Nikon D300 HDR out of the camera !!

Discussion in 'Nikon' started by Sosumi, Mar 2, 2008.

  1. Sosumi

    Paul Furman Guest

    Odd that red & green are opposite on the color wheel but diagonal in a
    sensor and double the number of green pixels. <scratching head>
     
    Paul Furman, Mar 5, 2008
    #41
    1. Advertisements

  2. Sosumi

    Tony Polson Guest


    The difference between light and pigment?
     
    Tony Polson, Mar 5, 2008
    #42
    1. Advertisements

  3. Sosumi

    frederick Guest

    The human eye is about twice as sensitive to green shades than other
    colours. AFAIK, that's why they did it that way.
    I guess (totally unsubstantiated) that one of the reasons why later
    iterations of sensors seem in many ways to defy physics in terms of
    clearly observable improved imaging performance in the face of physics
    that say it can't be done, might be big improvements in spectral
    performance of dyes in bayer filters.
     
    frederick, Mar 5, 2008
    #43
  4. Sosumi

    Focus Guest

    The only logical conclusion I come to, is that you would argue with a
    statue...
     
    Focus, Mar 5, 2008
    #44
  5. [fullquote removed. Please learn to quote responsibly.]
    Don't get me started on "*your* lack of imagination" and "just
    because *you* cannot come to any other conclusion".
    A statue that gave logical arguments would be ahead of 99% of
    the world population.
    A statue that at least would argue back in a semi-coherent manner
    would at least be a novelty and not worse than many here.

    -Wolfgang
     
    Wolfgang Weisselberg, Mar 5, 2008
    #45
  6. ["Followup-To:" header set to rec.photo.digital.slr-systems.]
    Which ones may that be?
    Even the very best dyes cannot pass more than 100% of the light
    they shall pass and block more than 100% of the light they shall
    block.
    Agreed?

    -Wolfgang
     
    Wolfgang Weisselberg, Mar 5, 2008
    #46
  7. Sosumi

    John Sheehy Guest

    They're not exactly opposite. Cyan is opposite red, and magenta is
    opposite green.

    Part of the range between blue and red on a color wheel would be totally
    invisible, if it were fully saturated, as visible light does not use a
    full octave.

    There is no meaning to the order of colors in the red/green*2/blue
    elements in a CFA. The only consideration is that since there are more
    greens (probably because noise and detail in the green channel are more
    noticeable), that the greens be equally distibuted, and they can only be
    done so with such a pattern:

    g*g*g*g*
    *g*g*g*g
    g*g*g*g*
    *g*g*g*g

    After you do that, the relative placement of red and blue are irrelevant,
    because the pattern is the same, just shifted some, if you swap blue and
    red.

    --
     
    John Sheehy, Mar 5, 2008
    #47
  8. Sosumi

    frederick Guest

    "defying physics" was perhaps not the right way to put it. "Exceeding
    expectation" based on what was understood to be a given reality might be
    a better way of putting it. In the "1 megapixel" days of digital
    imaging (not so long ago), there were many expert sites with deep and
    apparently concise calculations "proving" that we'd need 25mp bayer
    sensor digital just to match everyday 35mm colour print film. We didn't.
    Yes - agreed. You don't think spectral performance of dyes in RGB
    arrays is perfect now, I assume?
    Nikon's patented dichroic mirror sensor design might be an alternative,
    if it worked or could even be made in a commercially viable manner.
    Foveon seems a doomed and severely flawed concept IMO.
     
    frederick, Mar 5, 2008
    #48
  9. Sosumi

    Colin_D Guest

    That says nothing about the shape of the passband associated with the
    filters. The shape of the response curve determines the effectiveness
    of a filter more than does efficiency. Efficiency in terms of Light loss
    can be made up in the amplifiers; passband shape cannot.

    Colin D.
     
    Colin_D, Mar 5, 2008
    #49
  10. Sosumi

    Paul Furman Guest

    Ah so the artist's color wheel is anthropomorphisized, not a scientific
    relationship?

    I'm a landscape architect who has spent a lot of time rendering greens
    onto drawings and I can tell you it's just amazing how many subtle
    greens people can distinguish. We always complained there weren't enough
    shades of green colored pencils. My most recent reminder was quickly
    choosing a green from a Pantone color chip then getting the print back I
    was astounded how wrong the choice was. We went back to the Pantone set
    & there simply wasn't anything close to what would have been
    appropriate. We did find something close but it took a lot of hunting.

