Is curvature of field a good thing? (3 Viewers)

[Eric_D]

First, I do not mean angular magnification distortion (AMD): that is perceived as straight lines bending as you move them away from diametrically across the center of the field, towards an edge. Not pincushion or barrel distortion. That is not 'curvature of field', IMO.

Curvature of field is when you have to refocus if you move the binoculars such that the target is moved from the center to the edge of the field of view.
If you don't try refocussing, it can be perceived as unsharpness towards the edges, or a small 'sweet spot'.

In astronomy, everything is at infinity, near enough, so yes, they want a flat field. (After all, the night sky is a flat plane ... no ?)

Even photographers, when they set the lens at "50 meters", do not expect 50m object distance to be sharp from edge-to-edge: that would be a spherical focus surface, yet lenses are designed and tested on plane (flat) objects like brick walls.

Some binoculars, notably Zeiss Jenoptem, have very noticeable curvature of field, such that nearer objects away from the center are in focus.
Is this a bad thing? Have they deliberately left it in, perhaps to avoid compromising other things (FOV, off-axis sharpness)?

We rarely look at flat surfaces from a straight-on perpendicular angle.

Usually we are standing on a horizontal plane, where the foreground, below the horizon, is progressively closer to us as we look downwards.
I would even argue that we spend more time in streets and valleys, where things to either side are closer than central, distant objects.
We spend less time on ridges.

So "Is curvature of field a good thing?"
I have informally tested many (predominantly mid/low-end Japanese late-20th century Porro I ) binoculars, and found very few with the opposite, undesirable field curvature.
Is focussing closer at the edges seen as beneficial, a bit like the preference for pincushion rather than barrel distortion (AMD)?

Is it perhaps welcomed as (pseudo) 'Depth of field'?
I do not believe binoculars with the same '8x30' specification can really have greater or less depth-of-field: it is determined by the '8x30' numbers, lighting conditions and the user's eyes.

The entire discussion of edge sharpness often fails to recognize the most important optical element in the system: the human eye. Great eyes probably don't see much edge softness, whereas eyes that have lost the ability to readily accomodate see it immediately. I think the same can be said for our differing FOV perceptions.
John

 

[A2GG]

I don't think it's a bad thing; it doesn't bother me unless the sweet spot size is noticeably small and blurring is really bad outside of the sweet spot.
I don't look at the edges and re-focus to get them sharp. I frame the subject in the center and focus there. I mostly ignore the edges unless something catches my eye on the periphery, then I'll quickly move the bino to center it. I've used flat field and non-flat field binos and can get along with both. I went from a binocular with a 435ft fov to 372ft (I use now) and I adjusted. I'm more picky about other things like image color/tones, ergonomics and weight.
I don't get too concerned with flat field vs blurry edges, FOV size, CA and things like these that others may feel are important.

 

[Binastro]

Space is essentially curved, but so distant that it seems flat.

Photo lenses definitely try to achieve flat fields.
Even a well made Cooke triplet has an almost flat field.

Projecting slides from Gauss type 5, 6 or 7 element lenses the whole field, say 4ft or 5ft wide can be flat.

Professionals used large lenses stopped down to say f/32 to give flat fields.
I remember seeing a scene of a distant market place where the resolution was superb in an enormous image.

With digital photography, the lenses are redesigned as the emerging rays need to be almost perpendicular to the sensor, whereas film lenses can have angled rays as the film grain copes with this.

As to binoculars, I use the bottom of the curved field to view closer object without changing focus. It also involves vignetting.

Regards,
B.

 

[Eric_D]

From 'The End All FL Edge Sharpness Thread' page 4:

"projection onto the inner spherical surface of the eyeball (over a 60-70 deg. field) is not at all the same as projecting an image onto a flat surface, e.g., a camera sensor or movie screen.
... optimal design probably should (or does?) make use of a certain amount of field curvature, but these books do not address visual scanning or dynamics. "

The Eye-Binocular System

For myself, I find it difficult to get absorbed with measuring properties of different binoculars, when we somehow never get around to the underlying issue, i.e., the quality of the integrated eye-binocular system. Can it be assumed that by measuring the optical properties of the instrument alone, we can arrive at a refined understanding of how well the eye-binocular system works? Or, by making such measurements, can it also be assumed that optical design criteria were used that are independent of the observer's eye? In my opinion, the argument is illogical, i.e., post hoc, ergo propter hoc. We assume that the designers used what we find it convenient to measure.

A recent demonstration of the necessity to consider both the instrument and the observing eye was clearly demonstrated by Holger Merlitz regarding the design criterion for distortion. In our present discussion about astigmatism and field curvature, it might be noted that projection onto the inner spherical surface of the eyeball (over a 60-70 deg. field) is not at all the same as projecting an image onto a flat surface, e.g., a camera sensor or movie screen.

Although I am unaware of formal design guidelines, several authors on visual instruments have suggested that optimal design probably should (or does?) make use of a certain amount of field curvature, but these books do not address visual scanning or dynamics. I do know that the formal computer evaluation of image quality invariably includes the use of a sophisticated 'model eye'; hence, it's not much of a leap of faith that the same model is also used in off-axis design. Based on the facts presented by Henry (and Kimmo), and knowing that I don't need focus adjustment on the edges of my 8x32 SE, I would hypothesize that the field curvature is fairly well matched to my eyes.

Just comments, of course, and only intended for discussion.

Ed
PS. Like Kimmo, I also favor eliminating astigmatism, which as far as I can tell serves no useful purpose.

"PS. Like Kimmo, I also favor eliminating astigmatism, which as far as I can tell serves no useful purpose."

I had wondered whether I could rotate lenses in a pair of cheap/faulty binoculars to try to cancel astigmatism?
That kind of tuning could even be personalised to people whose eyes' prescription includes astigmatism?

