Results 1 to 13 of 13

Thread: Diffraction and Digital IR Photography

  1. #1

    Join Date
    Dec 2011
    Location
    WA
    Posts
    34
    Real Name
    Shane

    Diffraction and Digital IR Photography

    I am upgrading my D200 IR830nm converted DSLR to full frame to fully utilize the focal length of the 25mm/2.8 ZF-IR and 50/1.4 ZF-IR lenses, and as a bonus bigger prints. Unfortunately I have Nikon DNA so I am stuck with a FF Nikon, which currently limits the choice to the D700, D610 or D800 models.

    The D600 is excluded due to OLPF cleanliness issues.
    The D700 is also suspect due to the internal IR LED used for shutter monitoring/calibration (LED wavelength is 850nm and can create issues with long exposures).

    D200 10MP, 6.1 um pixel diam, 164 pixels/mm
    D700 12MP, 8.5 um pixel diam, 118 pixels/mm
    D610 24MP, 6.0 um pixel diam, 167 pixels/mm
    D800 36MP, 4.9 um pixel diam, 204 pixels/mm

    Since the relationship between the Airy disk diameter for 850/900nm IR light (rather than 550nm) and pixel size has a significant effect on f/ratio choice and subsequent DoF, it has raised some questions after reviewing several websites (including this one) discussing diffraction and f/ratio.

    1) When relating Airy disk diameter to pixel size, factors from 2x to 3x the pixel size are often used to account for AA blur and Bayer matrix effects. (Bart van der Wolf utilizes 1.5x to indicate the onset of diffraction effects). An IR converted DSLR has no AA filter but still utilizes a Bayer matrix.

    What would you recommend the best approach to account for this in the Airy disk relationship to pixel size?

    2) In the Cambridgeincolour diffraction tutorial I am somewhat confused by the presentation of the calculated Airy disk diameter and the visual presentation of the Airy disk superimposed on the pixel dimensions. The visual presentation shows the disk to be smaller than the calculated size AND a footnote states
    “Note: above airy disk will appear narrower than its specified diameter (since this is defined by where it reaches its first minimum instead of by the visible inner bright region).”

    This is where I would like some clarity. Isn’t the area bounded by the first minimum representative of the, “visible inner bright region” (it certainly incorporates it) or have you chosen to eliminate that fraction of the Airy disk that also falls below 20% of the relative field strength within the first minimum?

    The calculated Airy disk diameter shown in the calculator is based on the diameter of the first zero (minimum) of the diffraction pattern, which is
    1.21967 lambda/(2 R) or 2.43934*N*lambda
    This value represents the main “inner” core of the Airy disk and accounts for around 83.8% of the total field strength of the Airy disk. This would taper from bright in the center to black at the calculated diameter but still represents 84% of the total field strength.

    The visual Airy disk representation superimposed on the pixel size appears to be calculated using
    1 lambda/(2 R) or 2*N*lambda
    As such, this only represents relative field strength greater than 20% within the diameter of the first minimum, hence a smaller diameter. As such, isn’t this the brighter portion of the inner bright region?

    No problem with this, but
    WHY is it presented this way AND
    WHICH is the most important value when relating to the pixel diameter, the calculated first zero diameter or the visual representation only representing a portion of the field strength in the first zero diameter?
    Last edited by t6b9p; 8th October 2013 at 11:01 PM.

  2. #2
    xpatUSA's Avatar
    Join Date
    Feb 2012
    Location
    Texas
    Posts
    2,949
    Real Name
    Ted

    Re: Diffraction and Digital IR Photography

    If you're interested in an alternative view, you'll get one here:

    http://www.luminous-landscape.com/tu...solution.shtml

    I don't know WHY CiC's tutorial is the way it is, can't help with that.

    In many ways, using "the Airy disk" to explain resolution is an unfortunate method. Especially when calculated to five decimal places . Some texts on resolution, e.g. of lenses, don't even bother bother with the (approx) factor of 1.22. Other texts argue the need for twice that, or even more. A further stumbling block is that pixels are square or slightly rectangular and actual do not have a diameter. Then, most texts use a monochromatic point source of light at exactly 555nm in their examples and try to relate the resulting disk radius to a fixed pixel pitch. Taking IR as being from say roughly 750 to 1000nm (for silicon detectors) there's quite a range of possible disk diameters from a particular scene and almost zero probability of point sources of light. Also sharp edges have their blur (sinc) function to add to your viewing pleasure.

    All of which leads me to answer your second question by choosing the second part: "the visual representation only representing a portion of the field strength in the first zero diameter". You've Googled "Strehl ratio" of course?

    You'll probably know that scene details in the "Airy Zone" have very low contrast, 9% or less, so may not even be that worthy of intensive research.

    Like your good self, I do find the subject interesting but prefer these days to separate lens diffraction as a subject from sampling theory as applied to sensors. That is why I provided the link above. It's hard reading but, at least, it mentions different colors and gives a table or two.

