Re: Canon Develops New Camera-Lens Optical Element
Yours is a legitimate question, but one that has been addressed in this discussion thread. I will explain again:
Chromatic aberration in photographic lenses for consumer use can be broadly classified into two types: transverse and longitudinal. Nearly all such lenses exhibit some degree of both to varying extent and in varying combinations.
The transverse type of aberration occurs when the image magnification of the lens is a function of the wavelength of light. That is to say, the apparent focal length of the lens is not precisely the same for different colors of incoming light. This leads to color fringing that increases in severity as one goes from the image center to the image periphery.
The longitudinal type of aberration occurs when the plane of sharpest focus is a function of the wavelength of light. That is to say, one range of the color spectrum are in focus at one subject distance but other colors will be in focus at other subject distances; or even worse, one portion of the spectrum is in acceptable focus but other colors may not achieve acceptable focus anywhere. This leads to color fringing that is generally uniform in severity across the image plane, but tends to be reduced by stopping down, since the effect of a high f-number is to increase the depth of field, thereby permitting the planes of sharpest focus to overlap each other throughout the visible spectrum.
Generally it is easier to correct the first type in post processing, if the characteristics of the aberration are measured for the lens that created the image. This is because in this type, the plane of sharpest focus is the same, the aberration is independent of f-number, and the extent of aberration is not dependent on the subject-camera distance. The basic method of correction is to apply an inverse scaling transformation of image pixels as a function of their color.
The second type is very difficult if not impossible to correct adequately in post processing, primarily for the same reason that it is not generally possible to correct for missed focus. Deconvolution or various other sharpening algorithms may work partially, but these generally create unpleasant visual artifacts. The essential information has been lost: you can't de-blur what is blurry without more information about how the light entered the lens, and this information is dependent on the distance of the subject to the camera, as well as numerous other factors. Efforts to correct longitudinal chromatic aberration have largely focused on attempting to process out the color fringing as they register in the image file, rather than trying to model the behavior of the lens. Sometimes this works convincingly well. But it isn't really a substitute for designing the lens to avoid such aberrations.
The takeaway is that chromatic aberration occurs whenever a lens is designed in a way that causes an image point to be spread out as a function of the wavelength of light. If you've ever seen the OLAF test images produced by LensRentals, this is exactly what it will show: A collimated white light source is projected through the lens and a spot is produced that is very rarely just a white smear (let alone a white dot). It's almost always a multicolored pattern, which shows the presence of complex chromatic aberrations. See here:
http://www.lensrentals.com/blog/2014/04/olafs-lens-art
keriboi said:
Yours is a legitimate question, but one that has been addressed in this discussion thread. I will explain again:
Chromatic aberration in photographic lenses for consumer use can be broadly classified into two types: transverse and longitudinal. Nearly all such lenses exhibit some degree of both to varying extent and in varying combinations.
The transverse type of aberration occurs when the image magnification of the lens is a function of the wavelength of light. That is to say, the apparent focal length of the lens is not precisely the same for different colors of incoming light. This leads to color fringing that increases in severity as one goes from the image center to the image periphery.
The longitudinal type of aberration occurs when the plane of sharpest focus is a function of the wavelength of light. That is to say, one range of the color spectrum are in focus at one subject distance but other colors will be in focus at other subject distances; or even worse, one portion of the spectrum is in acceptable focus but other colors may not achieve acceptable focus anywhere. This leads to color fringing that is generally uniform in severity across the image plane, but tends to be reduced by stopping down, since the effect of a high f-number is to increase the depth of field, thereby permitting the planes of sharpest focus to overlap each other throughout the visible spectrum.
Generally it is easier to correct the first type in post processing, if the characteristics of the aberration are measured for the lens that created the image. This is because in this type, the plane of sharpest focus is the same, the aberration is independent of f-number, and the extent of aberration is not dependent on the subject-camera distance. The basic method of correction is to apply an inverse scaling transformation of image pixels as a function of their color.
The second type is very difficult if not impossible to correct adequately in post processing, primarily for the same reason that it is not generally possible to correct for missed focus. Deconvolution or various other sharpening algorithms may work partially, but these generally create unpleasant visual artifacts. The essential information has been lost: you can't de-blur what is blurry without more information about how the light entered the lens, and this information is dependent on the distance of the subject to the camera, as well as numerous other factors. Efforts to correct longitudinal chromatic aberration have largely focused on attempting to process out the color fringing as they register in the image file, rather than trying to model the behavior of the lens. Sometimes this works convincingly well. But it isn't really a substitute for designing the lens to avoid such aberrations.
The takeaway is that chromatic aberration occurs whenever a lens is designed in a way that causes an image point to be spread out as a function of the wavelength of light. If you've ever seen the OLAF test images produced by LensRentals, this is exactly what it will show: A collimated white light source is projected through the lens and a spot is produced that is very rarely just a white smear (let alone a white dot). It's almost always a multicolored pattern, which shows the presence of complex chromatic aberrations. See here:
http://www.lensrentals.com/blog/2014/04/olafs-lens-art
Upvote
0
:-\
