[henry link]

Recently there were some complaints on another thread that we‘re not being informed about what kind of “ED” glass is actually used in various new binoculars that claim to use a special low dispersion glass. That’s true and the manufacturers are not likely to reveal the exact ED glass type they use, but knowing that wouldn’t really tell us what we want to know anyway. What’s important is not the glass type, but whether the overall design actually accomplishes the goal of reducing longitudinal CA compared to a conventional design. We can see that for ourselves.

The photos below show the center of a CA target I use to evaluate both longitudinal and transverse CA. They were made by simply placing the camera behind the binocular eyepiece, then cropping down to a tiny area of the center of the field. This is similar to what you would see with the magnification of these 8x binoculars boosted to about 64x.

The top row shows the CA at the center of the field for three binoculars: left-CZJ 8x50 Octarem, center-Nikon 8x32 SE and right-Zeiss 8x42 FL. The purple fringe (red+blue) radiating in all directions is longitudinal CA. It’s quite obvious in the conventional Octarem and SE and slight in the FL. So, whatever glass is used in the FL does reduce longitudinal CA.

Most of the color fringe visible in the FL is transverse CA (lateral color) rather than longitudinal. Notice that the color is reddish in one direction, but bluish in the opposite direction. This form of CA appears in the photo either because the target is a little off center or the (hand held) camera is not quite in line with the eyepiece. The same kind of misalignment between eyepiece and eye is constantly occurring when binoculars are used hand held. The FL actually has a bit more transverse CA than the Octarem or the SE. The amount at the very center of the field is tiny, but it starts to become visible even at normal magnification only about 4-6 degrees off axis and becomes quite obvious closer to the edge of the field. Other expensive roof prism binoculars with complex objectives have considerably more transverse CA than the FL and all binoculars show more visible transverse CA than longitudinal. It’s the color fringing seen when dark birds are perched on a wire in front of a bright sky.

The bottom row shows the same three binoculars, but with the Octarem and the FL stopped down to the same 32mm aperture as the SE to simulate the stop down imposed by the eye in daylight. Now the Octarem is nearly as free of longitudinal CA as the FL. That’s because it’s stopped down from about f/4 to about f/6.7 (higher focal ratio = lower CA). The SE stays at f/4 so there is no improvement. The FL changes to about f/5 with slight improvement over CA that was already pretty well corrected. From this you can see that, for daylight uses, large exit pupil binoculars like 7x42, 7x50, 8x50 and 8x56 have always had low longitudinal CA, quite comparable to modern “ED” binoculars.

[FrankD]

Henry,

I thank you for the explanation. A question though....when I look at a distant horizon outside my home (mountain line against a light colored sky) and move the binocular so that the edge of the mountain is at the top edge of the field of view (might be a red/purple form of CA along the mountain/sky edge) and then move that same mountain line to the bottom edge of the field of view (usually the opposite CA color blue/green is present) then am I looking at lateral or longitudinal CA?

I always thought it to be the latter but after reading your comments I think it is the former.

Thanks.

[Steve C]

Thank's for that Henry. That looks to be very useful. Simple too.

[ronh]

Henry,
That is a great lesson! It's fortunate that the thin shadow line is visible, running across the target. It shows that this really impacts definition. We're sometimes led to believe that longitudinal CA is insignificant, and this demonstration shows that this is clearly not the case.

Still, even when I try to as hard as I can to center my eyes, transverse rears its ugly head DISTRESSINGLY close to the center of the field. Within a bird width, to be exact! Henry, do you think that transverse or longitudinal is the worse practical centerfield offense against view quality?

I have a quick non-quantitative test for longitudinal CA. Look for the difference in the best focus positions looking at a traffic light. All my binoculars (none ED) focus further back for red, as expected from a yellow-green optimized achromat objective. At night expecially, the difference is easy to observe.

In addition to the CA effects, your photographs also reveal contrast differences, well away from the color blur. The SE looks a little lighter black than the other two, which are about equal. Somebody "might" want to apply the newly developed technique to measure that in these photos.

