OK, this question is made mainly out of curiosity.

All around the place one could read "linear polarizers MAY cause problem with your AF SLR, use circular polarizers". Now I have perfect understanding what could go wrong: ratios between reflected/transmited light on insulating surfaces (if they are used for the beam-splitting) at different positions of the filter will vary and fool exposure metering and/or autofocus.

I wonder, do this kind of problems exists in real world? In particular do they exists with EOS 300?

Recently I've got for a short time general purpose (not photo) linear polarizer and tried to do a little bit of experimenting. It seems that there is no effect on exposure metering and regarding autofocus performance, it was deteriorated but may be the reason was that the filter was not quite clear (it blurred picture substantialy).

Before I started to the photoshop for further experiments, may be somebody knows the answer?

-- Gombo Tsydynjapov (g.e.t@newmail.ru), December 23, 2001

Answers

Your results won't be entirely predictable using a linear polarizer. Or at least they won't seem to be. If you have the polarizer turned one way, your exposures & AF will be correct. Turn it 90 degrees and your exposure & possibly your AF will be incorrect.

You can see the difference if you have acces to both types. If you use a circular PL, depending on your orientation to the sun, you will see the viewfinder darken as you turn the filter, and the exposure value will change a bit as well. Using a linear PL you will see the same darkening effect, but some of your exposure values will be different than they were with the circular PL. If you use color negative film, the printer may correct for this exposure errors and you might never notice.

The linear PL will effect the AF in the same way. In some positions it will work and in some it will not. You can always use the linear anyway. Set & lock your exposure and AF without the filter. Then attach the filter and adjust your exposure to compensate.

-- Jim Strutz (j.strutz@gci.net), December 24, 2001.