the problem.

Regards,

Jeff

[Glenn NK]

Seems that it wasn't until I applied -1 exposure compensation that the problem disappeared. I also applied bracketing during the re-shoot.

This lens (i.e. Canon EF85mm f/1.8 USM) seems to be extremely sensitive to light. Though I haven't received it yet, I bought a lens hood for this lens and will continue experimentation using the hood to see if this helps the problem.

Regards,

Jeff

From looking at the images and what you've said, it seems that exposure was a large part of the problem.

What did the histograms look like (not the camera histos, but the PP histos - were they blown out at all)?

Glenn