Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
RAW color multioliers vs color temperature
#1
Is it possible to compute color temperature given the three raw-photo color multipliers?

(The question comes from an effort to design another approach for underwater color balancing.)



Thanks,

Oleg.
#2
[quote name='olegk' timestamp='1347626225' post='20103']

Is it possible to compute color temperature given the three raw-photo color multipliers?

(The question comes from an effort to design another approach for underwater color balancing.)



Thanks,

Oleg.

[/quote]



well you need a white card or something similar to shoot against. Standard color meters have a white diffuser in front of them when you measure incident light.

If you don't have a white reference, how should the camera know whether you point it at some red, green, blue etc. and deduce the color temperature from it?

I guess a piece of white cloth you wrap around the lens dome could do it?

If a scene has decent representations of all colors, you can of course just look at the RGB histogram, and push the red and green channels, but that's more like guesswork.
#3
[quote name='photonius' timestamp='1347657772' post='20111']

well you need a white card or something similar to shoot against. Standard color meters have a white diffuser in front of them when you measure incident light.

If you don't have a white reference, how should the camera know whether you point it at some red, green, blue etc. and deduce the color temperature from it?

I guess a piece of white cloth you wrap around the lens dome could do it?

If a scene has decent representations of all colors, you can of course just look at the RGB histogram, and push the red and green channels, but that's more like guesswork.

[/quote]



Thanks, though what you describe is kind-of first step - obtaining the RAW image with suitable color balance written in it.



My question deals with a step somewhere later in the flow. I can read the three ( R, G, B ) color multipliers from the above mentioned RAW file. These numbers uniquely identify the "white balance" setting that would make any picture taken at the same depth around the same site look natural. Well, unless ambient light changes, but this doesn't happen too much. And now I want to apply this setting (to another image) in a RAW converter that "understands" only color temperature (Sony IDC). So I need the formula to convert color multipliers into the color temperature.



By the way, I tried a color-measuring method nearly identical to what you describe. To start from, I keep FLD filter on the lens when diving - to reduce extra green. I wrapped a piece of semi-transparent white polyethilene around the lens dome and tried to set custom white balance. Unfortunately it didn't work underwater. When I first did it on land, it worked as expected. It looks like the camera (Sony NEX C3) wanted to go beyound its color-temperature limit of 9900K. Maybe there was too much blue light at the site (Kessaria, Israel - depth of 3-9m), so that FLD filter wasn't appropriate? A reddish filter may do a better job, but I don't have it, so cannot try.

Interesting that the pictures were not a waste color-wise - I was able to restore acceptable white balance using both color temperature and curve adjustment.



Regards,

Oleg.
  


Forum Jump:


Users browsing this thread:
1 Guest(s)