    LumoLabs has an interesting slant to resolution which you might find interesting . .

    http://www.falklumo.com/lumolabs/art...ess/index.html

    .
    Last edited by xpatUSA; 9th October 2013 at 02:20 AM.

  3. #3

    Join Date
    Nov 2009
    Location
    Provence, France
    Posts
    906
    Real Name
    Remco

    Re: Diffraction and Digital IR Photography

    Crudely put, such differences in fudge factors between authors generally mean "We have no idea what is going on here"
    and this is not limited to diffraction or even optics . The 'no idea' is an exaggeration of course.

    But what we can say:
    The airy disk diameter (however you define it) increases linearly with the wavelength, so double the wavelength (λ), and you double
    the airy disk diameter. So, the limiting diaphragm will be 1-2 stops lower/larger than for visible light (taken as ~500 nm, IR ~ 1000).

    So the different representations and fudge factors are not important in this practical use (they disappear from the equation).

  4. #4
    xpatUSA's Avatar
    Join Date
    Feb 2012
    Location
    Texas
    Posts
    2,949
    Real Name
    Ted

    Re: Diffraction and Digital IR Photography

    There's some more stuff in this thread elsewhere:

    http://www.physicsforums.com/showthread.php?t=516322

    Something else that is rarely emphasized in articles is that the focal length used is the nominal (infinity) value. Any subject that is closer than infinity will have a bigger Airy disk or edge spread. No problem for IR landscapes, but for a 1:1 close-up shot the sizes double (because at 1:1 the distance is 1 + extension = F*(1 + m) = F*2).
    Last edited by xpatUSA; 9th October 2013 at 04:29 PM.

  5. #5

    Join Date
    Dec 2011
    Location
    WA
    Posts
    34
    Real Name
    Shane

    Re: Diffraction and Digital IR Photography

    Thanks for the references Ted. I had already seen the LuLa page and bookmarked for future thorough reading but had somehow forgotten about it, so thanks for reminding me. The Lumolabs was interesting too.

    Remco, I appreciate your response but it appears you thought I needed a general answer to the "basic" problem of going from visible to IR. I have been shooting IR and UV film/digital for over 30 years so "doubling the wavelength, doubles the Airy disk diameter" is basic "stuff". If you have advanced knowledge of diffraction I would (respectfully) appreciate if you could elaborate further with regards to my specific questions.

    Added - Thanks Ted for the additional info, we must have been posting around the same time as I just missed it.

  6. #6
    Administrator
    Join Date
    Apr 2008
    Location
    California, USA
    Posts
    1,473
    Real Name
    Sean

    Re: Diffraction and Digital IR Photography

    Thanks for the careful reading. Here are some clarifications that will hopefully make things clearer for you...

    Quote Originally Posted by t6b9p View Post
    2) In the Cambridgeincolour diffraction tutorial I am somewhat confused by the presentation of the calculated Airy disk diameter and the visual presentation of the Airy disk superimposed on the pixel dimensions. The visual presentation shows the disk to be smaller than the calculated size AND a footnote states
    “Note: above airy disk will appear narrower than its specified diameter (since this is defined by where it reaches its first minimum instead of by the visible inner bright region).”

    This is where I would like some clarity. Isn’t the area bounded by the first minimum representative of the, “visible inner bright region” (it certainly incorporates it) or have you chosen to eliminate that fraction of the Airy disk that also falls below 20% of the relative field strength within the first minimum?
    The visual example shows the airy disc precisely at the calculated size, not smaller. This is where perception vs numerics enter in though, as you say, and I think is the source of the confusion here. The problem with airy disk visual examples is with how the formula gets tonally mapped for display; the size of the disk is actually quite a bit larger than it might appear since the outer edges are very dim. I think the tendency is to just gauge the width based on the bright parts and not based on the location of the first dark ring. Yes, I agree that this can be misleading in appearance, but that's just the nature of how an airy disk appears and is defined.

    Quote Originally Posted by t6b9p View Post
    The calculated Airy disk diameter shown in the calculator is based on the diameter of the first zero (minimum) of the diffraction pattern, which is
    1.21967 lambda/(2 R) or 2.43934*N*lambda
    This value represents the main “inner” core of the Airy disk and accounts for around 83.8% of the total field strength of the Airy disk. This would taper from bright in the center to black at the calculated diameter but still represents 84% of the total field strength.

    The visual Airy disk representation superimposed on the pixel size appears to be calculated using
    1 lambda/(2 R) or 2*N*lambda
    As such, this only represents relative field strength greater than 20% within the diameter of the first minimum, hence a smaller diameter. As such, isn’t this the brighter portion of the inner bright region?
    The visual example is calculated correctly using 2.43934*N*lambda, not 2*N*lambda.

