14 THE PHOTOGRAPHIC IMAGE, BROCK
large-area threshold contrast sensitivity. The differences in the latter are mainly due to
gamma differences. It is interesting to note how such effects are sometimes reversed
higher up the curves, crossovers to the right showing the increasing effect of superior
micro-properties at smaller line-separations. The emulsion resolution at any image
contrast is given by the intersection with the abcissa of that value; for example at “0.2”.
This is some 25% higher than its resolution measured on Cobb test-objects mainly due no
doubt to the greater length space ratio of the sinusoidal test-objects.
In Fig. 8 the threshold curves for emulsions (2) and (6) of Fig. 7 are replotted
together with the C.T. curve for the 6" f/5.6 lens of Fig. 6. The latter is shown for low
contrast (log. luminance 0.2) as well as for black and white contrast. Some resolving-
power figures read from the graph are given in Table 2.
TABLE 2. Lines per mm. Resolution at different contrast.
High Speed Emulsion | High Resolution Emulsion | Ratio:
(a) (b) | b/a
High Contrast: 9 230
Perfect Lens 8: 2 2.6
High Contrast: ,
6” f/5.6 lens 45 71 1-5
|
Low Contrast: | : | 9 |
6” f/5.6 lens 27 4 1.8
Low Contrast: | | | ;
Perfect Lens 40 110 2.7
Considering first the emulsion with different contrast target imaged by a perfect
lens, we see that the high contrast resolution is approximately 2.6 times the low contrast
in both cases, the threshold curves being nearly parallel in this region. The 6” lens is not
good enough to yield the full emulsion resolution in either case, as its C.T. is low and
falling rapidly above, say 50 lines per mm. The C.T. curves for the 6” lens cross the
threshold curves more or less oblique according to the convergence of the latter and the
target contrast, and the horizontal distance between the intersections is correspondingly
affected. Thus it happens that the advantage of the high resolution emulsion is greater
at low contrast. It becomes even greater at still lower contrasts such as a log luminance
ratio of 0.08, before the two emulsions reach substantial equality at zero frequency.
The graph also shows that the lens C.T. should be kept as high as possible out to the
resolution limit of the emulsion; for example, doubling the C.T. in the 50 lines/mm region
would increase the lens/film resolution by some 60%, whereas extending it at the same
level to much higher frequencies would have no practical value.
Emulsion threshold curves have another application, as pointed out by Selwyn. With
such curves for one or more emulsions covering the required range of frequencies we
could determine the C.T. curve of the lens by measuring the combined resolving-power
for a range of testobject contrasts.
Fig. 9 shows a comparison, at low and high contrast, between two lenses of the same
construction and nominally equal performance *), with the threshold curve for a slow
emulsion. Lens A is better over the whole frequency band, but at low contrast its resolu-
tion is only 1.2 X better than that of lens B, whereas at high contrast the ratio is nearly
2.0. (This would be interpreted, according to taste, as proving the greater sensitivity of
*) based on curves published by Dr. Washer.
high con
the futi
them for
subjects)
clear th:
ence bet:
es wouk
on an e
as Supe:
detail o:
contrast
such as
photogra
Fig.
effect of
cus settii
lens, at
contrast,
the thre
for slo
emulsion
ting (A)
modulati
frequenc
sharp cu
100 line
(B) give
ulation
line regi
response
hundred
high con
focus se
qual res
on the s
(B), wit
cut-off,
resolutio
film. At
(B) is si
on both €
would .
superior
emulsion
Fig.10. 7
tings of
at high
trast, Wi
curves |
slow
