Doug Kerr
Well-known member
It is widely (but not universally) accepted that when using a camera equipped with a "white balance diffuser" to measure the chromaticity of the incident light (under the camera's "custom white balance" scheme), best practice is to place the "instrument" (i.e., the camera with the diffuser mounted) at the subject location.
(There are various opinions as to the best way to actually face the diffuser: toward the camera position for the shot, toward the principal light source, parallel to the "surface" of the subject, and so forth, but that is a minor detail.)
This of course makes sense. We are interested in having the instrument accept for measurement the same illumination that falls on the subject.
Nevertheless, we continue to hear reports of cases where, for whatever reason (generally convenience, or disbelief in the "common wisdom"), the measurement was taken from the shooting position, with the diffuser "facing the subject", with quite satisfactory white balance results.
This at first seems paradoxical. After all, the light reflected from the subject has a chromaticity that is greatly influenced by the "reflective color" of the subject. That is, of course, why a red sweater appears red in the image.
Thus, it might seem, a measurement from the camera position toward a subject wearing red would indicate a "reddish" chromaticity for the incident light. The camera would adopt a corresponding "color correction vector" and apply it to the actual shot, resulting in the subject's face having an unnatural bluish cast. Yet we don't seem to see this happening.
The explanation is that the "acceptance sensitivity pattern" of the typical diffuser is not a narrow one. It may well embrace almost the entire hemisphere. Of course, normally the sensitivity will decline with angle off axis, perhaps even somewhat following the classical "cosine" response of an "ideal" (Lambertian) transmissive diffuser. Nevertheless, except in the case of a subject that is "really wide" (the classical "red barn shot from close up" example), the light reflected from the subject does not make up a substantial fraction of the light "gathered" by the diffuser, and thus doesn't seriously influence its chromaticity.
Such a pattern, incidentally, would be desirable for our purposes, since (even with multiple light sources of disparate chromaticity, the dreaded "mixed light" situation) the light presented by the diffuser to the camera proper for measurement would exhibit the same chromaticity as the "illuminance" (in the full technical sense) on a subject surface at that location, oriented the same way as the diffuser.
In fact, when we make the measurement from the shooting position, we are really measuring the overall ambient light on the photographer, just as when measuring at the subject location we are measuring the overall ambient light on the subject. And in most cases, the color of the subject has no more effect on the chromaticity of the ambient light falling on the photographer than the reflective color of the photographer's shirt has on the chromaticity of the ambient light falling on the subject.
Accordingly, if the ambient light is roughly "global" (that is, has the same chromaticity as it lands any place in the battle zone, on a surface of any reasonable orientation), measurement of the chromaticity of the light falling on the photographer will give us a good estimate of the chromaticity of the light falling on the subject (which is what the camera needs to do its white balance work).
Does this dispose of the desirability of measuring at the subject location? No. In many cases, the ambient light is not "global", and a measurement of the chromaticity of the illumination on the photographer is not a good indication of the chromaticity of the illumination on the subject. But at least we see why measurement from the camera position works as well as it seems to in many cases.
Several curiosities in this area appear in connection with a new white balance diffuser, the Color Parrot, recently introduced by Drew Strickland (proprietor of the Pro Photo Home forums). It is intended, he says, to be used for measurement from the camera position for the shot. He claims that, used this way, it gives better results overall than competitive diffusers (maybe used in that same way). (These other diffusers are mostly said by their manufacturers to be intended for use at the subject location, as discussed above.)
Drew largely attributes this superior performance of the Color Parrot to the fact that it is "more targeted" than the other devices. He hasn't defined exactly what that means technically, but it seems as if he is saying that the Color Parrot has a narrower acceptance sensitivity pattern, causing it to concentrate more on the light arriving from the subject. Of course, for me, that doesn't seem desirable, given the concepts discussed above.
I have not yet seen any sensitivity pattern for the Color Parrot. However, I have seen the "white balance reference frame" taken by the camera with the Color Parrot in place, aimed at a model with a red sweater in front of a red studio background. The chromaticity of that reference frame was very nearly neutral over almost the entire frame.
Thus, I cannot help but believe that the "instrument" was essentially capturing the overall incident illumination on the camera position (just as with other diffusers), not substantially "targeting" the subject. Otherwise, we would expect the reference frame to have a substantially "reddish" cast overall.
I don't have any details of the Color Parrot's construction. But I have seen a picture of its various "layers" laid out. I don't see any obvious evidence of any optical component that would in fact cause a sensitivity pattern that varied substantially from the "cosine" pattern that would be exhibited by a classical diffuser.
The Color Parrot does seem to produce a non-uniform pattern of luminance across the reference frame, but that is of course unrelated to the acceptance sensitivity pattern.
But of course the supposed "targeted" directivity of the Color Parrot may well result from some optical principle that has not yet been disclosed and whose implementation is not obvious from the picture.
