capturing reds - success story

Discussion in 'General Technical Discussion' started by Iliah, Sep 11, 2005.

  1. Iliah

    Iliah

    Jan 29, 2005
    nowhere
    I was shooting red tomatoes and peppers today, just for fun. Daylight, Minolta colormeter suggests 5800K.

    First attempt: auto everything. Reds are blown out beyond recovery.

    Second attempt: auto WB, exposure compensation to get the red channel histogram just touching the right wall. Image underexposed, "correcting" exposure in post-processing brings way too much noise - it needed +1.3eV.

    Third and forth attempts: custom white balance with auto exposure and manual exposure compensation - same as above, red blown out or strong underexposure.

    Fifth attempt:
    - took custom white balance,
    - then shoot colour chart with it,
    - then shoot tomatoes, increasing exposure (against suggested by camera meter 1/20 all the way to 1/8!!!) and white balance on the camera so that blue channel just started to touch right wall on the histogram - I was to adjust colour temperature setting on the camera to 3800K,
    - opened colour chart shot, used click-white balance (minor change from preset), stored the values of white balance,
    - opened the tomatoes shot @ 3800K, applied the saved white balance - nothing is blown out, good colour and details in color channels,
    - took some shots using Auto WB, settings around daylight (direct sunlight, cloudy, +/- compensations), and manually setting 5600, applied white balance from colour chart shot - and was able to get the reds not blown out only by underexposing the image.
     
    Last edited by a moderator: Sep 11, 2005
  2. Gale

    Gale

    978
    Jan 26, 2005
    Viera Fl
    Thank You Iliah,

    Will have to read this a few times :>))
     
  3. kjoosten

    kjoosten

    79
    May 1, 2005
    Houston, TX
    Man, that's pretty labor intensive. I've had the same problem with rose shots (but gave up long before I would have figured this out). It would certainly seem that there is room for some camera intelligence to be developed.
     
  4. Iliah

    Iliah

    Jan 29, 2005
    nowhere
    Dear Kent,

    I think all what is needed - to lower colour temperature compared to the real one by 1000K - 1500K approximatly, include something gray in the test shot for future setting gray balance by clicking it, and watch the histogram for correct exposure. Seems pretty simple.
     
  5. Iliah

    Iliah

    Jan 29, 2005
    nowhere
    After some more experiments and :biggrin: thinking :biggrin: I get this: adjusting WB until all the channel histograms show the right end at nearly the same point; then adjust exposure until all of them start to touch the right end; then cover the reds with something gray for a shot from which WB will be copied to final shot. Next will be test of usefulness of blue filters for shooting reds.
     
  6. cwilt

    cwilt

    Apr 24, 2005
    Denver, CO
    If I read this correctly...

    When photographing reds in daylight. Skew the WB in camera and correct WB in post. Could that explain why I get better red rose shots with WB set to incandescent in camera and adjusted later?
     
  7. Iliah

    Iliah

    Jan 29, 2005
    nowhere
    Dear Charles,

    Setting to incandescent may be too low, but it explains things perfectly if you understand what WB is. For daylight red multiplier is bigger then blue, to compensate the absence of red part of spectrum. But all the details into the red are brought by "unwanted colour", cyan/blue. Intentional shift of white balance to lower values reverses the situation - blue is amplified more then red. This way we can expose blue channel properly without red being blown out.
     
  8. cwilt

    cwilt

    Apr 24, 2005
    Denver, CO
    Wouldn't I have to skew it more than someone at a lower altitude? My wifes rose garden is at 5700 feet. I stumbled across this with the D70 but thought it was due to IR/UV contamination.
     
  9. Iliah

    Iliah

    Jan 29, 2005
    nowhere
    Dear Charles,

    Right, with D70 you may need to compensate more then with D2X.
    D2X, 7000K (that is what you have at that altitude for daylight, if I'm not mistaken) :
    R/G = 1.93
    B/G = 1.4
    D70(s) / D50, 7000K:
    R/G = 2.65
    B/G = 1.36
     
  10. Iliah do you have pictures to show the progress of your actions?
     
  11. cwilt

    cwilt

    Apr 24, 2005
    Denver, CO
    Sounds about right. We are a little blue up here. Or is that from the lack of O2? :wink:
     
  12. TOF guy

    TOF guy

    208
    Mar 11, 2005
    Iliah,

    I don't get it :confused: ! You're shooting JPEGs now :frown: ?