There might be a downside: either anamorphic stretching with a different magnification horizontally and vertical; or different magnification between left and right eyes?
I have seen the latter in quite a few pairs of binoculars, including reputable makes/models. I have read of astronomers choosing unmatched eyepieces to give the night sky a '3D effect'!

 

[Canip]

I use the bottom of the curved field to view closer object without changing focus.

Exactly, and this effect allows companies such as Steiner to claim, e.g., that their IF Navigator 7x30 features a "sports auto focus", which is complete rubbish of course, but nevertheless it works well in the way described by Binastro, i.e. the strong field curvature allows observing closer objects without having to refocus by using the bottom of the FOV.

 

[Eric_D]

I found an interesting reference to curvature of field. Several people report it (or unsharpness) is not symmetrical if you rotate binoculars about their optical axis: sometimes left and right edges behave differently from top and bottom. This could just be accidental decentration: the sweet spot is off-center. But is it consistent between left and right eyepieces, and different samples of the same model?

Here ronh reports low curvature at the bottom of the field of his Fujinon FMT 6x30, even though it would arguably be a good thing there!

My 6x30 is weird. It has a 53 degree field, like the 7x50, but the eyepieces are not the same, being considerably smaller and with less eye relief. The funny thing is, I measure its curvature as 2D center to edge in the top of the field, but the bottom is almost perfectly flat, with a very sharp view all over. I notice this asymmetry only in the daytime, so it must depend on pupil vignetting somehow. I'd say it was screwed up, except both sides exhibit the same odd behavior, and the central view is about as good as anybody could ever want. So I play this funny correction to advantage, using the top part to view closer objects when I'm in a hurry, and luxuriating in the bottom half's large sharp area. Strange though, huh?
Ron

Henry Link reports the asymmetry does not follow rotating binoculars, but is eye-related: human factors.

Bill,

I've noticed the same difference between horizontal and vertical edge performance. In fact I've seen it to some extent in every binocular when I've looked for it. I don't think it is a characteristic of any particular binocular's optical quality, but has to do with differences in the way the pupil of the eye aligns with the exit pupil depending on whether we look to the side or up and down. To simplify things try using only the right eye and the 3 and 6-o-clock positions. I find that even if I rotate the binocular 90 degrees I still see worse edge performance at 6-o-clock. My speculation has been that looking to the side cuts off more of the opposite side of the exit pupil than looking down does. That means that more of the 9-o-clock side of the objective is blocked from forming edge astigmatism at 3-o-clock than the top of the objective is blocked when looking down.

I tried testing this idea today by blocking off parts of the objective of the 8x42FL. I found if I masked the top half of the objective the edge performance at 6-o-clock improved slightly but when I blocked off the lower half it improved much more to approximately equal the 3-o-clock position. Then I tried masking the 9-o-clock side of the objective and saw no significant improvement at the 3-o-clock position, but saw some small improvemant if I masked at 3-o-clock. In the end I found that masking ANY side of the objective improves edge performance somewhat. All this tends to suggest to me that the partial blocking of the exit pupil by the eye pupil position has a big effect on what we see when we try to examine the edge of the field, but I'm a long way from understanding just how it works.

The fundamental difference I see between the edge performance of the 8x32SE and the 8x42FL is that field curvature dominates in the SE and astigmatism dominates in the FL. You can focus a star to a pretty good star point at the edge of the SE, but the same star at the edge of the FL cannot be brought to a good focus. Instead it goes through the classic pattern imposed by astigmatism. The star is seen as a vertical line one one side of focus, horizontal on the other side and forms a little cross at "best" focus. BTW the edge performance of the 8x42FL is really the only thing that can be "star tested" at 8X. To see the true diffraction pattern of a centered star requires about 30x (60x is better) if the objective size is 42mm.

Henry

 

[Ignatius]

There are persistent rumours that if that annoying human factor could be taken out of the equation, earth would generally be a much nicer place and most problems would disappear or be solved as a consequence.

 

[jring]

Hi,

field curvature probably has to do sth with how the brain computes 3d images apart from stereopsis.

There was a very interesting video on youtube arguing this point quite well by showing footage of the same scenes shot with classic Cooke triplet cine lenses and corresponding modern Leica flat field lenses and the former indeed looked more 3d despite the fact that they obviously were images on a flat monitor and since this was from a regular one eyed cine camera, no chance for stereopsis or what we know as 3d effect in binoculars...

Unfortunately my old youtube link does not work any more (says the video is private)... so no need to search for my post from 7 years ago...

If somebody knows a current link for that video, please share...

Joachim

 

[Binastro]

It may be that the Cooke Speed Panchro f/2 was used.

The 50mm f/2 is I think a 7 element 5 group lens of double Gauss type.

These lenses, in many different focal lengths, were designed by Horace Lee in 1920 and were used in most movies for perhaps fifty years.

There are new production versions nearly identical to the ones from I think the 1950s but with modern housings.
Still made by Cooke optics, and used in the present day.

There are various series, but essentially similar designs with some more modern glass types.

The Deep Field Panchro is another very sought after lens.

The Cooke triplet is an earlier three element still photo lens.

There is also the Cooke triplet telescope objective, which is different.
I got the Zeiss Aero Triplet 120cm f/7 for Harold Ridley for his comet photography. This is a Cooke triplet type. Made in 1920, I think from the serial number.
Later versions had one element split, as the 3 element was too difficult to make and may have had some aspherising.
So this had four elements.

I did see the comparison video with the Cooke lens and Leica lenses.

Some of the Zeiss movie lenses are surprisingly low price, but I think made in Japan, possibly? by Cosina.

Regards,
B.