I'm sure we'll learn more about this as the product comes into use.
Well, just some food for thought.
(There are various opinions as to the best way to actually face the diffuser: toward the camera position for the shot, toward the principal light source, parallel to the "surface" of the subject, and so forth, but that is a minor detail.)
This of course makes sense. We are interested in having the instrument accept for measurement the same illumination that falls on the subject.
Nevertheless, we continue to hear reports of cases where, for whatever reason (generally convenience, or disbelief in the "common wisdom"), the measurement was taken from the shooting position, with the diffuser "facing the subject", with quite satisfactory white balance results.
This at first seems paradoxical. After all, the light reflected from the subject has a chromaticity that is greatly influenced by the "reflective color" of the subject. That is, of course, why a red sweater appears red in the image.
Thus, it might seem, a measurement from the camera position toward a subject wearing red would indicate a "reddish" chromaticity for the incident light. The camera would adopt a corresponding "color correction vector" and apply it to the actual shot, resulting in the subject's face having an unnatural bluish cast. Yet we don't seem to see this happening.
The explanation is that the "acceptance sensitivity pattern" of the typical diffuser is not a narrow one. It may well embrace almost the entire hemisphere. Of course, normally the sensitivity will decline with angle off axis, perhaps even somewhat following the classical "cosine" response of an "ideal" (Lambertian) transmissive diffuser. Nevertheless, except in the case of a subject that is "really wide" (the classical "red barn shot from close up" example), the light reflected from the subject does not make up a substantial fraction of the light "gathered" by the diffuser, and thus doesn't seriously influence its chromaticity.
Such a pattern, incidentally, would be desirable for our purposes, since (even with multiple light sources of disparate chromaticity, the dreaded "mixed light" situation) the light presented by the diffuser to the camera proper for measurement would exhibit the same chromaticity as the "illuminance" (in the full technical sense) on a subject surface at that location, oriented the same way as the diffuser.
In fact, when we make the measurement from the shooting position, we are really measuring the overall ambient light on the photographer, just as when measuring at the subject location we are measuring the overall ambient light on the subject. And in most cases, the color of the subject has no more effect on the chromaticity of the ambient light falling on the photographer than the reflective color of the photographer's shirt has on the chromaticity of the ambient light falling on the subject.
Accordingly, if the ambient light is roughly "global" (that is, has the same chromaticity as it lands any place in the battle zone, on a surface of any reasonable orientation), measurement of the chromaticity of the light falling on the photographer will give us a good estimate of the chromaticity of the light falling on the subject (which is what the camera needs to do its white balance work).
Does this dispose of the desirability of measuring at the subject location? No. In many cases, the ambient light is not "global", and a measurement of the chromaticity of the illumination on the photographer is not a good indication of the chromaticity of the illumination on the subject. But at least we see why measurement from the camera position works as well as it seems to in many cases.
Several curiosities in this area appear in connection with a new white balance diffuser, the Color Parrot, recently introduced by Drew Strickland (proprietor of the Pro Photo Home forums). It is intended, he says, to be used for measurement from the camera position for the shot. He claims that, used this way, it gives better results overall than competitive diffusers (maybe used in that same way). (These other diffusers are mostly said by their manufacturers to be intended for use at the subject location, as discussed above.)
Drew largely attributes this superior performance of the Color Parrot to the fact that it is "more targeted" than the other devices. He hasn't defined exactly what that means technically, but it seems as if he is saying that the Color Parrot has a narrower acceptance sensitivity pattern, causing it to concentrate more on the light arriving from the subject. Of course, for me, that doesn't seem desirable, given the concepts discussed above.
I have not yet seen any sensitivity pattern for the Color Parrot. However, I have seen the "white balance reference frame" taken by the camera with the Color Parrot in place, aimed at a model with a red sweater in front of a red studio background. The chromaticity of that reference frame was very nearly neutral over almost the entire frame.
Thus, I cannot help but believe that the "instrument" was essentially capturing the overall incident illumination on the camera position (just as with other diffusers), not substantially "targeting" the subject. Otherwise, we would expect the reference frame to have a substantially "reddish" cast overall.
I don't have any details of the Color Parrot's construction. But I have seen a picture of its various "layers" laid out. I don't see any obvious evidence of any optical component that would in fact cause a sensitivity pattern that varied substantially from the "cosine" pattern that would be exhibited by a classical diffuser.
The Color Parrot does seem to produce a non-uniform pattern of luminance across the reference frame, but that is of course unrelated to the acceptance sensitivity pattern.
But of course the supposed "targeted" directivity of the Color Parrot may well result from some optical principle that has not yet been disclosed and whose implementation is not obvious from the picture.
I'm sure we'll learn more about this as the product comes into use.
Well, just some food for thought.