Purple fringing - references to technical explanation?

Discussion in 'Digital Photography' started by Roland Karlsson, Nov 2, 2003.

  1. Hi,

    Purple fringing, a.k.a. chromatic aberration is a common
    problem for digital cameras. We have discussed it here
    many times. And the discussions usually diverge between
    a pure lens effect (chromatic aberration) and a sensor
    effect (non perpendicular rays hitting the sensor).

    Sometimes internal charge spreading (a.k.a. blooming)
    is discussed, but this is obviusly not true. Why
    should this effect spread from the center and outwards?

    Now, normal chromatic aberration does not look like
    purple fringing. It is usually looks like a geenish
    or redish fuzziness of edges, e.g. redish at an inner
    edge and greenish at an outer edge.

    Purple fringing chromatic aberration is non symetric,
    it is a broad and sharp purple "fringe" at the outer
    edge.

    Now ... someone must have made experiments with the
    real thing or have a 100% knowledge about this issue.
    Something reliable must be written somewhere.

    Do anyone know where to find the truth?
    Not speculations, not discussions but
    the actual truth, i.e. some document from
    a sensor manufacturer or a scientific report.


    /Roland
     
    Roland Karlsson, Nov 2, 2003
    #1
    1. Advertisements

  2. Charles Schuler, Nov 2, 2003
    #2
    1. Advertisements

  3. Thx for your reply.
    I have seen this - I don't doubt that it is correct.
    I just want an explanation why it is broad sharp purple
    areas instead of fuzzy red-green areas as chromatic
    aberration usuall is. I have heard several theories
    in this newsgroup why it is so. I even have my own theories.
    But, I am 100% sure that someone really knows. Instead of
    just speculating, I would like to know to.
    I certainly is not blooming IMHO. I can come up with
    no valid theory why blooming should cause several pixel
    wide purple areas on the side away from the center of
    the picture.
    You forgot the third alternative: an optical effect
    in the sensor and/or in the protection glass.


    Roland
     
    Roland Karlsson, Nov 2, 2003
    #3
  4. Roland Karlsson

    Nils Rostedt Guest

    Roland Karlsson
    This may be something one step above the usual webstuff. At least it gives
    some references.
    http://www.vanwalree.com/optics/chromatic.html .
     
    Nils Rostedt, Nov 2, 2003
    #4
  5. Roland Karlsson, Nov 2, 2003
    #5
  6. Roland Karlsson

    jriegle Guest

    I've owned lenses used with film SLRs that have purple fringing. It was
    slightly noticeable with a Nikor 300mm f/4 ED lens and a Tokina AT-X 400mm
    f/5.6 APO. I also noticed it stronger in a older Vivitar 200mm f/4 lens. The
    purple fringing would appear around bight (white) objects surrounded by dark
    areas. In the case of these lenses the fringing was more of a haze around
    the light area and occurred even in the center of the image. In this case it
    was caused by LCA (Longitudinal Chromatic Aberration) which means different
    wavelengths of light focus at different planes. Stopping the lens down will
    reduce the effect of LCA.

    TCA (Transverse Chromatic Aberration) means different wavelengths of light
    focus at different focal legths. This means that color fringing can be
    observed near the edges of the image becoming less towards the center and
    none at the center. For example, If red light is focussed at a slightly
    longer focal length on the focal plane than other wavelengths, you will see
    a red fringing on objects and that fringing will face the outer edge of the
    image. There will be a bluish fringe on the opposite side of the object. I
    find many tele-zooms and some tele-primes to exhibit this, but it is usually
    a red/blue fringe. Stopping the lens down will not correct the effect of TCA

    I imagine a lens could be "overcorrected" in such away that the red and blue
    wavelengths focus at a longer focal length giving us that dreaded purple
    fringing. In images that aren't so overexposed, you should see a yellowish
    color just to the inside of purple fringing. Since the fringing is most
    apparent from strongly exposed areas near the image edge, the yellow part is
    not always apparent, but I have seen it in my digital pix. I owned a Vivitar
    19mm f/3.8 lens that had a blue/yellow fringe near the edge (used with
    film).