    Thierry :redface:
     
  13. Iliah

    Iliah

    Jan 29, 2005
    nowhere
    Dear Thierry,

    Impossible insult :wink: . I even do not shoot compressed NEFs for such experiments. But don't you know that D2X uses analogue "preconditioning", that is amplification/multiplication?
     
  14. TOF guy

    TOF guy

    208
    Mar 11, 2005
    Oops, sorry :Little: :Angel: ! I should have guessed :rolleyes: ! This is for the D2X (and D2Hs, I assume ?) .

    This is one feature I hope will make it to the D200 :tongue:

    Now, you've mentioned this as a technique (adjusting WB so that each channel independently just barely stretches to the right of the histogram) to avoid overexposing any channel. Will this also minimize noise in each channel as well ?

    If I'm good my wife will let me buy the D200 :Angel: :Angel: :Angel:

    Thierry
     
  15. Iliah

    Iliah

    Jan 29, 2005
    nowhere
    Dear Gilles,

    Look at these two screen captures. Note that exposure is absolutely the same. See the spike at the right side of "correct wb" shot. No amount of manipulations I tried helped to get data out of it. All I was able to do was to move that spike to the left.

    [​IMG]

    [​IMG]

    Dear Thierry,

    Of course correct channels exposure helps to supress noise.
     
    Last edited by a moderator: Sep 13, 2005
  16. TOF guy

    TOF guy

    208
    Mar 11, 2005
    Iliah,

    Comparing the images in the top and bottom frames :Dizzy: :

    The speculars don't have the same color :confused: . So you still have to adjust WB on the second pic. Wouldn't this modification leads to red overflow :eek: ? A curve could be used to prevent that, but then one has to deal with preserving contrast, shadows, ... kind of the same issues that one face with too much DR.

    So it seems that you have more data, but also more work to do in PP :Depressed . Of course the final pic will be better :biggrin: .

    Or am I missing something (I don't have a D2X, therefore I'm only making hypothesis here, nothing borne from experience ) :Unsure: ?

    Thierry
     
  17. cwilt

    cwilt

    Apr 24, 2005
    Denver, CO
    Dear Iliah,

    What if you replaced the red channel in the 5600K image with the one from the 3800K image?

    Or...

    Shoot at the proper WB with proper exposure and allow the red to blow. Bracket for the red channel on the next shot. Replace the blown red with the properly exposed red.

    Thoughts?

    Kind regards,
    Charles
     
  18. Iliah

    Iliah

    Jan 29, 2005
    nowhere
    Dear Friends,

    Whatever you are shooting, at the time of shoot (if not tethered) you are limited to JPEGs - as what you see is the histogram from JPEG (thumbnail). To set the exposure correctly based on histogram the only way is to skew WB first. After the exposure is determined, you can restore the correct WB if you are going to shoot NEFs. With some converters and some cameras you will have the most active colour channel slightly blown out that way, but only very slightly. Decreasing exposure 1/3 of the stop solves that.

    Dear Charles,

    Combining 2 exposures may not be necessary if things are done as I suggested here. While combining exposures wind is our enemy.

    Dear Thierry,

    WB was restored by just loading WB from colour target shot.
     
  19. Why not just shoot in RAW ???

    Thanks
    JohnG
     
  20. cwilt

    cwilt

    Apr 24, 2005
    Denver, CO
    The wind must be rough there if can move those tomatoes. :wink:

    Lets see....

    Skew WB to match RGB Channels. Set proper exposure based on this histogram. Take shot. Include WB target of choice. Take another shot for WB reference. Open WB shot and set WB based on target. Save WB numbers. Open image and apply WB from reference.

    Is that correct?

    Wouldn't that underexpose the blue channel and result in a similar image to under exposing with proper WB to begin with?

    HELP! I'm caught in a loop! :confused:
     
Loading...
Similar Threads Forum Date
A workflow from Capture to Final Backups General Technical Discussion Mar 7, 2014
memory/battery swap changes capture mode!?!? General Technical Discussion Feb 23, 2012
D7000 Download Capture Raw to PS3? General Technical Discussion Feb 20, 2011
3D (3 dimension not Nikon D3!) Project - frame capture question General Technical Discussion Feb 5, 2010
Capture and those Red Rectangles! General Technical Discussion May 3, 2005