paul wrote:
>There is this thing called ExpoDisc which sounds similar. On one of the
>forums I'm on there is currently a lively discussion about its relative
>merits:
>
>http://www.photocamel.com/index.php/topic,2888.0.html
>
>It's even dragged in ExpoDisc themselves to comment.
>
>Apparently it's not too tricky to create one - can't find the reference
>at the moment, but I think it just consisted of some semi-opaque
>material across a filter or lens cap.
>
An interesting thread. Sort of scattered and the issues not thought
through. The way I see it:
The original use was for commercial movie production. The need there for
absolutely consistent color balance for shots that may be seen cut
together but are actually taken at different times and places is very
high. Most still shooters don't need that level of accuracy, but I'm
sure there are projects where it is more that worth it.
Obviously lots of different materials have been tried, and various
experimenters have opinions about their favorites. There was apparently
a shoot out using various materials a while ago which was won by an
older style Pringles cap. What I don't know is what the criteria were.
Expodisc claims to hand tune each one with color analyser.
The posts about where to point the thing to do custom white balance were
pretty amusing. Even the ExpoDisc post didn't really state the reason of
their position with any clarity.
The answer is pretty straightforward. Consider two subjects each of a
single color, one pure white and one say, yellow. The white one is
illuminated with light of the same color yellow as the other subject and
the yellow one is lit with pure white light. A each will then appear
identical in color to an observer and WB settings taken from them will
be the same. Images taken with their respective WB settings will show
apparently identically colored subjects.
Now consider WB settings taken by pointing the camera towards the light
source. The white source will result in little correction and the yellow
subject will appear yellow in an image taken with that WB. The yellow
source will result in a WB setting with major compensation. An image
taken with that WB setting will show a white subject. An image taken
with that WB setting will show a white subject. That's why you measure
the incident light, not the subject.
Of course, WB taken from the subject will quite often be pretty good
because many common subjects have a mix of colors that balances out not
far from neutral. But it should not be much, if any, better that the
auto WB on a competent camera, since that is just what it has to do to.
It has no choice but to assume that the overall average of the of the
actual subject colors in the scene is neutral, since it has no other
information. I suspect that the better ones are more sophisticated than
that, limiting the proportion of the scene that can come from each
color's sensors to create the assumed neutral setting.
One exception is the Olympus E-1, which has an ambient light sensor on
the body which is used together with the light reflected from the
subject to determine auto WB. I don't know how well it works on average,
but it can throw things off in some situations.
Moose
==============================================
List usage info: http://www.zuikoholic.com
List nannies: olympusadmin@xxxxxxxxxx
==============================================
|