Thanks for the good work. I will be more annoyed than usual by CA for a few days.
Ron

ps, I deleted and rewrote an earlier response, to try to make more sense.

[marcus]

Is chromatic aberration the 'donut' that I'll see around the edge?

[henry link]

Originally Posted by FrankD

Henry,

I thank you for the explanation. A question though....when I look at a distant horizon outside my home (mountain line against a light colored sky) and move the binocular so that the edge of the mountain is at the top edge of the field of view (might be a red/purple form of CA along the mountain/sky edge) and then move that same mountain line to the bottom edge of the field of view (usually the opposite CA color blue/green is present) then am I looking at lateral or longitudinal CA?

I always thought it to be the latter but after reading your comments I think it is the former.

Thanks.

Frank,

What you are seeing is transverse CA or lateral color, different terms for the same aberration. If the eye is perfectly aligned with the eyepiece lateral color is completely absent for a small area in the center of the field and becomes increasingly prominent toward the edges. It looks and acts differently from longitudinal CA in several ways.

Firstly, it forms only on edges that face toward the center or the edge of the field and shows a different color on each side of an object. Longitudinal CA shows the same purple fringe radiating in every direction from an object and is just as prominent in the center of the field as it is toward the edge.

Transverse CA is also affected by pupil position (longitudinal CA is not), so that it may appear at the very center of the field if the eye's pupil is misaligned and at the same time disappear from an area on the opposite side of the field from the misaligned pupil. Since objects are constantly moving and the pupil constantly decentering in hand held binoculars, the appearance of transverse CA has a flickering, now you see it, now you don't, quality which I think explains most of the subjective disagreements as to how much CA particular binoculars exhibit.

Henry

[henry link]

Ron,

I remember your stop light test from Cloudy Nights.

The only time I've been able to definitely see longitudinal CA as a color fringe in low power binoculars is looking at Sirius or Venus (if it's high enough to avoid atmospheric color). But I see lots of transverse CA in bright daylight and, as you say, distressingly close to the field center in some binoculars. In those I feel like I can't achieve a completely relaxed sharp image because of it.

Please don't take the photos I posted too seriously. They were done (quick and dirty) only to show an approximation of the CA I can see at the eyepiece with boosted magnification. I needed to adjust the exposure so that the highlights wouldn't be blown in the 8x50 and 8x42 images, so there are brightness differences from that and focus was very critical. The SE might not be in quite as sharp focus as the other two. BTW, that shadow is a good example of single line detection vs resolving two lines. It's probably no more than 4-5 arc seconds wide, but I can easily see it at 8x. My ability to separate lines at 8x would be more like 11-12 arcsec.

Henry

[henry link]

Originally Posted by marcus

Is chromatic aberration the 'donut' that I'll see around the edge?

I'm not sure. At the very edge lateral color mixes together with other edge aberrations like astigmatism and field curvature.

[michaelboustead]

Henry

Color fringe is a problem in high contrast situations? My Swarovski 8x32 seemed to have a lot more problem hawk watching than my usual birding in forest and forest edge.

Mike

[henry link]

Mike,

Yep, but if you really want to look for it you'll find it in dappled sunlight in the woods or branches against the sky.

Henry

[henry link]

Here's one more image of my test pattern which shows the progression of off-axis transverse CA.

The center of the field (Zeiss 8x42 FL) is the cross on the left. The slight CA surrounding it is longitudinal. Each white bar is placed at 2 degree intervals. In this enlargement you can already see a little transverse CA on the first bar, just 2 degrees off-axis. Notice that the left side of the bar is red and the right side blue. This photo looks worse than what you would seeing looking through the binocular but is accurate in showing the steady increase in color fringing away from the center. At normal 8x magnification in sunlight, with my eye perfectly centered, I can't detect the longitudinal CA on the cross. I begin to notice a narrow color fringe of transverse CA on the bar placed 4 degrees off-axis.

[ronh]

Henry,
Thanks for another fun photograph. Yuk, in fact, and that's a Zeiss FL! Please don't do this to a Leica Trinovid like mine, okay?

