Understanding field curvature, please

[Chazphoto]

Advertisement (gone after registration)

Dear Leicaphiles

The discussions on the APO-Summicron 50mm ASPH emphasise it’s flat field esp compared to the curvature on the 50mm Summilux-ASPH.  So I went looking to understand this better and have emerged with some comprehension and some questions.  Googling led me to a few optical scientific resources which didn’t help a lot and to Roger Cicala’s two-part article on DPR, which did.  Best of all, in the sense of being most practical, was our brother in light (at least in the past) Tim Ashley who did much more hard learning than I and was kind enough to distill it and have a guest note from Mr Cicala to top it off.  

Based on atom’s excellent work, I realised the obvious thing we all knew from high school optics, which is that the distance from centre of lens to centre of subject is shorter than from edge of subject to centre of lens.  In effect, the simple lens he posits, maintains focus in a concave curve and the complex lenses we all use seek to flatten that as much as possible (amongst many design goals).  

So, with as flat a field as the 50AA achieves, what is the effect on a photograph?   I assume that photographing a building at right angles to the camera at f2 , is where the flat field would shine.  But I don’t know if the field of the Summilux is flat at f2 at the same distance from the subject that you’d have to be at to get the shot.  

Secondly, how does the flat field work if you photograph a receding vertical view, in a diagonal across the frame?  I was inspired to think of this by a photo in the 50AA images thread, where resolution is maintained throughout the bricks of the houses on that diagonal.  I assume you’d have to stop down anyway, but does the flat field play a part in this scenario?

Third, what about a landscape.  Inherently more variable distances than architecture, but still benefiting from high resolution across the frame. Tim Ashley has some illustrations of this exploring the runway shape field of the 28mm he’s using.   If that lens had been the 50AA, what would have been the result?

Any comments gratefully received, including that this is over-thinking it!

cheers

 

Chazphoto

[Robert Blanko]

The odd effect of field curvature is most obvious if you have a blurred background and the blur depends on the field curvature. For instance the Voigtländer 50/1.0 may show this mostly unwanted effect under certain conditions, see here where the blur in the background is reduced at the corners https://www.fredmiranda.com/forum/topic/1741370/3.

Same with sharpness. One may need to refocus to achieve maximum sharpness in the corners - while reducing sharpness in the center and vice versa.

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[Chazphoto]

Hi

Yes, Roger Cicala’s article in DPR was clear enough about the way one could choose focus to maximise sharpness if you know the field curvature of a lens at a particular aperture.  I’m far from sure of how one would do that in practice and pretty sure that I’d rather not!

I’m really not clear whether field curvature changes with focused distance.  For example, people say that the curvature of the 50mm Summilux ASPH is helpful for portraits but those would be at 0.7 to 1.5m.  Would it exhibit field curvature if photographing a building straight-on at 10m and f5.6?

What seems clear is that the APO-Summicron will not have field curvature at either distance.  I would like to know how that and the two second examples I gave interact    More out of interest to learn than anything else.

cheers

Chazphoto

 

[250swb]

31 minutes ago, Chazphoto said:

 I would like to know how that and the two second examples I gave interact  

 

 

Field curvature applies if the photograph isn't very interesting, and then at least you have field curvature as a fall back discussion.

[Robert Blanko]

My gut feeling is that field curvature is hardly noticeable at apertures of 5.6 or more and at infinity, as it could be hidden by the comparatively large depth of field - i.e., the DOF is so large that field curvature would not make a significant difference.

[jankap]

See this one, if you want to try the effect: http://www.rokkorfiles.com/24mm VFC.htm

Edited May 29, 2022 by jankap

[Chazphoto]

Advertisement (gone after registration)

Well, that was interesting.   I had a Minolta Dynax 9 for years and loved some of the high performance and unusual lenses that Minolta had produced over the years, including the variable soft focus 135mm, which I never had   I didn’t know about the variable field curvature lenses!

However, I asked our friend Google a slightly different question about this subject and found some useful articles: one from PetaPixel and one from BH Photo (will post links when I’m on a pc) which help by answering the question of when would you want to use a flat field lens - and by implication when not.   From those articles, I take the following points (errors being mine not theirs).

1.  Flat field lenses traditionally used for macro lenses to enable reproduction of flat artwork   Ditto enlarging lenses, microscopes, other scopes etc (and lots more learned articles on those than photography!).

2.  The comment is repeatedly made that these lenses tend to be high contrast and high resolution and it isn’t explained whether that is a consequence of the flat field or designed-in as well.  Therefore, the usual warnings about the risk of using such lenses for portraits.

3.  The BH Photo article gets to the point about different DoF effects with flat versus curved field lenses, including at different apertures, with examples   Hard to really see the effect on the iPhone, but the background treatment is clearly different.  The article leaves it to the viewer to determine which is preferable aesthetically.

