Segura said:
:)

Dead Thread now . . .

then let it die :P

:)

Dead Thread now . . .

Segura said:
Nothing wrong. It is caused by optical vignetting at a wide aperture. It is called cat's eye bokeh.
Reference 1: http://toothwalker.org/optics/vignetting.html#fig3
Reference 2: http://toothwalker.org/optics/bokeh.html

Very helpful link about "cat's eye bokeh" Segura ... Over 30 years in photography and there is still stuff to learn ...

vidrazor statement 1: "I'm not buying that rap about off-axis falloff. If his explanation were true, edge centers would be just as dark as edge corners."

I assume you are referring to this image

Still wrong, but maybe this will help you understand. I added a blue oval to my simple example. This should represent the off axis vignetting on both the left and the right (looks close to a catseye doesn't it?). Again if you notice more is clipped from the corners, not the edge centers as you believe would be the case. And this line is not a hard line, instead it is a gradual loss of light from the center of the image circle to the ends, but since the edges of the sensor are CLOSER to the edge of the image circle, they are darker.

vidrazor statement 2: "The only reasons I see for falloff are that light travels further to reach the corner, and also strikes element surfaces at sharper angles, losing additional light."
Do you have any evidence to support this statement? Seems like you are taking a guess, but not really understanding why. I believe you are confusing this with Spherical Aberrations mixed in with the separation of light, or Chromatic Aberrations. Stand in front of a mirror. Now stand to the far left. Now the far right. Is there a lot of light falloff? The sharp edges don't matter . . .
REF 3: http://toothwalker.org/optics/spherical.html
REF 4: http://toothwalker.org/optics/chromatic.html

vidrazor statement 3: "Stopping down collects only light rays traveling in a more parallel direction, evening out illumination"
Wrong again. A 50mm lens on an FX sensor has a field of view of 46°. It will collect light within this field of view. Now imagine a cone of light starting from the aperture blades at 46°. If you are shooting wide open, then your cone is much larger (still 46°), but it is too large to clear the lens barrel. Now imagine stopping it down to f/22, where the opening is much smaller. Therefore you have a much smaller cone of light, which is able to clear the lens barrel. Now there should be a certain aperture where the amount of light is small enough to completely clear the lens barrel. This is the maximum aperture you can use where there is no vignetting from light fall off.

You can visualize this in the following image. The red bars indicate the exit pupil (aperture). In this image, the lens at an f/11, and no light falloff for this field of view occurs. If you were to open the aperture up more, your cone would be fatter (same angles) and be clipped by the barrel. It is called natural vignetting, and has nothing to do with parallel light (does this even exist?)

Believe what you want . . .

>>Do this test:<<

Actually, you're reinforcing my point to a degree. The light has to travel further to reach those edges. That's was my point. Remember that 35mm optics on a 24x36mm frame actually fill an area OUTSIDE of that frame, and there is probably a similar situation with DX-design lenses. So while your illustration is correct, the falloff depends more on the lens type and particular design.

The authors point on the link you post claims aperture effect from lateral light travel, as illustrated by the side shot of the lens. That's what I'm not buying.

More abbreviations to confuse more people - I love it! And I thought OOFH was the sound I make when trying to bang my way through a crowd.

Sorry someone had to say it :)-

Excellent links Segura. Makes a lot of sense if you think it through that the circle of confusion will become less and less a circle as you move away from the point of focus.

Both of those pics of kids are fantastic. The first just gives her that halo of lights and your image gives a sense of motion. Both are well documented in your two links.

I can go home now, I learned something today.

Oh, yeah, word of the day... OOFH

aetas said:
Did you try this test. Or did you test it again when you got home.

Who are you asking this to, the OP? And you forgot to use OOFH!

Did you try this test. Or did you test it again when you got home.

vidrazor said:
If the conditions were more or less identical and you weren't getting these effects before, are you using any in-camera processes (or stopped using) like CA correction, sharpening, etc that you may not have used (or used) before?

Final test to rule out any in camera processing. Look through your camera at some bokeh like the original shot and you will see the same results. Why? Because you are looking through your lens at wide open aperture, so you will see the effect before you take the picture.
If you have a DOF preview button to change the aperture, you will see it clears up if you stop it down as well.
This rules out any of the issues that vidrazor pointed out.

Now go out and take some pics!

vidrazor said:
I'm not buying that rap about off-axis falloff. If his explanation were true, edge centers would be just as dark as edge corners.

They are man . . . you just don't see it because of the crop. Do this test:
1. Draw a circle on a piece of paper (this is the circle created by the lens)
2. Draw a rectangle inside the circle (this is the area of the sensor)

Now what part of the rectangle is closes to the circle? Is it:
A: The corners
B: The center of the sides of the rectangle (as you believe)

This is why vignetting is more pronounced in the corners, since they are closer to the edge of the image circle.