Not long ago, you tried Stokes' test on a Zeiss FL, and I remember you got a result quite different from what is expected if the transverse CA was dominated by an achromatic objective. In that case, the lateral color, like the longitudinal, would be in the form of a folded-back or secondary spectrum: yellow and green on one end, red and blue on the other.

It isn't surprising that the Zeiss FL is not much like an achromatic astronomical telescope with a simple narrow field eyepiece, but this photograph gives some insight to what's going on. It shows primary spectrum. Curious! In my normal glass binoculars, the fringe is yellow-green on one side, and purple on the other. So it seems like either the objective of the FL is fancy pants indeed, not just a compressed-spectrum achromat, or this color effect is from the eyepieces, which in some sense are not "achromatic". Do you know where this color effect arises?
Ron

[FrankD]

Thank you for the explanation and the pictures to go along with it. I do believe I have a better understanding of the two terms and how they affect the images in the various models I have on hand.

[elkcub]

Originally Posted by henry link

... Longitudinal CA shows the same purple fringe radiating in every direction from an object and is just as prominent in the center of the field as it is toward the edge.

... Since objects are constantly moving and the pupil constantly decentering in hand held binoculars, the appearance of transverse CA has a flickering, now you see it, now you don't, quality which I think explains most of the subjective disagreements as to how much CA particular binoculars exhibit.

Henry

Henry,

Your demonstration certainly shows that purple is prominent and radiates in every direction. It might be worth mentioning/emphasizing however, that purple is not a real color in the sense that there is such a spectral frequency. Rather, it is the combination of red and blue, which lie at opposite ends of the spectrum. This combination is perceived as purple.

I was struck by the fact that the upper left panel (CZJ) clearly suggests (to my eye) two layers of color in the fringe, one being more reddish than the other. This is largely missing in the lower panel. It then occurred to me that I might use a feature in the Art Director's Toolkit (ADT) to tease out the composition of these two color layers, and also determine how they might blend perceptually if seen from a sufficient distance. The same analysis was done for the lower left panel.

In the attachment, Color 1 was picked off the layer near the upper white horizontal edge, and Color 2 from the next layer out. (The whites and blacks vary systematically, incidentally, but that's another issue.) The "Calculation" in each panel shows a 50/50 combination of the two layers, and the RGB weightings respectively. The total width of the fringe is roughly 4-5x larger for the upper panel than the lower, incidentally, so a purple fringe would be much more prominent at a given distance.

For the upper panel, the first fringe layer is close to an equal red-blue combination, but the second layer is darker and blue dominant. This may be because of the camera's focus. Ideally, to minimize color fringing, focus would be mid-way between red and blue, — as the eye does by accommodation to neutralize chromatic blur. My guess is that the camera was focused on the far red side, thereby increasing the size of the blue blur, and creating the appearance of a second layer. If this reasoning is correct, I guess it highlights a cautionary note for using a camera to characterize (or predict) perceptual effects. Unlike the camera, visual accommodation (i.e., focus) changes continuously.

In the lower panel there is less evidence of double layers, and the fringe retains a similar color balance from the white border outward. The purple simply darkens on the way out.

I agree with your second observation above, and would add that even with a stationary subject the eye is also constantly moving, not just angularly but also by way of nystagmus, accommodation and pupilary responses. Part of the reason that observers differ about CA effects, I think, is also because they are all somewhat different in these oculomotor behaviors.

Ed

[henry link]

Thanks for your comments and measurements, Ed. I also noticed the greater defocus of blue compared to red giving the Octarem CA a two layer look.

I can tell you it's pretty difficult to accurately focus a camera while pointing it through the eyepiece of a binocular, so it's probable that I inadvertently focused the Octarem closer to red. I assume my eye seeks best focus around the point of maximum photopic sensitivity (440nm-460nm), but focusing is crude on the ground glass of a viewfinder so I can easily imagine missing. It's also possible that blue really does swing farther out of focus than red in the Octarem's optics. Either way I only hoped these photos might crudely illustrate some principles that are hard to describe verbally, so I was happily surprised by how close they came to matching what I actually see through the binoculars at boosted magnification.

