What the heck are chromatic aberrations?

fishfish Registered Users Posts: 2,950 Major grins
edited May 9, 2005 in Technique
Many moons ago our favorite whipping boy, wxwax, complained about one of his P&S digicams having the dreaded purple fringing. At the time, I didn't own that same model of camera and didn't really think much of it, as I figured whatever chromatic aberrations (CA) were, they must be an artifact of the specific camera he was using.

As I continue along my journey of photographic enlightenment, and acquire more sophisticated equipment, I find that I'm more curious about the issues that digital cameras have over film cameras -- and there are quite a few. One of which, is CA. Granted, I didn't really worry about it, until a certain well-known and widely hated (:lol3) photog from NYC who shall remain nameless (andy) mentioned that one of my beloved images exhibited signs of CA. EEK! How can this be? Not only is the dSLR nothing like the camera that the waxinator was using, but for gawdsakes...it probably cost five times as much, including lens. So instead of coming off like I was completely misinformed, I simply responded to said NY photog, "yeah, I know...oh well." :uhoh

And so in order to come clean, I did some major research on CA and found a number of very interesting, but highly technical, websites that explain the phenomenon. Now I can say in complete confidence, that I not only know what CA is, but can also recognize it, and maybe even explain it to someone.

I've included some excellent resources at the bottom of this post.

Let's start with a simple definition of chromatic aberration: An optical aberration caused by a lens bringing different colors of light to a focus at different points.

Further, Chromatic aberration is caused by the dispersion of the lens material, the variation of its refractive index n with the wavelength of light.

Since the focal length f of a lens is dependent on n, it follows that different wavelengths of light will be focused to different positions. Chromatic aberration of a lens is seen as fringes of color around the image, because all colors cannot be focused at a single common position on the optical axis. However, there exists a point, called circle of least confusion, where this effect can be minimized.

15408809-L.gif

Here's a classic example of CA (not my photo):
15410087-L.jpg


What makes it worse?

* The effect is worse at wider angle.

* The effect is worse at lower f numbers. (wider opening for the lens). You need to examine the particular shot.

* The effect is worse near the edges of the field. You need a bright object on a dark background to see it.

* The effect is only noticable when you have a sharp lighting contrast; a white object on a black background, for example or objects that contain opposite colors that are involved.. Abberation does not involve ALL of the light in question, just some of it. This means that you usually need a dark background to see the effect. On two brighter colored objects, a little of the colors involved will be mixed, but the tendencey is just to blur the image and your eye can't see the actual colors. This is the most important condition for seeing the effect.

* The colors of objects can make a difference. Objects that don't have the colors in them that are subject to aberration don't show the effect. Likewise, you can't see the violet fringe against a violet object.

* The color of the light can make a difference. Some lights that have pure spectral colors (a laser pointer or mercury vapor light are examples) will not show the effect because they don't contain the violet and green colors that are affected. Or, if they do, then the entire image is shifted, and you can't notice it.




For further research into CA, I recommend the following resources:

http://www.olympusmicro.com/primer/java/aberrations/chromatic/
http://www.vanwalree.com/optics/chromatic.html
http://members.shaw.ca/quadibloc/science/opt0505.htm
http://www.kineticbooks.com/physics/17372/17403/sp.html




as you were. :1drink
"Consulting the rules of composition before taking a photograph, is like consulting the laws of gravity before going for a walk." - Edward Weston
"The Edge... there is no honest way to explain it because the only people who really know where it is are the ones who have gone over."-Hunter S.Thompson

