Question for you optical gurus...

Discussion in 'Digital SLR' started by Mark², May 7, 2007.

  1. Mark²

    Mark² Guest

    I know people sometime use "star filters" to create similar effects...
    ....but what exactly is going on when these bright-light star effects are
    created with no such filter?
    What causes it?
    That is an un-edited image (save for a slight levels adjustment and size
    I'm assuming the number of points is dictated by the number of aperture
    blades(?)...but I don't know.
    I don't spend a lot of time number crunching, but I know there are folks
    here who know this stuff inside and out (Littlewood? Bill? Roger?).

    Just curious... And thanks in advance to anyone who can shed light on this
    one (pun intended).
    Mark², May 7, 2007
    1. Advertisements

  2. Mark²

    Paul Furman Guest

    Yep, you are stopped down to f/11. It won't do that wide open where the
    opening is perfectly round (or with rounded aperture blades?... maybe
    not perfectly round?). I'd guess it's an incremental thing, not positive.
    Paul Furman, May 7, 2007
    1. Advertisements

  3. achilleaslazarides, May 7, 2007
  4. Mark²

    TheDave© Guest

    I lost a photo contest once because the judge thought I "cheated" by
    using a star filter when I hadn't. Virtually identical type picture as
    your example, just different location.
    TheDave©, May 7, 2007
  5. Mark²

    Jeff R. Guest

    Jeff R., May 7, 2007
  6. Mark²

    Mark² Guest

    Mark², May 7, 2007
  7. Mark²

    Bill Funk Guest

    I've noticed the same thing, and since it changes with the aperture
    used (the points are stronger with a smaller aperture), I'll have to
    go with the idea that it's caused by the aperture blades and the
    difraction they cause.
    I never checked to see if rounded blades (for a more round aperture at
    smaller apertures) lessens the effect, but I would imagine so.


    Hillary Clinton refused to fly in a Gulfstream II
    private jet provided for her in South Carolina
    and demanded the more luxurious Gulfstream III.
    She was on her way to California. She can't be
    seen in something the valets are embarrassed
    to park.
    Bill Funk, May 7, 2007
  8. Mark²

    acl Guest

    It's diffraction. Diaphragms with an even number of blades result in a
    pattern with the same number of rays, while those with an odd number
    result in patterns with twice that number. For diaphragms with more
    blades, there are more rays but they get dimmer. It is not hard to
    imagine having so many that the end result is a circular aperture, and
    what the effect would be...
    acl, May 7, 2007
  9. I dont believe it has anything to do with the number of aperature
    blades... It depends on the way the light refracts on the glass of the
    lens. Assuming it is happening without a filter of any kind...
    Scratches on the lens will affect this as well...

    Ghosthunter1013, May 7, 2007
  10. Mark²

    acl Guest

    You are wrong, the number of rays is either equal to the number of
    blades (if even) or twice the number of blades (if odd). It has to do
    with the properties of two dimensional fourier transforms, see
    (go down to where it says fourier transform).

    As for "It depends on the way the light refracts on the glass of the
    lens", no, it doesn't in any meaningful way.

    Or, if you have two lenses with different numbers of blades, you could
    try it and actually check if it depends on it, no?
    acl, May 7, 2007
  11. ...what exactly is going on when these bright-light star effects are
    TTBOMK the stray light rays are caused by reflection from the
    edge of the diaphragm blade. The reflection can be seen by
    playing with a diaphragm and a small light in a dark room.

    There are, obviously, diffraction effects happening at the same
    time but diffraction will produce circles and bands around the
    image of the street lamp that become nested polygons as the
    lens is stopped down and deviates from circularity.

    In pin-hole photography the reflection from the sides of the
    hole causes havoc with contrast and flare. The best pinholes
    are made dimpling a sheet of brass with a pin-point and then
    carefully sanding away the convex side of the dimple - this
    results in a knife edge pinhole that will give good contrast.
    I find it is almost impossible to do.

    As the diaphragm is stopped down the hole becomes less circular
    and more like a pent/hex/sept/.../agon. The flare rays become
    more directional the more the opening deviates from a circle.