4.  This, plus clues from the Karbe interview on the APO 50mm, leads me to conclude that relative to aperture, DOF falls off faster with the flat field lens.  He talks about this as the plane of contrast having a faster transition to out of focus areas, but that seems to me to be two ways of explaining the same thing.  In other words, to deal with my question 2, you’d have to stop down further with a flat field lens to get the diagonal row of houses within the DoF - unless you know your lens’s field curvature and do something clever with where you focus to take advantage of it.  This sort of also makes sense to me if we think of flat field lenses as slicing a vertical plane through the image, whereas a curved field lens takes a shallow crescent (assuming a mild case of simple concave curvature) - the difference between a madeleine and an ice cream scoop.  As you stop down they both give you more DoF, but the softer scoop shape grows faster, and less predictably. 

5.  The PetaPixel article has descriptive examples of how you might do this (not photos).   They confirm my view that this would be hard work; more akin to using a view camera on a tripod than a Leica M (though with live view and a touchscreen you’d get there okay).

6.  With macro lenses we were always told to stop down and I always forget to do that enough!  When I use the OUFRO with the Summilux 50mm pre-ASPH I struggle to get the subject in crisp focus (handheld and breezy conditions, often) but the curvature issues, obviously now, play a part and also yield a pleasing background.

7.  That field curvature is more an issue with wides than teles makes sense also, if I think of the simple lens diagram and consider the narrower field of view - naturally less angle in which the curvature can develop.  

Cheers

Chazphoto

 

[Chazphoto]

Useful articles here:

https://petapixel.com/2016/12/21/field-curvature-tricky-problem-photography/

and here:

https://www.bhphotovideo.com/explora/photography/tips-and-solutions/flat-field-lenses-and-why-they-matter-when-shooting-close-ups

Chazphoto

 

[Chazphoto]

I was driving back from walking the dogs the other morning and passed an old school that has been converted to a house; a building I drive past nearly every day.  It has a slightly churchy feel, with a modified cross shape in plan.  It occurred to me that photographing the two arms of the cross, so that the centre of the house was in the centre of the photo, would be an example of where a concave field would be useful, because the image would retain sharpness as the two wings of the house come towards the camera.  The trouble is, how do you know that your lens has a curve which approximates to that 90 degree angle of the walls (45 deg each side of the perpendicular) and extends far enough towards the camera?  Although I struggle to understand how, it is clear that field curvature can be convex from the camera's position (useful for keeping spherical objects in focus, I gather).  In my worked example of the cross-shaped house, if I focus at the centre of the image and junction of the walls, a flat field lens with zero DoF would not keep the wings in focus, but extending the DoF would do that without the mental gymnastics of learning the field curvature of your lens.  

May be all this doesn't matter a lot.  I don't usually notice a photograph of a building where I have out of focus wings due to these issues.  As Robert says above, stopping down seems to fix most things, but with higher resolution sensors and screens and prints, may be not so much (and I suspect the view camera masters of yore, did pay attention to these issues).

Last note: I wrote "madeleine" two posts above, meaning mandolin and my iPhone keeps auto-correcting "lens" to "Len's" which is jolly irritating.

Cheers

 

Chaphoto

[Chazphoto]

I have no idea if this set of notes is more to myself than of use to anyone (and I claim no expertise) but it is great to be able to record my thoughts here.

I used a APO-Summicron 50mm for a few days this weekend past.  I was taking alternative photos with it and my pre-ASPH Summilux 50mm.  I noticed a lot of differences but the flatness of field wasn’t one of them.  I was reminded, with one test, of how little DOF you get with a 50mm at a metre or two distance to subject.   I was photographing some foxgloves in a diagonal line receding from the camera and, focusing on one, I stopped down to try and get more in focus.  DoF increased but not enough to bring every stem into full (apparent) focus.  I did not experience any material difference in this between the two lenses, though I still am left with the feeling  from the four days that the 50AA DoF falls off faster than the pre-Asph.  Any field curvature with the pre-ASPH didn’t make a difference with this line of subjects, so the net result was that the flat field of the 50AA didn’t matter one way or the other.  

As a comparison, I photographed some buildings at f5.6 and (for material purposes) infinity and with a midpoint further away from the camera than the leftmost building (though that would be 15-20m away anyhow).  There was a lot of patterned brickwork of various types to use in order to check resolution.  The APO lens did resolve more detail, generally, particularly towards the edges of the frame though the pre-ASPH wasn’t far behind.  I did not notice field curvature or flatness having any impact one way or the other.  In fact, the most material difference was the colour tone of the new lens compared to the older (re which a separate post will follow on another thread).

Where has all this field curvature learning and thinking left me?  More informed on what is involved and why it might matter.  Confirmed in my prior, uninformed, view that it doesn’t impact my kind of photography that much (noting we’re talking M10 here for resolution).  However, for a wide field of view lens, a flatter field than is usual for that focal length may be useful.  Also aware that the 50AA has virtues, of which / to which the flat field will be a contributor. 

best

 

Chazphoto