Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site onfcanim.UUCP
Path: utzoo!watmath!watnot!watcgl!onfcanim!dave
From: dave@onfcanim.UUCP (Dave Martindale)
Newsgroups: net.rec.photo
Subject: Re: Product Quality: Color prints from slides (contrast buildup)
Message-ID: <14708@onfcanim.UUCP>
Date: Fri, 25-Oct-85 02:58:30 EDT
Article-I.D.: onfcanim.14708
Posted: Fri Oct 25 02:58:30 1985
Date-Received: Sat, 26-Oct-85 06:20:15 EDT
References: <298@tekig4.UUCP> <349@vaxwaller.UUCP> <5746@tekecs.UUCP> <1505@utcsri.UUCP> <334@tekig4.UUCP> <2141@amdahl.UUCP>
Reply-To: dave@onfcanim.UUCP (Dave Martindale)
Distribution: net
Organization: ONF, Montreal
Lines: 55

In article <2141@amdahl.UUCP> ems@amdahl.UUCP (ems) writes:

>Something which has always bothered me, when you copy a print onto
>a slide (an perhaps other copies as well) you get 'contrast buildup'.
>It seems to me that if the contrast range of the print is *less*
>than that of the original subject, then the contrast range of the
>slide should also be *less* than the original.  So why does the
>slide show *higher* contrast?

There are two things to consider here - the contrast range that can be
*represented* by a particular medium, and the contrast ratios of "good-
looking" images in that medium compared with the contrast ratios of the
original.

A good print material is only capable of about a 30:1 range of reflectances.
Thus, when printing, you set your exposure so that some area of the print
comes out white, and anything in the image that is darker than 1/30 of this
is mostly lost in the shadows.  Transparency material, on the other hand,
is capable of 100:1 contrast range or more.

But representable contrast range has nothing to do with contrast buildup.

For a particular photographic material, if you plot the log of reflectance
(or transmittance) against the log of the illumination level, you will get
an S-shaped curve with a centre section that is roughly straight.
The slope of this curve is called "gamma".  If it is 1.0, then a given
change of luminance between two parts of the image will produce the same
change in the photograph - there is no change in contrast (provided both
luminances fall in the linear part of the curve, of course).  If the gamma is
2.0, for example, then a 2:1 luminance change in the scene will produce a
4:1 change in the final image.

Why would you want a gamma of other than 1?  Well, it seems that a gamma
of 1 is about right for prints, that will be looked at by reflected light.
But for images that are projected (slides, movies) or self-luminous (TV),
they appear subjectively "flat" or low in contrast if they have a gamma
of 1, reproducing luminance ratios correctly.  I believe that this has
to do with darker tones appearing subjectively lighter than they are
because the image is seen in a dark environment.  Anyway, an overall
gamma of about 1.5 seems to give these types of images enough additional
contrast so that they look correct.  So direct-positive films, television
systems, and negative-positive movie film systems are all designed to
produce a final gamma of about 1.5.  And they work well if you photograph
real images.  But if you photograph a slide to duplicate it, you have now
increased the total gamma to 2.25 (gammas multiply) and it looks too contrasty.

To avoid this problem, you need to do duplication with a film that has a
gamma somewhere down near 1.0.

In real life, things are a bit more complicated than the description
above, since the low-light end of the response curves of many films have
a greater gamma than at higher light levels in order to compensate for
the loss of contrast due to lens and camera (or enlarger or printer) flare.

	Dave Martindale