Quote Originally Posted by ic-racer View Post
The typical 'in camera' exposure index test checks that the system of lens, aperture, film, shutter and meter will produce an adequate exposure of the film. Using a flash for the test negates some of the benefit of the test.
Quote Originally Posted by wiedzmin View Post
at Ken Nadvornick - flash at full power test - 95% the same, 5% 1/10th stop lower - ok, thank you for suggestion

at ic-racer - I'm using the same camera, film etc , only difference is that I would like to use flash with gel to match color of the sunlight in afternoon ~4500K, could you explain more what using flash will negate?
Don't want to steal ic-racer's thunder, but I think what he's referring to is that in a film speed (EI) test you normally want to include all of the (major) real world variables that will have a significant impact on your real world exposures. Soup-to-nuts, as it were. And two of the biggest variables with potentially large error spreads are your shutter and your meter. (Especially your shutter.)

By formulating the test using electronic flash, you will effectively remove both of these potential error-inducing variables from your results. Exposures will be determined by aperture alone without regard to the shutter (camera body or lens) you will normally be using in the field. Likewise, your regular continuous-light meter (in camera or handheld) will not be used during testing either since the light source is electronic flash.

For example, I've done EI tests also. One of the things I did in preparation was to use a Calumet shutter tester (no longer available new) to pin down the actual speed of the shutter I intended to use. That became my standard against which other shutters were later calibrated. At least to the extent necessary to give reasonable, usable results. No need to go overboard.

I followed a similar procedure for the meter I used. But meters can be a little more difficult. Most are non-linear across their range. And given the differences in sensor types, age, calibration drift over time, etc., most are not similarly non-linear. One of the ways to factor this particular potential source of error into testing is to make your meter reading using light levels similar to those you will be using in the field. Meaning, film speed tests (the so-called Zone I test) would be metered against a black surface, while the development time tests (Zone VIII) would be metered against a white surface. By not including your normally used meter in the testing chain, this error might later—and frustratingly—creep into your real-world exposures.

None of this will necessarily invalidate your tests. But it's always good practice before running an experiment to try to anticipate in advance the possible sources and sizes of error which may creep into your system. That way when things don't turn out perfect you might be better positioned to interpret your data and still glean useable information from it.

After the inital "Oh my god, that can't be right" it's always nice to be able to look back at your notes and discover "Ahh. Never mind. I can explain that. Factor it out and this test is still valid. I don't really need to redo everything."