Announcement

Collapse
No announcement yet.

Measuring and fixing the Magenta/Green Cast of my CFLs

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

    Measuring and fixing the Magenta/Green Cast of my CFLs

    I recently bought some Alzo CFL bulbs to test. I want to figure out how much of a hue is being introduced so that I know how to properly gel it/set my camera to get the correct colors.

    I set my white balance to 5500K, what Alzo claims their bulbs are, and took a picture of a white/black/grey card. I then import the JPEG image into LR (the jpeg has the baked camera settings in it, LR does funky stuff when importing RAW images). I eye-drop the grey card, some adjustments are made (increase to +31 magenta and +4 blue). So, I bumped up my color temp to 5700K and shifted the magenta in camera to +4. This all but removed the green shift entirely and was correctly white balanced.

    Is there any disadvantage to fixing the color in camera like this as opposed to using gels?

    Thanks!

    EDIT: I also noticed that if I eye-dropped different spots of the grey card, I get different readings (about a +5/-5 tint and temp difference). Is this because the exposure was not even across the entire surface and is it negligible?
    Last edited by SDub; 06-28-2016, 09:33 PM.

    #2
    Is there any disadvantage to fixing the color in camera like this as opposed to using gels?
    Well, the disadvantage would be if you have any other light sources present. If there's daylight coming in too, it'll now look "off". Or if you're mixing the Alzos with properly-balanced fluorescents or LEDs or tungsten or anything else.

    If you use gels, then you retain the ability to mix those lights with any other light source.

    In effect, what you've done in camera is a manual white balance, just -- well, you did it really, really manually. The process you did is pretty much what a camera's internal white balance procedure does.
    ..
    The AU-EVA1 Book - The DVX200 Book - The UX180 & UX90 Book - Lighting For Film & TV - Sound For Film & TV

    Comment


      #3
      The way I dial in my iffy daylight sources:

      Shoot a camera raw still with a gray card; open in Photoshop camera raw. Set WB with the eyedropper.

      Double check the RGB values of the gray card - they should be pretty even (like 155r, 155g, 155b) or within a few points.

      Take a look at the temp and tint sliders - these equate to cto/ctb and +green/-green. Take a guess or use an online color temp-to-CTB value converter. Gel the lights with your best guess and shoot again.

      You may find you need to further tweak your gels in 1/8 or 1/4 amounts, but after a couple shots, you should find what you need (IE, that particular fixture or globe needs 1/2 CTO and 1/8 minus green). Write the combo on the globe box or stick a piece of tape on the fixture.

      Now, if you need all your lights to match, you'll know the gel combos to get everything "pretty dang close".

      I do plenty of interviews where manually white balancing to me key still looks fine (particularly darker rooms). I have an HMI that's 6000k, and often I don't gel that if there's a window in the background - gives a really nice warm look to the background that's often appropriate. Sometimes window light is really much cooler and it looks fine. But sometimes, you need all your stuff to match.

      Comment


        #4
        Yeah, unless you're going for a particular "look" or effect, I would gel the lights, as opposed to shifting the whole camera, so that you can try to match the ambient light. Most of the time when I'm shooting with my Astra's, I'll check the WB of the room, if I'm working with the ambient/available light, and then dial the Astra's in as close as possible, then usually re-balance for them.

        Comment


          #5
          Ah, thanks Michael Carter for explaining your process. I mostly shoot at night (to avoid daylight influence) and use only these alzo bulbs so I guess just doing a manual white balance w/a grey card as opposed to setting the color temperature with a kelvin value is a better idea, like Barry implied.

          Thanks, guys!

          Comment

          Working...
          X