After shooting of a fisheye lens earlier this week, I noticed that what is in focus does not seem to lie on a flat plane as it would for a rectilinear lens. Now, I am aware that lenses usually exhibit some field-curvature but in my experience it is minimal compared to what I see with a fisheye lens.

Does a fisheye lens inherently focus on a curved-surface or a flat plane?


1 Answer 1


The difference between a wide angle fisheye lens and a wide angle rectilinear lens is equal area projection versus straight line projection. Uncorrected, they will both demonstrate field curvature.

All simple lenses will demonstrate field curvature based on the angle of view the lens provides. Of course the sensor/film size is also involved in the angle of view yielded by a particular focal length. When used with a 35mm sized full frame sensor, an uncorrected long telephoto lens, such as a 400mm one with a field of view (FoV) of only 5° will have field curvature the shape of a 5° arc of a sphere. An uncorrected lens with an FoV of 45° such as a 50mm one will, likewise, have field curvature the shape of a 45° arc of a sphere. As you can see, by the time a lens such as an 8-15mm fisheye is considered, the FoV on a FF sensor approaches 180° and the field curvature of an uncorrected lens would be the shape of half a sphere!

Most lenses used by modern photographic equipment are not simple lenses. They are compound lenses with several elements combined into several groups. Most of the additions beyond a simple lens are to allow for things such as close focusing, zooming, and to correct optical aberrations. One such aberration that is usually corrected to one degree or another is field curvature. This is fairly straightforward on a lens with a narrow field of view (FoV) because the curvature is much less pronounced than with a lens with a wide field of view. How much, if any, of the field curvature is corrected depends on each individual lens' design.

A lens such as the Rokinon 8mm T3.8 Cine Fisheye for Canon is corrected very well and yields almost a flat plane of focus when used on an APS-C sensor for a diagonal FoV of around 167°. At the other end of the spectrum, a single meniscus lens with the same FoV would have a very pronounced field curvature. Most designs are somewhere in between.

  • \$\begingroup\$ Does this implies that the field-curvature is influenced by sensor-size? Since the field-of-view is larger? This seems very surprising ! \$\endgroup\$
    – Itai
    Nov 21, 2013 at 13:42
  • \$\begingroup\$ Not really. The sphere is the same size for a particular lens. How much of that sphere is in the FoV depends on the size of the sensor. If you modify the shooting distance to compensate for the different sized sensor then you change the relationship between the focal length, focus distance, and the depth of field that will make the same amount of field curvature less noticeable. This is one of several reasons why the same lens on a FF camera demonstrates softer edges than when it is mounted on a crop body, assuming the center of the lens is what is focused on a flat target in both cases. \$\endgroup\$
    – Michael C
    Nov 22, 2013 at 6:44
  • \$\begingroup\$ Remember, this is in the context of an uncorrected simple lens. But the same holds true of a corrected lens in the sense that regardless of the shape of the plane of sharpest focus (whether spherical, flat, something in between, or even an irregular shape created by the correcting lens elements) the angle of view that is directly related to the sensor size determines how much of that shape is within the part of the image circle projected on the sensor. \$\endgroup\$
    – Michael C
    Nov 22, 2013 at 6:53

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.