By Moshe Ronen
|
|
For the E-20 the information relating to WB is under ‘Vendor Original Information’, items: ‘WhiteBalance’, ‘Red WB bias’ and ‘Blue WB bias’. For the C4040 the information has a different format, it can be found under items ‘LightSource’ and the last of the three ‘Contrast’ items.
The following table and graph show the red and blue biases set by the E-20 for the various WB settings.
Without actually knowing the meaning of the biases, the numbers may be useful to get a feeling for where Auto WB or One-Touch (OT) WB placed a specific illumination condition on the illumination color temperature scale.
By examining the numbers in the EXIF information, it is also interesting to compare how Auto and OT interpreted a given situation. Noting the numbers obtained with OT for frequently occurring conditions and comparing them with the chart, one may decide on the best preset for these conditions, and save the hassle of OT.
It is interesting to note that the color temperature scale is a one dimensional variable and therefore cannot cover all possible colors (it’s the color of the light radiated by a black body heated to that temperature, in degrees Kelvin). The bias system used by the camera is two dimensional and can therefore correct for lighting situations not covered by a pure color temperature scale.
We can see from the chart and graph below that the values for 4000K – ‘White fluorescent’ – setting stick out from the smooth graphs. In the sample pictures which follow we also see that the setting for 4000K is different than all others.
Having said that, we may conclude that
presets may not always be capable to provide the color correction we can
obtain with OT.
With all other settings from 3000K to 7500K we see a gradual transition from very cool (blue) to very warm (yellow-red), while the 4000K image is all magenta, quite out of line with all the others.
Obviously this setting, intended to be used with fluorescent light, has a strong correction of magenta, to compensate for greens in fluorescent lighting. It’s called ‘4000K’, but I think that in fact it does not belong in the color-temperature line, rather, it is a special setting for fluorescent light.
I just took an OT measurement under a fluorescent desk lamp and obtained values of ‘Red bias’ 430 and ‘Blue bias’ 364, quite close to cloudy-daylight and different from the values the camera sets for fluorescents (R: 372, B: 562). It only shows that defining a light as ‘fluorescent’ is not enough, there are many varieties of ‘fluorescent’.
Following are sample images taken indoors,
with daylight coming from a large window on the left, changing the WB preset
from one shot to the next. The last image is with OT, it looks a bit yellowish
to me, Auto or 4500 - 5500 seems to render a more natural result. Contrary
to the accepted belief, it seems that Auto does a pretty good job.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
C4040 Preset Values
The following table shows the possible
WB settings on the C4040 camera, and the corresponding EXIF data:
|
|
|
Sunlight | Daylight | Blue +1 (17) |
Overcast | D55 | (+/-) 0 (18) |
Tungsten | Incandescent Light | Unknown (9) |
Fluorescent | Fluorescent | Blue +4 (13) |
Auto | Unidentified | Blue +6 (11) |
One Touch | Unidentified | Blue +7 (10) |
As seen, the values for OT and Auto
for this particular situation are quite close, as may also be seen from
the following images.
|
|
|
|
|
|
|
|
|
|
|
|
Unlike the E-20, the ‘Fluorescent’ setting for the C4040 seems to be quite in line, we have a gradual transition from cool to warm colors, as we go from Tungsten to Overcast. The Fluorescent settings seems not to have an aggressive green correction, as it has with the E-20. In this case it also seems that Auto does a good job (the shirt and trousers are a neutral grey!).
The C4040 has one more WB adjustment control in the menu, with which I did not play: ‘White Balance Adjustment’. By playing with this adjustment you could probably tune presets to correspond to specific One-Touch settings. The E-20 does not have this fine tuning feature.
I cannot know for sure, but it seems as if color correction of the C4040 is ‘one dimensional’ – along the temperature scale, while with the E-20 it is two dimensional.