For the photography gurus here...

[Gewehr98]

With the Big Thaw in full effect here, I was out taking snapshots of the moving water near the house with my Nikon D-200 Saturday.

I was running my favorite 70-210 f4.0-5.6D zoom lens, with "standard" 62mm 1A UV filter.

This image, at 116mm focal length, f4.2, 1/2500 sec, ISO-200, shows some goofy colors from the reflected sunlight playing off the water ripples. (I resized it down to 1280 x 857 for the sake of our dial-up members at APS)

The effect shows up in several of the flowing water images I took that day.

Is that a function of the UV filter, or simply a prismatic effect from the sun's angle on the water? 

Have I got too wide an aperture, so the shallow depth of field is causing bokeh colors?

It's not unpleasant, but darned if I can nail down the cause.  I will head out again and try the same shots sans UV filter. 

If I add a polarizer, I know I could get them to go away, but so would the reflection of the sun...

[freakazoid]

 

[Brad Johnson]

It's not unpleasant, but darned if I can nail down the cause.  I will head out again and try the same shots sans UV filter. 

Prismatic effect due to variations in surface geometry.

Add a polarizer and change your orientation relative to the sun's position.

Brad

[Monkeyleg]

A polarizer may work for some of those reflections, but not all, as the polarizer just works on one plane.

I'm sure I never saw that happen on film, which makes me wonder what would cause it to happen on digital.

[zahc]

This is why I don't think color is ever going to catch on.

[Monkeyleg]

C'mon, zahc, I wasn't saying film would have done a better job, I'm just wondering how digital might render those reflections differently.

It's like gradations. In film, you can have a subtle gradation, say from 30% CMYK to 32%, and it will go in almost immeasurably small increments. Digital can't handle subtle gradations, so the value may go from 30% to 30.1% or 30.2% to 30.3%, etc. The result is that you can see steps in the gradation. Digital's gotten better at it, but it still does it.

[TommyGunn]

PINK TIDE!
It looks like the water is reflecting the forum's border colors.

[Gewehr98]

I'll try the polarizer.

It was neat looking at all the images and seeing the exact same color gradient in the reflections from top to bottom. 

I *thought* it might have been a function of the lens elements and internal coatings, or even some color fringe errors.

The only problem with that idea is that this is a crop sensor DSLR, and I'm only using the sweet spot in the middle of the big full-frame lens.

[230RN]

There's a pink unicorn in the background which is reflected when the angles are just right in the ripples.

Or someone's Easter Bonnet with lotsa pink ribbons.

Actually, I was thinking that somehow during the exposure period the reflections picked up red from the spectrum --red and then white, wwhich would make  pink, but that seems about as likely as the unicorn theory. 

Also, during the exposure, the results are scanned from top to bottom, right?, so it might have been a timing effect with the water ripples shifting during the exposure scan.... I think this might explain the top-to-bottom effect. (As opposed to the old chemical cameras, where the "data" is collected over the whole film plane at the same time.)

Does this give any hints for anyone to explain it better?

Terry

[Brad Johnson]

I'll try the polarizer.

It was neat looking at all the images and seeing the exact same color gradient in the reflections from top to bottom. 

I *thought* it might have been a function of the lens elements and internal coatings, or even some color fringe errors.

#1 is your key to prismatic effect.  The color gradation follows the spectrum as seen through a prism.

Chromatic aberration will usually show up as color seperations throughout the entire image, or as a concentric problem getting better/worse as you move from the outside of the image to the center.  Coating problems will normally manifest themselves in a specific spot (depending on the illumination source's position) as a flare or partial reflection of internal lens mechanics.

Brad

 

[Gewehr98]

Another potential culprit: "Spherochromatism"

(Although it usually manifests itself on faster lenses, around f1.4 or so...)

A good discussion about the phenomenon brought up this example:

[AJ Dual]

Could any of it be from NIR in the sunlight, showing up in just the right angles?

Like when you watch the business end of a TV remote through a DSLR, or even your phone camera, in theory, you'd think the CCD and it's controller should interpret the IR as deep crimson red, stuffing it as close to the human end where it can be seen, but instead it always renders it as bright pink, or violet-white etc.

[RoadKingLarry]

This is why I don't think color is ever going to catch on.

http://www.youtube.com/watch?v=LtIA0BP-A7w

[Gewehr98]

Supposedly, this is an issue with spherical lenses running wide open, with bokeh regions shifting magenta before the plane of focus, and green behind it. It can be remedied by going to a higher f-number, using a smaller aperture to force ALL the colors to focus at the sensor plane. It's also referred to as "bokeh CA".

I just bought a Constant f4 70-210 Tele off Goodwill the other day. I will have to see if it duplicates the effect...

[Gewehr98]

I was able to duplicate the effect with a 70-300 f4-5.6D lens today. 

Bumping the aperture got rid of it, just like the textbook folks said it would.