Path: utzoo!attcan!utgpu!watserv1!watcgl!imax!dave
From: dave@imax.com (Dave Martindale)
Newsgroups: comp.graphics
Subject: Re: photographing screens
Keywords: film reciprocity error
Message-ID: <1989Nov30.173755.814@imax.com>
Date: 30 Nov 89 17:37:55 GMT
References: <3777@hydra.gatech.EDU> <18678@watdragon.waterloo.edu> <1406@ulmo1.mt.luth.se> <24526@cup.portal.com> <264@fe2o3.UUCP>
Reply-To: dave@imax.com (Dave Martindale)
Distribution: na
Organization: Imax Systems Corporation, Oakville Ontario
Lines: 47

In article <264@fe2o3.UUCP> rusty@fe2o3.UUCP (Rusty Haddock) writes:
>
>How 'bout checking the reciprocity correction for ``long''
>(< 1/15-sec) exposures?  Sure, it'll vary from film to film
>but I haven't seen anyone mention this yet. Still, it might
>be a good idea to verify that you may or may not need to
>increase the exposure or add a filter here and there. 

The reciprocity data for films assumes a constant exposure level
during the exposure time.  If you're making a 1 second exposure
of a normally-lit scene, the light reaching the film is pretty dim,
and reciprocity compensation is needed.

But if you are shooting a CRT, the light actually reaches the
film as 60-70 short pulses of light each second, each of which is
considerably brighter than the average light level.  I would expect
that little or no reciprocity compensation is needed.  (I've never
used any).

The limiting case of this effect happens with digital film recorders.
The entire exposure may take 5 minutes, but each point on the film
is exposed to light of each colour exactly once, with an "exposure time"
that is determined mostly by the decay time of the CRT phosphor -
somewhere in the microsecond to millisecond range.  If reciprocity
compensation is needed, it's due to very short exposure time, not
long exposure.


About using spot meters:  If you have a spot meter that will reliably
meter a flickering source (the Pentax digital will not, the Minolta
digital will, and any analog one should), measure a full-intensity
white area in the middle of the screen, then give about 2.5 stops
more exposure than that.  Or, if you meter understands making
highlight readings, do that, and it will make the 2.5 stop compenstation
internally.

Either way, you ensure that the highest intensity that your CRT
can produce will fall at the upper edge of the linear portion of
a slide film's response curve, guaranteeing a good exposure with
maximum shadow detail.  (Negative films have a longer range, and
you might be able to get better results with more than 2.5 stops
increase).

This method is actually better than measuring an "18% grey" patch.
The actual brightness of an "18% grey" patch depends on how accurately
you do "gamma correction" for the monitor, while white is white no
matter how messed up your gamma correction is.


Brought to you by Super Global Mega Corp .com