Circular polarizer in digital cameras
Don Stauffer
circular polarizer, since metering and autofocus are supposed to be
polarization sensitive.
I don't understand how a circular polarizer would give the same effect
as the linear polarized filters I am used to. Those used to limit sky
brightness or reflections off of transparent surfaces as you rotated the
filter. Won't a circular polarizer merely shift plane of polarization?
Helpful person
If I understand correctly how this works you buy a filter that
consists of a linear polarizer followed by a retarder. The linear
polarizer takes care of the surface reflections and the retarder gives
circular polarization for the following optics.
The retarder should be achromatic. Other comments?
Phil Hobbs
Plastic retarders are probably achromatic enough--they're usually zero
order.
'Circular polarizer' is a horrible misnomer, but we're stuck with it.
If there were a retarder on both sides, it would change RH and LH
circular into orthogonal polarizations, select one of them, and
transform it back into one circular polarization. That would be a real
circular polarizer. As Richard says, though, there's only one retarder,
so it matters which way round you put it. Both are useful, but they do
different things.
(1) Used with the retarder facing the object, and with a matching filter
over the light source, it's useful for reducing glare.
Specularly-reflected light comes back with the opposite helicity (RH
turns into LH and vice versa) (*), so if both filters are the same
(passing RH circular, say), the glare will come back in LH and be
absorbed by the polarizer on the lens. (Back when computer displays
were shiny CRTs, you could get circular polarizing filters that reduced
the glare off the glass this way. They were pretty good at the time.)
(2) Used with the polarizer facing the object, a circular polarizer
works just like a normal polarizing filter, except that the transmitted
light is changed into circular polarization so that the polarizing beam
splitter and the Savart plate anti-aliasing filter on the CCD can work
properly.
Cheers
Phil Hobbs
(*) The helicity doesn't reverse if the angle of incidence is greater
than Brewster's angle, because the P reflection coefficient changes sign
there.
--
Dr Philip C D Hobbs
Principal
ElectroOptical Innovations
55 Orchard Rd
Briarcliff Manor NY 10510
845-480-2058
hobbs at electrooptical dot net
http://electrooptical.net
Helmut Wabnig
Don Stauffer
Thanks, guys.