Comments

  • gregneilgregneil Registered Users Posts: 255 Major grins
    edited February 6, 2005
    Brilliant! Thanks for the not-too-technical lesson. It all makes sense now. Any ideas as to why it would be a problem associated with digital cameras and not film camera?
    There's a thin line between genius and stupid.
  • fishfish Registered Users Posts: 2,950 Major grins
    edited February 6, 2005
    I think it's because of the large density of pixels in a CCD or CMOS sensor, where the light propagates to nearby pixels. I'm not certain about this...need to do some research. Film doesn't have this problem, because it doesn't have discrete pixels. I would assume that CA is possible on film, just not that readily apparent.
    "Consulting the rules of composition before taking a photograph, is like consulting the laws of gravity before going for a walk." - Edward Weston
    "The Edge... there is no honest way to explain it because the only people who really know where it is are the ones who have gone over."-Hunter S.Thompson
  • dkappdkapp Registered Users Posts: 985 Major grins
    edited February 6, 2005
    Good post. I could never tell the difference between CA & PF (Purple Fringing) so I just chalk them up as the same. Am I right, or just happy in my own ignorance?

    I came across some great tutorials awhile back when researching the f828. There seemed to be some discussion going on when it was released about this. I think I saved off some of the links & tutorials teaching you how to remove the purple & save the pic. I'll see if I can find them & add them here.

    Dave
  • gregneilgregneil Registered Users Posts: 255 Major grins
    edited February 6, 2005
    fish wrote:
    I think it's because of the large density of pixels in a CCD or CMOS sensor, where the light propagates to nearby pixels. I'm not certain about this...need to do some research. Film doesn't have this problem, because it doesn't have discrete pixels. I would assume that CA is possible on film, just not that readily apparent.

    That makes perfect sense to me. Kind of like the difference between analog and digital audio recording... With an analog image you're capturing everything the lens can see, whereas a digital image takes sample points. The CA is probably more aparent in digital because when the aberration hits a sensor, that sample will record the aberrant color (purple) for the entire region, wherease with film, it all blends together nicely. I'm definitely talking out of my rear end here, though. I wonder if once we get digital cameras with enough discreet sensors, we'll see less CA?
    There's a thin line between genius and stupid.
  • digismiledigismile Registered Users Posts: 955 Major grins
    edited February 6, 2005
    I don't believe the issue here is Digital vs. Film at all. CA is a result of the refractive properties of a glass lens or lens system.

    Check out Fish's first link to Olympus Microscopy. What the example shows is that an incoming beam of white light (all wavelengths) is refracted into its component parts instead of being focused into a single point and therefore displaying the original white light.

    I remember studying refraction of lenses back when the earth's crust was cooling ...:D CA has literally been around for a couple of centuries.

    So does a digital sensor show more CA than film? Not sure, because the whole point of CA is that the refracted light is not hitting the exact same spot. Possibly a CMOS might exaggerate the CA of a given lens.headscratch.gif

    Maybe there is a Phd Doctor in the house that can set us all straight :D

    Brad
  • wxwaxwxwax Registered Users Posts: 15,471 Major grins
    edited February 6, 2005
    Good thread, fishmonger. nod.gif

    Not a technical response.... but here goes anyway.

    Chromatic Abberration is a function of the lens, not the recording medium.

    However, CA is more obvious on digital cameras than film cameras. Why?

    Both media are composed of tiny receptors that are exposed to light. But film has vastly more receptors than a digital sensor.

    In all high contrast shots, CA is present. The lens makes sure of that. But thanks to film's vast numbers of individual grains, it is able to faithfully translate the minute band of blue light and keep its size in perspective relative to the entire image.

    A digital sensor has relatively fewer receptors. So splits in the light become more exaggerated: a single blue pixel accounts for a greater percentage of a digital image, than does a single blue grain on a film image.

    How'd I do? lol3.gif
    Sid.
    Catapultam habeo. Nisi pecuniam omnem mihi dabis, ad caput tuum saxum immane mittam
    http://www.mcneel.com/users/jb/foghorn/ill_shut_up.au
  • DavidTODavidTO Registered Users, Retired Mod Posts: 19,160 Major grins
    edited February 6, 2005
    wxwax wrote:
    Chromatic Abberration is a function of the lens, not the recording medium.

