Just a few thoughts on some recent tests I had my studio lighting college class do. I had them shoot under 5 different lighting conditions; direct sun, open shade, tungsten, florescent and strobe. I had them keep the white balance on direct sun.

I had them correct the files to a neutral grey patch on a MacBeth color chart. When corrected, the files were not all the same. Several colors did read differently on the conversions.

I do not use the auto WB. It simply does not know what it is looking at. It does not know if there is a predominant color in the scene and can only correct to neutral.

The same goes for in camera metering. No matter how sophisticated, it can only meter to an overall neutral grey. It does not know if you are looking at a scene with a lot of dark tones or an overall white scene like a snow covered landscape. It will try to make it grey.

I think these basic issues are why we hear so many complaints about meters being "wrong". It is also why I still use an incident light meter and a color temp meter for my work.

My advice is to skip auto WB and make your "best guess" as to the proper WB setting.

Jim Cavanaugh
www.cavphoto.com

When we were shooting film, the only options we had to contol white balance was to use filters on the lens or correct in printing. Films are balanced, arguably, for between about 5000K and 5800K depending on manufacturer.

So why don't we just all run around with our digital cameras set at between 5000K and 5800K like we were used to with film, throw an occasional 81A or 81B filter on, or fix it later as needed? In bright sun without an 81a, I find myself very often liking the setting right around 5600-5800K. Why not just leave it set at that?

With the D1 series cameras, this is pretty much what many people did because Auto WB worked so poorly. We'd try to make small corrections when we thought it was brighter or the sun was lower, etc. I know from experience though that cloudy, cloudy -3, direct sun, or whatever setting I picked as what I thought was needed often was not - even though the lighting conditions seemed quite familiar. Images with significantly off WB also often exhibited over or under exposure even though routinely metering the same subjects the same way.

Auto WB on the D2X does work and it works incredibly well.
Better than I can guess (often enough that I'd rather not), and faster than I can preset using a card or ambient meter.

Again, the problem is not the auto WB capability of the camera. I also believe it helps the Matrix metering function better and more predictably since WB is tied into the metering. Jpegs, and nef's opened in capture look great. The same nef's opened in ACR, with some subjects though, are predictably mis-interpreted by the software and often way off!

I did a file search of my entire computer and eliminated all but the correct ACR .8bi file version 3.3 from everywhere except the correct location - including zipped downloaded files, slightly renamed files, etc. Hasn't made a bit of difference.

On the upside, though, I tried shooting some test shots with preset WB taken using the ambient meter. Results only slightly cool for my taste but darn close. It's not that hard to do, either, and you can store six (I think)different presets with comments. I expect I'll use this feature a lot more for some of my scenic work especially. Thanks for the pointer, Carol.

If you use the pre set WB from the shooting position is that not the same as you would get with Auto WB???? I though the preset using the function button was kind of like the exposure lock. You move to the light that is hitting your subject and preset then move back. If you are in the same light as your subject doesn't Auto WB set based on the prism anyway????



Foley

AWB uses a combination of measurements including the prism ambient meter and the light coming throught the lens. This combination of readings from three sensors generally results in very accurate WB.

The really cool thing that I learned from Carols post is that the ambient sensor on the pentaprism can be used much like a hand-held color temperature meter that inputs the correct WB setting for the light falling on the subject.
This is way more convenient, for me, than measuring reflected light using a gray card. Like a hand-held incident light meter, you do not have to be at the subject position to take an accurate reading - you only have to be sure that the light falling on the subject is the same as the light hitting the sensor. This means that won't likely get an accurate measurement if, for instance, your subject is in sun and you (and the camera) are in shade.

I wish the presets had +/- adjustability like the other settings so I could warm it up just a touch - but it works great and I can see myself using it in a variety of situations.

Obviously it must have sleep walked Dave