red and green donuts

BugEZ · Post by **BugEZ** » Mon Jul 21, 2014 3:57 pm

The photo below is a crop of one of the images in a stack of a doli's eye. the "A"region is toward the lens and the "C" region is away from the lens. B is on the focal plane.

What I can not explain is why there are "red donuts" surrounding the black centers of the "red" ommatidia in the area toward A, and green donuts in the centers of the green ommatidia toward C. I would have expected the donuts on the same side, both toward A or both toward C. Any clue about what I am seeing? Is the "B" region actually on the focal plane or is it above or below the center? Because of the high contrast our software tools select it for the stacked result...

K

rjlittlefield · Post by **rjlittlefield** » Mon Jul 21, 2014 9:00 pm

This looks like a severe case of longitudinal chromatic aberration away from the focus plane. What's happening -- I think -- is that in one direction red goes OOF faster than green, and in the other direction it's the reverse. The color cast in the dark area depends on which color has spread fastest into the area. I don't know whether the chromatic aberration is due to your lens optics or the doli's eyes (or some some combination of both).

Yes, I take the B region as being the part that's actually on the focal plane.

--Rik

jswatts · Post by **jswatts** » Mon Jul 21, 2014 11:30 pm

Yep. Ditto Rik.

enricosavazzi · Post by **enricosavazzi** » Tue Jul 22, 2014 12:53 am

I also agree that this is most likely axial (=longitudinal) CA. It would be interesting to know what lens/tube lens combination you are using. Others with the same equipment would then be able to try and reproduce this problem.

BugEZ · Post by **BugEZ** » Tue Jul 22, 2014 3:23 am

Many thanks to all for your thoughts. Longitudinal chromatic aberration is an interesting theory.

Enricosavazzi wrote:

It would be interesting to know what lens/tube lens combination you are using. Others with the same equipment would then be able to try and reproduce this problem.

For this photo I used my Olympus UMPlanFL 10X/0.3NA Infinity corrected objective, mounted in front of my 200mm f4 Pentax telephoto. The lens is in front of my Pentax K100D. I don't have specs on the doli's optics

. There is an empty filter shell to provide a bit of stand off between Iris and the tube lens, and an Iris that I keep wide open.

I suppose a test of a black and white hatched target would show if the phenomenon originates is in my optics or the dolis...

One thing we have to consider is that the doli's lens is mounted in front of a rather long shiny tube/wave guide that is much longer than the lens is wide. When we are "focused" I am not sure what part of the eye we are seeing. The sides of the tube or the target at the end where the photo receptors are located? The nice focused hex pattern suggests the focus is near the surface, but with the insect's lenses tossed in I am less certain.

K

BugEZ · Post by **BugEZ** » Tue Jul 22, 2014 4:55 am

Quick study of a black and white subject. This confirms some longitudinal color aberration.

First the subject with an arrow pointing to a tiny dot on this paper scrap.

crop of dot in focus

Focus in front of the dot...

Focus behind the dot...

Mystery solved I think...

K

enricosavazzi · Post by **enricosavazzi** » Tue Jul 22, 2014 6:26 am

BugEZ wrote:Quick study of a black and white subject. This confirms some longitudinal color aberration.
[...]
Mystery solved I think...
K

Yes, this is the typical way axial/longitudinal CA looks like.

However, we don't know yet if it comes from the microscope objective, or from the tube lens, or from an "unlucky" combination of the two. Axial CA is not reduced by closing the lens aperture (although it can be hidden by an increased DOF), so the cause of this CA could be the 200 mm lens in spite of it being substantially "stopped down" by the microscope objective.

Perhaps other users of the Olympus UMPlanFL 10X/0.3NA could comment on whether they have experienced the same problem with other tube lenses. If they haven't, then it might be appropriate to try a different tube lens.

rjlittlefield · Post by **rjlittlefield** » Tue Jul 22, 2014 9:02 am

BugEZ wrote:When we are "focused" I am not sure what part of the eye we are seeing. The sides of the tube or the target at the end where the photo receptors are located? The nice focused hex pattern suggests the focus is near the surface, but with the insect's lenses tossed in I am less certain.

You can confirm what you're focused on by scanning through your stack from outside in (closest to the camera lens working back). Whatever comes in focus first has to be at the surface of the eye. There may be a little bit of texture visible on the front surface of some ommatidia, but I expect that the first strong features you'll encounter are those bold hexagons which are the boundaries between ommatidia at the surface of the eye. The black "pupils" are actually reflections of the hole in your lens. Those will be in focus some distance in back of the ommatidia, following the rules for convex mirrors. From the image you've posted, it looks like the ommatidia are strongly curved so that the pupil is focused at a depth not much different from the ommatidia boundaries.

--Rik

BugEZ · Post by **BugEZ** » Tue Jul 22, 2014 3:51 pm

Rik and Enrico,

Thanks for your comments. I'll scan through my images of some "hairy eyed" dolis and look for evidence of hairs. Those will be at the surface.

Thanks!

K

Charles Krebs · Post by **Charles Krebs** » Wed Jul 23, 2014 2:17 pm

I regularly use several of the Olympus MPLFL infinity objectives (with an Olympus microscope head or a Nikon tube lens). They are not "Apo" but are very good.

My experience with them would cause me to believe that the effect seen on the initial "eye" image appears way too severe to be an axial chromatic error from this series of objectives (unless you have a real "dud").

The second example (the dot on paper) is actually a pretty severe test. I would expect to notice some axial "color" error in such a test unless you were using an Apo objective.

As Enrico mentioned, you can't entirely eliminate the possibility of the tube lens being used as contributing to this. Even if it is not, the amount of color error the "dot" series illustrates does not seem excessive for a Fluorite (but non-Apo) objective. (And to me would not account for what you are seeing in your first example).

BugEZ · Post by **BugEZ** » Wed Jul 23, 2014 6:07 pm

Charles Krebs wrote regarding Olympus MPLFL objectives:

My experience with them would cause me to believe that the effect seen on the initial "eye" image appears way too severe to be an axial chromatic error from this series of objectives ...

I have been very happy with this objective. The working distance is adequate. The sharpness is great. I get good results with most subjects photographed with this lens.

I think the red and green donuts are quite interesting. Since the out of focus artifacts don't end up in the final image the longitudinal chromatic aberration (if it is that...) does not appear to be a problem.

Many thanks for your input!

I did another experiment to see if I can better judge the location of the corneal surface of the fly's eye. I dabbed an eye with a magic marker to try to make a visible mark at the surface. I then examined the images in the stack to see if I can better gauge the cornea's surface relative to the spot picked by Zerene and the above/below bits. Bottom line is that Zerene does a fine job. The slice it picked to contribute to the final image of the ommatidia were the ones with the sharpest contrast of the marker boundary. If there is a difference in elevation of the surface and the stuff in the final image I would say that it is within the bounds of the depth of field of the optics.

crop from first frame with focus below marker blob at surface. Note area in upper right...

second frame with focus at surface (I think...)

third frame with focus above surface

Dmap result...

Keith

rjlittlefield · Post by **rjlittlefield** » Wed Jul 23, 2014 9:13 pm

This is an interesting problem. I still think it looks like LCA, but I agree that the camera & microscope optics do not have nearly enough LCA to explain what we're seeing. That puts responsibility on the doli's eyes. We normally think of the "black hole" as being a surface reflection of the lens pupil, but I wonder if in this case some of the reflection is actually coming from layers below the surface and thus has an opportunity to be affect by CA in the doli's eyes. Looking at the last image posted by BugEZ, I'm struck by how much more red appears in the black holes of the yellowish ommatidia than of the greenish ommatidia. Looking back, I see the same effect in the first image and also there's more green in the black holes of the greenish ommatidia. I considered that could just be spectral filtering, yellow = green+red so more red in the yellow, more green in the green, but then it still seems like there's a serious shortage of green in the OOF yellow ommatidia. Seems to me like the CA is in the doli's eyes, but at this point I can't explain the details.

--Rik