Interesting, my Sony CCD-3300 came with a cap that has a white center for color balancing. I immediately took it off and used another cap.

I think you are suppose to point the camera at the light source and then white balance.

A diffuse ""filter" (a white diffuser?) captures all the ambient light (Lambertian sensitivity) Pointing to the lightsource is OK if e.g. there is not a red wall close to you...or near the subject.

Put the camera in the subject position, turn it around to point at the camera shooting position. The light on your coffee filter is now these same as that on the subject. In an auditorium, set it all up beforefand, and do a trial if poss. Stage lighting is notoriously fickle, and of course can vary from scene to scene.
If you are out of doors, just turn the camera around on the tripod, and ensure it is in the same light (or shade) as the subject.
Works fine.
N

For those of you still looking for the original post in Fred Miranda's forums, here's the link:
http://www.fredmiranda.com/forum/topic2/92389

I'm still a bit 'uneasy' about WB with incident light as opposed to reflected light. I thought the WB with incident light is OK for photo cameras, but not for camcorders. Perhaps someone could comment on this, thanks.

There is no difference between incident and reflected measurements as long as the reflected spectrum is the same as the one transmitted through the diffusion filter. In both cases the cam gets "white" light on the CCD's.

I was thinking about holding the filters together with a needle point hoop. Then you can just hold it in front of the lens. Will that work??? There are small hoops that won't take up much space. I am going to town today and will try and find one. WBing here in Alaska is easy in the winter but soon we loose our natural cards from heat ;) I'll be bak.....

(Link about hoop and more below)
http://www.dvinfo.net/conf/showthread.php?t=63738

For accurate colour rendition the white balance needs to be set to correct for any colour cast in the light FALLING ON the subject, not the light reflected back from it. Consider white light falling on a red vase. The incident light is white while the reflected light is red. White balancing on the reflected light will attempt to reduce the red and return it back to white again. But this is exactly the opposite of what we're really trying to capture by photographing the vase in the first place - it's a RED vase and we want to see it as red, not as white or gray. That's why the best way to white balance with reflected light is to balance on a neutral gray or white card placed at the subject position. It's known to be colourless so the balancing process tunes out any colour in it. To use incident light for balancing, the camera or light meter with a diffusor has to be moved to the subject position with the light source's falling on it and and aimed back at the camera position, then moved back into position for the shot after balancing. Reflected balancing on a white card is usually more practical.

White balancing on a white card at the subject position is of course the best way to go, but got to the subject position is not always possible...Reflected light measurement is not such a straightforward story as the red vase example. A red vase will not change the WB which will be kept to an average value..Saturated colors are skipped in the calculation.

While the white balance algorithm may very well skip saturated colours, it's still based on the notion that an "average" scene will contain an equal balance of colours that when averaged will equal white or gray. It's this assumption that presents the problem with balancing on reflected light using a diffusor from the camera position because it is often simply not true and in many scenes colours from one side of the spectum or the other will dominate. If it's a choice between using a diffusor from the camera position or just guessing, then by all means use the diffusor. But IMHO it should be a method of last resort.

Reflected scene light WB techniques are not at all based on "avarage scene light". The analysis is much more complex. The basic systems search for very low saturation areas in the picture and use these values (and only these) for correction. These are also the areas where we "see" if the white balance is off, not on a saturated red vase...If there are no desaturated areas the WB is locked to an average color temperatures. More sophisticated approaches also analyse the IR content in the scene, some analyse flicker properties. Out of all this information, the WB is calculated