Have I a Camera problem
Last weekend I attended the Jim Clark Rally and when I photographed this orange coloured car. The colour of parts of the car changed to yellow as seen in the photos I have attached. I used 7D with 70 - 200L lens. All the colours with the rest of the cars were perfect.
Today I photographed an Orange tipped butterfly and the tips of the wings that should have been orange were yellow. Once again using 7D with a different lens.
I would love to know the reason for this as it completely ruins a shot.
Bob
1
2
3
Today I photographed an Orange tipped butterfly and the tips of the wings that should have been orange were yellow. Once again using 7D with a different lens.
I would love to know the reason for this as it completely ruins a shot.
Bob
1
2
3
0
Comments
1) Notice that the orange colors are most pleasing, in this sequence, when the car is mostly in the shade. The shade is a color temperature of 7000-8000 degrees Kelvin, and the camera is properly sensing and using that color temperature for white balance.
2) In images 2 and 3 the car is getting more into direct sunlight (5000-5400 degrees Kelvin), but some of the car is still in shade. The camera is still using the shade color temperature to set white balance, which means that those parts of the car in direct sun are now incorrect white balance.
3) Worse, the parts in direct sunlight are getting much more illumination than the parts in shade, but the camera is only partly compensating because you are in evaluative-metering mode. This is causing the car parts in direct sun to be overexposed, blowing out the Red channel "and" Green channel.
4) Worse still, the car appears to be a "fluorescent orange" paint color, meaning that in direct sun it is "super" bright as the sun's ultraviolet rays strike the paint. This means that the paint didn't just reflect available light, but it also converted some of the UV into visible light, meaning that the orange could actually glow in direct sunlight. This further contributed to the overexposure of the orange paint that was in direct sunlight.
5) The camera was also set for ISO 400, contributing to the problem because it has less dynamic range than say, ISO 100.
Under the circumstances, with auto-exposure, auto-white-balance, ISO 400, mixed lighting with some direct sun and fluorescent orange paint, the results are fairly predictable to be not so good.
You just have to choose your battles and set up the camera accordingly. What I mean by that is that it would be best, in the future, to select either shaded or fully sunlit, realizing that mixing sunlight and shade will cause problems. Obviously, shade works better for this car and its paint.
In direct sun, and considering this paint scheme, it would be best to underexpose to protect the red channel. A low ISO would also yield the broadest dynamic range.
Finally, if you shot RAW files, you might have a chance recovering "some" of the overexposed area by careful processing including a very low contrast setting in the RAW conversion and some highlight recovery too, and then very careful processing afterwards to preserve as much of the red and green channels as possible.
Here are histograms of just the Red channel and Green channel of a bright, but not overexposed, portion of the car's hood/bonnet from image number 1, where the hood/bonnet is still in shade (the Blue channel is no problem in either image):
... and a similar patch of the car when the paint is in full sun (image 3):
Note how both the red and green are nearly saturated, but definitely clipped, in image number 3.
Moderator of the Cameras and Accessories forums
Thanks once again Ziggy very well explained.
I have attached another shot in the sequence which is in more sunlight. Thanks again Sand.
Bob
1