    Quote Originally Posted by t6b9p View Post
    No problem with this, but
    WHY is it presented this way AND
    WHICH is the most important value when relating to the pixel diameter, the calculated first zero diameter or the visual representation only representing a portion of the field strength in the first zero diameter?
    The first zero and the visual representation are the same, and this is typically the standard way of assessing the size of an airy disk. However, diffraction is not an all-or-nothing effect, and instead happens gradually and to varying degrees. No single number is therefore necessarily correct; each value instead represents a standardized way of comparing between devices and sensors. The same applies to whether X or Y is a more important value for the circle of confusion when considering depth of field; each value just represents a standardized definition of how blurred something can be before it becomes "unacceptably sharp." This is why I have also added a second diffraction tool at the bottom of that page to show the varying degrees of diffraction for a given sensor size and resolution.

  7. #7

    Join Date
    Dec 2011
    Location
    WA
    Posts
    34
    Real Name
    Shane

    Re: Diffraction and Digital IR Photography

    Sean, thanks for following up with a clear explanation to my questions. Since originally posting I have done quite a bit more reading and was starting to come to the conclusion about the difference between the calculated and visual presentations, so your timing is perfect in clarifying that for me.

    Over the years I have found your tutorials to be well presented and have helped me in certain areas of interest, keep up the good work.

  8. #8

    Join Date
    Dec 2008
    Location
    New Zealand
    Posts
    17,662
    Real Name
    Have a guess :)

    Re: Diffraction and Digital IR Photography

    Quote Originally Posted by t6b9p View Post
    Sean, thanks for following up with a clear explanation to my questions. Since originally posting I have done quite a bit more reading and was starting to come to the conclusion about the difference between the calculated and visual presentations, so your timing is perfect in clarifying that for me.

    Over the years I have found your tutorials to be well presented and have helped me in certain areas of interest, keep up the good work.
    He's definitely a man at the summit of his game!

  9. #9
    xpatUSA's Avatar
    Join Date
    Feb 2012
    Location
    Texas
    Posts
    2,949
    Real Name
    Ted

    Re: Diffraction and Digital IR Photography

    t6b9p,

    Do you have a name?

    I am developing a spread sheet and, out of interest, I plugged in your D800 example at f/8 with 830nm light at 40 lp/mm:

    Diffraction and Digital IR Photography

    All theoretical, of course. It calculates your sensor MTF as a function of normalized detail frequency. It calculates the perfect lens MTF as a similar function. It allows you to enter some lens imperfection. Then multiplies the MTFs to get a total MTF of the captured detail frequency (54% in the above example).

    You may find this of interest:

    Diffraction and Digital IR Photography

    It shows how a selection of 6 real lenses behave with different aperture settings. The horizontal axis S normalizes the frequency part in place of the more usual lp/mm. S represents the spatial frequency Vf at the image plane divided by the zero contrast frequency Vo. A value of 1 is equal to the zero contrast frequency for the aperture of interest.

    Lens #1 at f/2.8 would be for portrait shooters perhaps and #6 at f/11 for landscapes? The lenses are pre-digital age, so-called 'legacy' lenses.
    Last edited by xpatUSA; 13th October 2013 at 03:35 PM. Reason: added more blurb

  10. #10

    Join Date
    Dec 2011
    Location
    WA
    Posts
    34
    Real Name
    Shane

    Re: Diffraction and Digital IR Photography

    Ted

    I had seen this spread sheet posted elsewhere but not with D800 parameters.

    thanks
    Shane

  11. #11
    xpatUSA's Avatar
    Join Date
    Feb 2012
    Location
    Texas
    Posts
    2,949
    Real Name
    Ted

    Re: Diffraction and Digital IR Photography

    Quote Originally Posted by t6b9p View Post
    Ted

    I had seen this spread sheet posted elsewhere but not with D800 parameters.
    Hi Shane,

    Interesting, that. Do you remember where you saw it? Thing is, it's my design on my HD, never posted anywhere in it's current form. The only thing that made it "D800" was that I entered a pixel pitch of 4.9um.

  12. #12

    Join Date
    Dec 2011
    Location
    WA
    Posts
    34
    Real Name
    Shane

    Re: Diffraction and Digital IR Photography

    Ted
    both the spreadsheet and charts are familiar and you posted it I believe on Lula, I will see if I can find the thread.

    Heres the charts http://www.luminous-landscape.com/fo...?topic=76371.0

    and the spreadsheet http://www.luminous-landscape.com/fo...topic=76371.40
    Last edited by t6b9p; 12th October 2013 at 11:31 PM.

  13. #13
    xpatUSA's Avatar
    Join Date
    Feb 2012
    Location
    Texas
    Posts
    2,949
    Real Name
    Ted

    Re: Diffraction and Digital IR Photography

    Thanks Shane,

    I had forgotten about those (getting old, ya know). The spreadsheet there was an early version; I'm still working on it from time to time.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •