native color temperature in RAW mode

We all know that a film is formulated for a specific color temperature and that it is important to match the light sources to the film using filters when necessary. Underwater we use daylight films with daylight strobes and usually call it good enough.

The same is true of digital cameras. Sure we have white balance settings that alter the final output, but the imager itself has a fixed color temperature, just like film, and that data is available to us when we shoot in RAW mode.

The question is: What's the imager's native color temperature and why does it matter? It matters because ideally we want the white light that enters the lens to be the same color temperature. Doing so maximizes the performance of the imager because all three color channels are used to their greatest extent.

I think there is probably some variation between cameras (Foveon?) but it is likely that the temperature is close to the point at which red and blue are of equal strength. This is likely to be below the temp that our strobes provide, and means that our color balance starts out off and gets worse as the light travels through water.

In testing my D100, I found that an 81EF was slightly too strong to optimize the channels with the builtin strobe, but the difference would be made up by just a few feet of water. Of course, I need to use the real strobes, but at this point I intend to use a permanent 81EF filter for macro. Some might say "there goes a 2/3 stop of light" but that's in fact what we want. That 2/3 stop of light is blue and green that prevents our red sensors from being used to their fullest extent.

Any thoughts?

Just to add on, the DS125 strobe claims a color temperature of 4800K vs. 5700K for the DS50. The difference is the same as an 81C filter. The difference between the 81EF and 81C is about an 81A which most are familiar with.

This gives you an idea of the range of adjustment I'm talking about. The DS125 might be the ideal temperature for digital if it were not for the minimum of 3-4 feet of water. Maybe your camera is a little different.

Tell me if I don't understand. With my D100 I still set the WB I want when I shoot in RAW. When I open the RAW file in PS from NC5 the software lets me change the WB or keep it at the preset (usally Cloudy -3 in my case). So there is not a preset color temperature (WB) with the D100 in RAW mode. You set the WB you want. RAW allows you to change that WB after the fact.

I use Bibble rather than NC5. It does just as you describe plus it provides the option of disabling white balance or using one-click white balance. The RAW data is not white balanced, but the file includes the camera setting so the translator can do it, so what you said is exactly right. What I want to know is what the native color temperature of the camera is. This will be what the white balance is converting "from".

I just did a test w/ the Finepix RAW Converter EX and the embedded color temperature of a shot taken w/ the camera WB set to auto yields a temp of 6500Kelvin.

Cheers

James

Are you saying you feel the camera itself is 6500K or that auto mode yielded 6500K?

The camera itself....I think....ugh! It was set at Auto to take that shot - and the converter opens it at 6500K

Cheers

James

I decided to mess around with my 5050 and got some puzzling results.

I took pictures of a portable light table with a cold cathode lamp (measured 7600K, slightly green). First I took a picture in manual with the smallest possible aperture (f8.0) and and very fast shutter. Then I put it in auto and let it decide (chose f4.0). In both cases WB was set to cloudy. I also measured the internal strobe (6500K) and used it. I omitted the strobe results since they were basically identical to the light table ones.

I then loaded the two RAW images three times each into PS using the Olympus plugin. The first time (top) was with all processing disabled. The second (middle) is with WB and "color adjustment" and the third (bottom) is with "auto image process". These are the options Olympus provides. The result is below:

The right column is brighter as expected. The middle row is cold but neutral with a bias toward green as my color meter predicted. The bottom row shifts the bias slightly toward magenta and significantly brightens the image. The top row is obviously green but the red and blue components are virtually identical.

I had hoped to interpret the native color temperature from the top row, but you can see it is too far from neutral to make sense. I wonder what the explanation is for this. As it stands, I don't know what you're supposed to do with the top option. Could this be typical of Bayer pattern processing that Oly simply didn't factor out, or does this suggest an imbalance in sensitivity between R,G, and B? Perhaps it's just a plugin bug?

A poster on another forum (that shall remain nameless) brought this link to everyones' attention.

The reason I post it here is to note that the native color temperature of this CCD is 4500K.