    How about a color system that has more bits of green distinction than
    red or blue?
     
    Paul Furman, Mar 6, 2008
    #50
  11. No reason why it should, unless like Isaac Newton you're an occultist
    with strong prior reasons for believing in the magic of
    octaves. That's why he invented the rather subtle extra distinction at
    the blue end in order to make the number of colours in the spectrun
    the magic seven, when six would have been more reasonable.
    The artist's colour wheel is intended to represent the colour mixing
    and complementary relationships in the colours we see. Since we see
    colours by using three colour filters in our eyes, it appears to us
    that there are three primary colours. For example, mixing yellow and
    blue to produce green isn't a fact derived from the physics of the
    electromagnetic spectrum, it's an optical illusion derived from the
    kind of colur filtration ond processing our brains and eyes use to
    distinguish colours. So in that sense the colour wheel is
    anthropomorpshised.
    The subtlety of our green perception is probably a consequence of the
    fact that one of our eye colour filters is green dominated, and the
    old evolutionary bargain between plants and fruit eaters that the
    plants would tell us when the fruit was ripe by switching from a
    camouflaged green to a red which stood out from the green leaves.
     
    Chris Malcolm, Mar 6, 2008
    #51
  12. Ah, yes, the CMYx-Bayer patterns. They've fallen out of use,
    haven't they?

    -Wolfgang
     
    Wolfgang Weisselberg, Mar 6, 2008
    #52
  13. Sosumi

    Colin_D Guest

    CMYK filters are not the point. My remark applies to ANY filter
    designed to pass or block parts of the spectrum.

    Colin D.
     
    Colin_D, Mar 6, 2008
    #53
  14. Sosumi

    Martin Brown Guest

    In part because defects in the chroma subsampling algorithm leads to
    problems in a saturated red image with fine luminance detail.

    Actually human colour vision does cover more than a full octave. Violet
    Hg line at 365nm is pretty bright purple (though most folk cannot see
    the 334nm line) and at the other end the eye remains sensitive out to
    700nm or so. And for some out as far as 780nm but with much reduced
    sensitivity. The eye is actually sensitive to the light at these longer
    wavelengths, but the "red" signal is interpreted as the difference of
    "yellow" and "green" sensors so sensitivity is poor. See for example:

    http://www.rwc.uc.edu/koehler/biophys/6d.html

    It isn't helped by the fact that the human eye has sensors peaked in
    blue, green and yellow. Red is a creation of our brain interpreting the
    signals.
    It is only approximately true for normal human colour vision. People
    with various forms of colour blindness see things differently.
    That is quite commonly done (or rather was in the old days of limited
    palette images). The other thing we are exquisitely sensitive to are
    flesh tones. Wrong colour usually means something bad.
    The green channel carries about half the luminance signal. There are
    other masks that give higher transmission typically CMYG or Bayer using
    two different shades of green (ISTR Sony christened one Emerald)

    http://www.dpreview.com/news/0307/03071601sonyrgbeccd.asp

    I found by accident the following URL with a large but interesting PDF
    that covers a lot of the basics and some other digital processing tricks
    too.

    http://www.star.bme.hu/clit1/Pixelgraphics/Digital_Imaging.pdf

    Regards,
     
    Martin Brown, Mar 6, 2008
    #54
  15. Sosumi

    Dudley Hanks Guest

    I am in no way an expert on HDr, nor have I been reading all the messages in
    this thread. But, I'm wondering how we can refer to an enhanced image
    straight out of the box by this term.

    It is my understanding that HDR is what we get after we overlay a number of
    shots which have captured a portion of the dynamic range of a given scene.
    The idea is to end up with an image that contains a greater vibrancy than
    can be captured with current sensor and printing technology. With the
    Nikon, and similar systems by other manufacturers, doesn't the camera just
    kind of guess at what very dark shadows and very light highlights would look
    like if the sensor could actually get the job done?

    Hence, shouldn't we actually refer to images produced by these systems as,
    perhaps, EDR (Enhanced Dynamic Range)?

    Wondering,
    Dudley
     
    Dudley Hanks, Mar 6, 2008
    #55
  16. Sosumi

    Pete D Guest

    Thats what Pentax calls it.
     
    Pete D, Mar 6, 2008
    #56
    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.