    Imaging Resource evaluates CA on every camera that they review--where appropriate. That means that they don't on SLR cameras, since it IS a function of the lens. But if the camera has a built-in lens, he'll evaluate the CA. Great reviews on that site.
    Moderator Emeritus
    Dgrin FAQ | Me | Workshops
  • gregneilgregneil Registered Users Posts: 255 Major grins
    edited February 6, 2005
    In all high contrast shots, CA is present. The lens makes sure of that. But thanks to film's vast numbers of individual grains, it is able to faithfully translate the minute band of blue light and keep its size in perspective relative to the entire image.

    A digital sensor has relatively fewer receptors. So splits in the light become more exaggerated: a single blue pixel accounts for a greater percentage of a digital image, than does a single blue grain on a film image.

    Yes.

    And even if some scientist type proves you wrong, I think I'm sticking with your explanation.
    There's a thin line between genius and stupid.
  • erich6erich6 Registered Users Posts: 1,638 Major grins
    edited May 8, 2005
    All,

    I have a Ph.D. in Imaging Science so I can speak to what's happening with chromatic aberration in film vs. digital. Now keep in mind that while I have this degree I'm a complete novice when it comes to real photography (go figure). So I'm the epitome of book-learning vs. practical know-how.

    Anyway, as some of you already stated chromatic aberration is completely a function of the optical system in front of the sensor. As Fish explained, it is due to different wavelengths of light coming into focus at different planes. Therefore, the amount of aberration is independent of how the light is being recorded.

    Now, digital sensors have pixels and as, Vxmax correctly stated, these are generally larger in size than the film elements that record light. What happens with digital cameras is that we get aliasing (errors due to not having enough sampling resolution). What does this have anything to do with color fringing? Well, an edge imaged unto a CMOS/CCD sensor with a perfect lens will still have a different response at shorter wavelenths than at longer wavelengths (shorter wavelenths are harder to resolve). That means that while the light coming from the edge making up red and green colors gets sampled just fine the blue portion is undersampled. That's why the fringing is purplish-blue. This is NOT chromatic aberration.

    What does chromatic aberration do with a CMOS/CCD sensor? Well, it really is a crab shoot (that's a technical term) because it can actually help you or hurt you depending on which wavelength is out of focus. If the shorter wavelengths are out of focus then you actually should get better color representation because all wavelengths are getting "blurred" at about the same rate (assuming the color pixels in the sensor are all the same pitch). However, if the shorter wavelengths are in focus then the other colors may still be resolved but now you're really susceptible to how well your sensor can sample the short wavelengths.

    Erich
  • colourboxcolourbox Registered Users Posts: 2,095 Major grins
    edited May 8, 2005
    fish wrote:
    I would assume that CA is possible on film, just not that readily apparent.

    Oh, it's possible. When I was a student I bought a cheap zoom (all i could afford) for my Pentax K1000 film camera, and it certainly produced some pretty ugly chromatic aberration. That was probably the first piece of gear that taught me that you get what you pay for with lenses.

    When color fringing turns up in digital photos, I try the chromatic aberration compensation control in Adobe Camera Raw that cancels out the CA. I think it does it by attempting to scale each color channel slightly to get them to line up again.
  • luke_churchluke_church Registered Users Posts: 507 Major grins
    edited May 8, 2005
    Hi,

    I think:

    Purple fringing and CA can also be mixed in with an effect called blooming, where over-triggered CCD cells 'leak' electrons into nearby cells resulting in an increase in brightiness over sharp contrast edges.

    This can increase the visibility of these defects by brighting them.

    I also think that there is some difficulty associated with the micro-lenses on digital cameras that tends to increase CA and similar effects.

    The combination of these makes it appear more visible.

    There is also the issue that people look at digital photo at a lot larger magnification. It's so easy to hit 'View pixels', much easier than getting your magnifying glass out :-)

    Hope this is of interest,

    Luke
  • ian408ian408 Administrators Posts: 21,939 moderator
    edited May 9, 2005
    Fish,

    Nice post. Marc Muench summed it up nicely when talking about PS.
    "Every lens has them".

    Oh, and we missed you at the shoot out.

    Ian
    Moderator Journeys/Sports/Big Picture :: Need some help with dgrin?
Sign In or Register to comment.