Ron,

I've noticed that the colors of transverse CA vary with different binoculars, but they are always nearly complimentary: yellow/purple, red/green, orange/blue. I'm open to suggestions for what causes transverse CA in binoculars. We know some of it comes from the eyepiece, but that doesn't seem to fully explain why some binocular designs have more than their fair share. Those designs seem to have a few things in common. They use complex objectives (usually 4 elements including a moving focusing element) and roof prisms. At the moment my money is on the complex objectives.

Henry

[ronh]

Henry,
A reflecting telescope is a case of pure eyepiece induced color. Naglers handle fast mirrors well, and have impressive wide and astigmatism-free fields, but they're among the worst for lateral color, magenta and chartreuse, a putrid clash but complementary. It's not anything you'd notice on dim deep sky type targets.

The only eyepiece that's completely free from lateral color by design is the 300 year old Huygens, and the penalty is a narrow and deeply curved field, and good only at long focal ratios, quite the opposite of the Nagler. Huygens observed planets with ridiculously long singlet lensed refractors that were hard to track, so he must have thought they were really the stuff.

So, it's the usual set of tradeoffs to fit the intended usage, etc.

Back to birdwatching, Zeiss seems to have greatly improved color without messing anything else up, and at competitive prices, an enviable achievement. How much longer can I resist?
Ron

[elkcub]

Originally Posted by henry link

Thanks for your comments and measurements, Ed. I also noticed the greater defocus of blue compared to red giving the Octarem CA a two layer look.

I can tell you it's pretty difficult to accurately focus a camera while pointing it through the eyepiece of a binocular, so it's probable that I inadvertently focused the Octarem closer to red. I assume my eye seeks best focus around the point of maximum photopic sensitivity (440nm-460nm), but focusing is crude on the ground glass of a viewfinder so I can easily imagine missing. It's also possible that blue really does swing farther out of focus than red in the Octarem's optics. Either way I only hoped these photos might crudely illustrate some principles that are hard to describe verbally, so I was happily surprised by how close they came to matching what I actually see through the binoculars at boosted magnification.
...
Henry

Not to quibble, Henry, but I believe max photopic sensitivity is about 555 nm, which registers perceptually as green. No?

If there is any interest in such esoterica, the mechanism for the eye's unconscious vergence control seems to be based on minimizing blur, rather than optimizing color response. For this reason, focusing on monochromatic light doesn't improve accuracy over focusing on white light, nor does neutralizing the eye's axial CA with a corrective lens. In other words, the oculomotor feedback mechanism uses the blur from axial CA to adjust focus. It's a fair and general assumption that it also uses the same mechanism when dealing with the image from any coherently coupled optical instrument.

When focusing an incoherently coupled optical instrument, such as the image on the ground glass of a camera, I don't know what criteria are used, but minimizing retinal blur would certainly be a possibility. In this case, however, it is also possible to impose any conscious criterion one wishes, — which highlights an essential and important difference between coherent and incoherent coupling.

Photographic methods provide a valid understanding only up to the point where unconscious oculomotor factors assert an influence on image perception.

Ed

[Steve C]

How much of the observed CA might be attributed to the camera?

[henry link]

Originally Posted by elkcub

Not to quibble, Henry, but I believe max photopic sensitivity is about 555 nm, which registers perceptually as green. No?

Yes, thanks for the correction Ed. My brain knew to hit the "5" key, but my hand disobeyed.

As for how the eye/brain reaches focus when it's presented with chromatic aberration I bow to your greater knowledge of how that works.

I agree about photographic methods. I've only posted photographic examples when they correspond pretty well to what I've seen at the eyepiece.

Steve,

I don't think the camera adds much. The photos look pretty similar to what I see at the eyepiece if the magnification is boosted to 64x. Remember the camera lens (80mm) is stopped down to the exit pupil of the binoculars which would make it f/20 for a 4mm exit pupil. Notice how little longitudinal CA there is in the Zeiss FL photos. The camera must add less than that.

Henry

[Surveyor]

Henry,

I have been following this discussion with interest and, quite frankly, a lot of surprise.