    With digital cameras there will be other directional artifacts
    resulting from the array structure of the sensor. I imagine was made with a
    digital camera.
    Nicholas O. Lindan, May 7, 2007
  12. No, diffraction will not produce circles around the image of the lamp.
    Please use google to find out what the effect of the shape of the
    aperture is on the image.

    As for it being reflection off the edge of the blade, think about it:
    if it was that, the length of these rays would be unequal if the lamp
    (or whatever) was not in the middle of the frame.
    Can you explain how these artifacts you have in mind would be formed?
    I have a hard time imagining this, and see nothing in the photo you
    posted to suggest any effect of the periodicity of the recording
    achilleaslazarides, May 7, 2007
  13. I lost a photo contest once because the judge thought I "cheated" by
    That's happened to me several times throughout my life....It really frosts
    my you know what, too. Stupid people jumping to erroneous conclusions that
    have a profound effect on your future.......
    William Graham, May 7, 2007
  14. Like this, maybe?

    Won't look near as pretty: the lamp not being a
    point source, and having a pretty messy spectrum,
    the thing will be a fuzzy dot. It won't be a set of

    The rest of this isn't worth responding to...
    Nicholas O. Lindan, May 7, 2007
  15. Mark²

    DoN. Nichols Guest

    It is in Australia, and you are probably in the US, based on
    your e-mail domain (

    Sometimes the DNS takes a while to find something, and sometimes
    it is even necessary to click on the URL (or the reload button) several
    times before you finally reach someplace. It is rather unpredictable.
    But I got to them with no problem -- from Vienna VA, USA, so try again.

    For that matter, to avoid the need for the DNS lookup, try
    substituting the IP address (which I have just looked up) in place of
    the server's domain name.

    IP Address:

    Modified URLs:

    Good Luck,
    DoN. Nichols, May 7, 2007
  16. Mark²

    acl Guest

    It's not worth responding to, eh? Did I somehow insult you, are you
    always like this, or did you just decide that I just don't know what I
    am talking about?

    Anyway, the link you posted to shows diffraction through a circular
    aperture. The diffraction pattern is related to the fourier transform
    of the aperture shape (rather, of the transfer function, but we take
    that to be 1 inside the aperture and 0 outside). The fourier transform
    of eg a square is going to have four "rays" coming out (of course
    they're not rays, they're modulated with a period related to the size
    of the aperture). It's trivial to check that the fourier transform of
    a shape with n-fold rotational symmetry will also be n-fold
    rotationally symmetric, but working out more details needs explicit

    Since I suppose you'll dismiss what I wrote again, look, for example,
    you can see some java applets to calculate the diffraction pattern
    through circular, rectangular and triangular apertures (go to where it
    says "rectangular and triangular apertures"). But then, if you don't
    believe what I say without bothering to investigate, why believe
    anything you find on the internet?

    Anyway. I can go into as much detail about this as you want, in fact.
    Somehow, however, I doubt you'll now google more carefully and say
    "oops you're right, sorry for the tone of the response" :)
    acl, May 7, 2007
  17. Mark²

    Paul Furman Guest

    Yet another glaring example here from yesterday's sailing shoot:
    Paul Furman, May 7, 2007
  18. Mark²

    Mark² Guest

    Wow. With so many widely-varying responses, it appears that I'm not the
    only one who didn't feel certain of exactly what's going on there... :)
    I actually find the effect quite appealing...expecially (out of
    stubbornness) given the fact that no special filter or effect was used or
    Mark², May 8, 2007
  19. Mark²

    Scott W Guest

    Do a 2D FFT of an octagon and you will see a pattern very much like
    what mark is seeing. This does not prove that it is diffraction but it
    is about what I would expect from an octagonal aperture.

    Scott W, May 8, 2007
  20. Mark²

    Jeff R. Guest

    Thanks Don.
    *I* am in Australia, as are the two photos, but my servers are stateside.

    I take it you viewed both pix OK?
    Jeff R., May 8, 2007
    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.