remely high reso-
ion. These curves
w that focus set-
v& depends to some
ent on the kind of
ail being photo-
iphed and the emul-
n in use. There are
ses which show this
ssover effect to a
ater degree and at
er frequencies, so
t even on Super
at low contrast
focus is not al-
ys unique.
Given the neces-
y data, these ex-
ples could be mul-
lied, but enough has
n said to show that
phenomena of pho-
graphie resolving
ver can be given a
sonable — explana-
| in terms of the
quency response of
components of the
tem. It will also be
uw that there may
considerable reduc-
is in modulation at
quencies well below
resolution limit,
| that there is no
que relation be-
en the high and
frequency per-
mance. The grosser
'repancies which
"ht be brought in
the lenses are how-
r reduced by the
;riction of the work-
range due to the
h frequency cut-
Fig. 12.
ıtrast-reduction at
ves in the produc-
| of a photograph.
THE PHOTOGRAPHIC IMAGE, BROCK 17
off of the emulsion, and by the low contrast of the object detail, since lenses tend to con-
verge to equal performance at very low frequencies.
So far we have used C.T. eurves for lenses only, to explain phenomena of photo-
graphie resolving-power, as this could be done without any reservations. Some problems
in measuring and using C.T. curves for emulsions have already been mentioned, but the
general shape of such curves is illustrated by those in Fig. 11, selected from published
examples. The curve for the high-speed panchromatic emulsion is somewhat similar to
that for the 6” f/5.6 lens of Fig. 6, indicating that such a lens and emulsion contribute
about equally to the image losses, as indeed was formerly thought.
The great attraction of frequency response methods, however, is that they can be
applied (in principle, at least) throughout the successive stages of the photographic
process. Given the contrast transmission of each stage, and plotting on log scales, we can
derive the overall response of any combination of stages by addition of ordinates. Fig. 12
is an attempt to show this. Labelled C.T. curves are given for the taking lens, the nega-
tive film emulsion, an the positive film.
Approximation and estimation had to be resorted to, because exact data were not
available for the lenses and materials used in photogrammetry. However, it is considered
that the basic curves are sufficiently close to reality to give a good picture of situations
that could arise in practice. Adding ordinates for the taking lens and negative emulsion
we get the negative response curve which naturally shows a greater loss of modulation
than either lens or emulsion alone, particularly at higher frequencies. Assuming a fairly
good diapositive emulsion, and no printing losses, the further loss of detail at the positive
stage (addition of ‘negative’ and ‘positive film’ curves) is less serious. However, the cur-
ves only show the modulation characteristics for the materials used at optimum density,
and taken over the full density range of a paper print the losses would be far more serious.
If the diapositive is to be made by projection, we require a further addition for the lens
characteristics, in this case assumed to be the same as the taking lens. The loss is notice-
able but not excessive.
Diagrams such as this show why there is always a loss at each photographic stage,
even though the following stages might have superior resolving power to the first. Since
no stage has a sharp cut-off in its frequency response such losses are inevitable. The
final curve in Fig. 8 shows heavy loss of modulation at 10 lines per mm, even though the
worst individual component of the series shows some response out to 100 lines/mm. From
the general shapes of the curves it could be said that to avoid all loss of modulation when
cascading these processes, each would need to have a limiting resolution an order better
than that of the stage it accepted.
There is no reason why the cascading operation should be confined to the purely
photographic stages. In principle the frequency response can be determined for every
operation involved in taking air photographs, (including an analysis of the frequencies
present in a perfect image at the working scale) and it is desirable that this should be
done, so that the relative importance of all stages may be assessed on a common standard.
It will be appreciated that this brief account of frequency response applied to photo-
graphy has presented the subject in a very simplified way and has ignored many things,
e.g. the importance of phase, and the distortions introduced by gammas greater than
unity. These matters must await detailed study as experimental data becomes available.
3. Measurement of contrast transmission.
The preceding section may be regarded as pleading the cause of contrast transmis-
sion, and has tried to show its attractions, which could be said to consists largely of ex-
plaining in more elegant ways effects which are already known. To complete the intro-
ductory picture before attempting to assess the value of frequency response in photo-
grammetry, a brief account will now be given of methods of measurement.