From your description of 2 degrees separation, I assume the white bar width to be 0.5 degree. Is this in object or image space (AFOV)? What are your target dimensions? I think I will make one to match yours.

I say surprised because the dispersion, if the bar is 0.5 degree wide, appears to be almost .1 degree wide (on the upper left image) and even wider off axis. Is this what you see with the eye (what ratio of CA to bar width do you normally see)?

TIA
Ron

[henry link]

Ron,

I'm afraid those photos are pretty useless for accurate measurements. I only intended for them to be illustrational. The second one (transverse CA in the Zeiss FL) was made with a different focal length lens and at a different distance from the target compared to the photos in the first post.

In the photos below below I tried to do a little better, but these photos can't be compared to the others because the sunlight is coming from a more oblique angle. The white tape should be about 30 arc minutes wide in image space (AFOV). The left photo is a Nikon 8x32 SE at full aperture. You can see that some lateral color has crept into the image even though I tried to get the cross perfectly centered. Perhaps it shouldn't be too surprising just how much longitudinal CA there is conventional binocular optics like this. This is are after all an f/4 crown and flint doublet. It really doesn't even qualify as an achromat.

The right photo is the same binocular stopped down to 22mm to simulate the effect of the stopped down eye on the binocular optics in the bright sunlight of the test set up. Thanks to the change in focal ratio the longitudinal CA cleans up pretty nicely. I think we can tolerate the very colorful optics of binoculars because the magnification is so low and also because we mostly use binoculars stopped down in the bright lighting conditions that would make the full aperture CA really visible. It may seem odd, but at normal magnification in this brightly lit test set up I can't see the color fringe of the longitudinal CA at all.

Lateral color is a different story. I see lots of it on the test target at normal magnification.

If you want to make a similar target, I used a piece of 1 1/2" lattice from Lowes, painted flat black. The white strips are 1/2" plastic tape placed at 2" intervals. I use the 8x32 SE (60 degree AFOV with minor distortion) to calibrate the distance needed for the separation between the leading edge of each white strip to represent 2 degrees of AFOV in an 8x binocular.

Henry

[ronh]

Does the camera's color sensitivity curve match the eye's?

[Surveyor]

Henry,

Thanks for the commentary. I have a lot better handle on it now and got me to thinking of tests I routinely make now. I went back and looked at some photos from the past year or so. If I give this some thought, there is probably a lot of CA information to be gathered from the existing tests, if I take the time to photo them or make notes about what I see.

Attached are photos of my positive USAF glass used for MTF target, just set up at 8x, not mounted in the collimator.

Next was a picture of the negative USAF, used for resolution measurements, when I was checking the Baigish porro but I did not try to keep the picture on axis. Looks as if the axis would be somewhere in group 3. For reference group 5, element 1 is about 8 arc seconds per line.

The third picture is through the collimating lens at about 90-95x. The optical axis of the Monarch 8x36 is the red x, the optical axis of the collimating tube is the center of the 5 arc minute grid.

Thanks for the idea, it may lead me to better understanding.

Have a good day.

Ron

[Surveyor]

Henry, your CA target reminded me of a program that I tried 4 or 5 years ago for MTF but could not get it repeatable enough on my test bench. The program did have a interesting side application but since CA does not bother me, I never looked into it. From memory, the program would take an edge and extract some CA information from it. I never learned to use it.

I have attached a sample plot from a LXL photo and a txt table, the bottom portion of the text file applies to the CA part. It is basically the same as your idea.

I will look for the description and maybe some instructions and post or email to you next week.

I am not going to take time to get into it again right now, I want to finish the three projects that have been dragging on for too long now.

Have a good day.

Ron

[elkcub]

Ron,

I can read your text file into Excel quite easily as well as SPSS. What do the column headers mean exactly?

Cy/mm,LW/PH,MTF(nchan),MTF(corr),MTF(R),MTF(G),MTF(B),MTF(Y )

and

x (pixels),R Edge, G Edge, B Edge, Y Edge,CA (max Delta)

How is the edge boundary and pixel size around the edge determined?

I assume these data are from your 8x20 LX L. Right?

Thanks,
Ed