This should explain the falloff:
Dark black circle is the full image circle. Red circle is the image circle without light falloff. Grey shaded area is vignette. Black rectangle is the image frame.

Thanks segura. I have never heard the phrase OOFHs Im going to use it all the time now though so i apologize for future posts. Its neat to see people spending the time to get links to explain some of the things they are saying. There have been many times on forums when a link has explained what I was trying to say better then myself.

[Reference 1:]

I'm not buying that rap about off-axis falloff. If his explanation were true, edge centers would be just as dark as edge corners. The only reasons I see for falloff are that light travels further to reach the corner, and also strikes element surfaces at sharper angles, losing additional light. Stopping down collects only light rays traveling in a more parallel direction, evening out illumination.

As for the OP remark about noticing this particular phenomenon, do you have the earlier images without this effect? Have you compared focal length and F settings to see if they match? What approximate distances from the OOF sources versus the subject were you compared to the distances of these new images? Similar, different?

If the conditions were more or less identical and you weren't getting these effects before, are you using any in-camera processes (or stopped using) like CA correction, sharpening, etc that you may not have used (or used) before?

@Segura
Good links! Never knew someone named it just knew that was they way it was. Love the photo as well!

Nothing wrong. It is caused by optical vignetting at a wide aperture. It is called cat's eye bokeh.
Reference 1: http://toothwalker.org/optics/vignetting.html#fig3
Reference 2: http://toothwalker.org/optics/bokeh.html
I am actually a bit surprised that no one who replied seemed to know what it was. I recommend reading both links, you will probably learn a thing or two. Not trying to slam anyone, just giving an opportunity to learn something . . . I have much to learn still too.

Here is a shot I took on a 70-200mm VR II with the same effect. You can see the eyes are circular around the center of the lens. This was shot at f/2.8. If you stop it down to f/4 it should go away, but that may not give you enough DOF.

Here is a group on Flickr
http://www.flickr.com/groups/catseye/

Again, nothing wrong with your lens . . .

Also since you are shooting wide open, if it was a sticky aperture blade, you would see it . . . the aperture wouldn't move anyway since you are . . . shooting wide open. So that rules out the blades.

I also think there's nothing wrong with that lens. Nice to see you around again, Phil. Your daughter is adorable.

Stuck or broken blades do happen - I had my 85mm 1.8 blades come loose once and it created odd squarish highlights as well as some darker banana peal shape shadows across the image.

I didn't see anything like that in his other images so I do not think it is happening here.

I do think that there is nothing wrong with the lens or camera.

Thank you for the analysis, TaoTeJared, I mentioned a stuck blade before I noticed these were taken wide open. A sticky blade causing problems on wide open shots would be rare enough to nullify the "most common" statement.

Though, a blade so sticky that it remains stuck for seconds (minutes even) would be easy to visually diagnose. Most sticky-blade problems are caused by slow blades and not truly stuck ones.

Ah they are right. When you do your test try to use a somewhat round light. I could see that making a difference at close range. It would be hard to get a round shape out of a rectangular light...Right?

I looked at many of your photos on your Flickr stream (very beautiful family by the way) but everything seems to be coming out fine and nothing is wrong with your camera at all.

~"I have discounted the fact that the elongated shape of the lights have anything to do with it because the ovals seem to make almost a halo shape around my daughter, and not something more random."

I'm not sure if discounted is the best word or what you are meaning by it... This is absolutely normal and you are correct that since the light source is not perfectly round. Also shooting down at an angle and some camera shake can cause some of that as well. Having fill flash can exaggerate the background shakes as well. The closer you are to light sources the more this can exaggerate.

I noticed you are shooting In AF-C - there is a setting that is AF priority or Shutter priority. That means that either the shutter will fire if the focus is locked or not. I'm guessing you have it in Shutter priority which will fire if the focus is NOT locked.

I also noticed on some of the "almost" sharp photos that you are shooting at very slow shutter speeds. Handhold anything below 1/60th (1/80th for me) you will start to see the shakes move in. A few were shot at 1/4s which I would always expect shake. I use Auto iso and shut the shutter speed at my focal length, (10-20% higher for DX)

I'll be sure to provide an update...any strange problem deserves a resolution!

Did I say work? I meant working...outside or something. :)

Im guessing from you last post its still under warranty? Good news tell us how it all turns out when you get home from work. And what are you doing on nikon rumors at work=)

hm...I was hoping it wasn't something mechanical, but that would make sense. Thank heavens for Nikon warranties!

A stuck aperture blade is the most common cause of out-of-round bokeh.

Thanks broxibear.

That is very interesting...I almost always keep my camera in AF-S mode...not sure why I ever put it to C. But I didn't even notice about AF Fine Tune. That may be why I am having a hard time with focus.

Now I just have to wait a few hours until I get out of work to go home and give it a try. Very good catch!