    Now the delimma. If this is caused by TCA, it shouldn't get corrected upon
    stopping down, yet stopping down does reduce the effect! Optical engineers
    know that aberations are classified in orders. I can only guess that this is
    some higher order aberation (most "color fringing" aberations are lower
    order). Any optical engineers out there?
    John
     
    jriegle, Nov 3, 2003
    #6
  7. I think Olympus mentioned the angle problem when they came out with the
    Cxx40 lens. In their previous compact lenses, wide angle photos had the
    light coming in at such a steep angle that there would be about 10
    pixels of colored radial blur at the corners. Maybe you can find their
    old PR and lens diagram related to the fix.

    I've had the Oly 2000, 3030, and 4040. The purple fringing wasn't the
    typical lens chromatic aberration. There was a whole lot of purple
    (red+blue) but no sign of a green tint. A lens problem would show a
    balanced spectrum; green wherever red and blue were in the wrong place.
     
    Kevin McMurtrie, Nov 3, 2003
    #7
  8. Roland Karlsson

    Martin Brown Guest

    Simple lateral colour looks like this. But even when that is fully
    corrected (at two or three wavelengths) there is still some residual
    colour error.
    Have a look at the descriptions of chromatic aberrations of achromat and
    APO lenses in astronomical telescopes. They are much more sensitive to
    higher order chromatic aberrations because of the high magnifications
    used. And the purple fringing is very difficult to eliminate.

    Essentially one compromise made is that when the image is sharp and flat
    field for the mid-green it is slightly defocussed off axis somewhere in
    the red and the blue (and the error usually gets larger with radial
    distance). This shows up as a purple halo around bright objects and
    specular highlights. The simple form of chromatic aberration (lateral
    colour) is usually well enough corrected that you don't see it often in
    modern lenses. However you do see the second order effect.
    Not necessarily you can get a whole host of different effects depending
    on the exact combination of the error(s). It is usually a pretty good
    hint that if it gets worse as you move radially away from the optic axis
    it is due to the effects of aberrations in the lens.

    Regards,
     
    Martin Brown, Nov 3, 2003
    #8
  9. Roland Karlsson

    jam Guest

    Roland,

    After being away for a few months, I'm dismayed to see that there's
    still no concensus answer to this question. I used to think that
    blooming was involved, but you've cured me of that. The radial
    distribution and the observation of purple fringing in film images
    seem to point to some high order lens aberration rather than to a CCD
    effect or to run of the mill chromatic aberration, but alas, we still
    lack a definitive explanation.

    Let me complicate things further. Our friend "Chris from Germany" used
    to swear that a certain UV filter (a B+W, I think) substantially
    reduced purple fringing on his G2 and G3, and I trust his observation.
    (He's an astute photographer.) Skylight is loaded with UV, and purple
    fringing often appears around dark objects seen against skylight. That
    would suggest a role for UV, but purple fringing also appears around
    bright lights with presumably less UV output.

    If the UV lead is spurrious, could his filter somehow affect a high
    order lens aberration?
    --
    Jeremy McCreary
    Denver, CO
    www.cliffshade.com/dpfwiw/
    -------------------------------------------
    | Hi,
    |
    | Purple fringing, a.k.a. chromatic aberration is a common
    | problem for digital cameras. We have discussed it here
    | many times. And the discussions usually diverge between
    | a pure lens effect (chromatic aberration) and a sensor
    | effect (non perpendicular rays hitting the sensor).
    |
    | Sometimes internal charge spreading (a.k.a. blooming)
    | is discussed, but this is obviusly not true. Why
    | should this effect spread from the center and outwards?
    |
    | Now, normal chromatic aberration does not look like
    | purple fringing. It is usually looks like a geenish
    | or redish fuzziness of edges, e.g. redish at an inner
    | edge and greenish at an outer edge.
    |
    | Purple fringing chromatic aberration is non symetric,
    | it is a broad and sharp purple "fringe" at the outer
    | edge.
    |
    | Now ... someone must have made experiments with the
    | real thing or have a 100% knowledge about this issue.
    | Something reliable must be written somewhere.
    |
    | Do anyone know where to find the truth?
    | Not speculations, not discussions but
    | the actual truth, i.e. some document from
    | a sensor manufacturer or a scientific report.
    |
    |
    | /Roland
     
    jam, Nov 15, 2003
    #9
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.