22 THE PHOTOGRAPHIC IMAGE, BROCK 
to settle the somewhat sterile arguments. Is there in fact any positive evidence showing 
a reversal or discrepancy between high and low contrast tests of photogrammetric lenses 
when used with the normal emulsions of air photography? Or is there any evidence of 
lack of correlation or reversal of orders when comparing high contrast tests with air 
photographs? The writer can bring forward no such evidence, but he has no high con- 
trast tests to call on, low contrast tests having always been accepted in England as the 
obvious and natural tests to do. Others may have evidence, but is it significant that high 
contrast tests are still done in the laboratory of a manufacturer who has probably made 
bigger advances in lens design than any other in the world? 
Advocacy of the high contrast test has often rested on the ground that it is more 
accurate. This statement could be queried, but was by-passed many years ago by Selwyn’s 
comment: “What is the point of measuring accurately something you do not want to 
know?” Howlett also has convincingly stated the case for doing this kind of test under 
conditions that simulate practical use as closely as possible. In the new language, we 
want to compare our lenses on the part of the frequency response curve that will be used 
in air photography, not on some other part which will never be used. The writer natu- 
rally sympathises with this point of view and cannot understand the argument that high 
contrast targets and fine grained emulsions should be used because they give a better 
differentiation of lenses. To differentiate them in an unused region may convey infor- 
mation to the designer (though purely optical tests would appear to do this better); it 
cannot be of much value to the user or his protecting agency. There is much more to be 
said in favour of a system that indicates quality at sizes larger than the low contrast 
resolution limit. 
Target contrast can affect accuracy in different ways. The density differences of 
the low contrast target must be held within very close tolerances, say 0.2 = 0.02, if con- 
sistent results are to be obtained. This is not easy, and the high contrast target gains by 
its complete freedom from such limitations. 
It is maintained by some that the accuracy of reading results is much better with 
the high contrast target, i.e. that the resolution end point can be more precisely fixed. 
It is not immediately obvious that this should be so; resolution is in both cases a point of 
zero contrast, and the small change upwards to just-visible contrast is the same whatever 
the original target contrast. If, however, the frequency response curve of the lens drops 
more rapidly at the higher frequencies which will be made visible by the high contrast 
target (and in general this will be so) then the end-point will be more sensitive to a 
change of image size, but then we return to Selwyn’s question. In fact, however, is it 
agreed that high contrast targets are easier to read? In the writer’s laboratory, some 
people at least have a strong preference the other way. 
A point which is not usually mentioned is the influence of graininess. The high con- 
trast target pushes the point of resolution down to lower size, hence graininess must be 
more obtrusive. What effect has this on accuracy? 
Finally there is the effect of inaccuracies in exposure. The curves of resolution 
against exposure for different target contrasts show that exposure is more critical at 
high contrast, for maximum resolution, and practical experience has confirmed this. 
42.3. Target shape. 
Although Cobb test objects have been in use in Gt. Britain for so long, the writer has 
no personal bias in their favour and can think of no good argument why they should be 
retained. They were introduced by a rapid decision in World War II and have persisted 
mainly to preserve continuity with existing results. In some respects they are an un- 
fortunate compromise, being too short to have the full attributes of lines, yet without the 
non-directional quality of the Howlett annulus. Their directional quality is indeed used 
by expert lens testers who like to match the astigmatic focal surfaces revealed by reso- 
lution te 
and the 
would be 
First, it 
grading 
the resol 
practical 
In t 
comparii 
to avera 
as the r 
Lon, 
lution te 
spurious 
grain cl 
seems to 
lution te 
itless ex 
4.24. S 
In T 
industria 
some th: 
discussec 
Targe 
its, say 
lengths. 
size, not 
and the 
size. In ( 
facture 
gives pel 
feétive « 
graphic « 
the effec 
Collim 
terms, pl 
the longe 
in Gt. B 
Light- 
rain in d 
which is 
sunlight 
testing if 
millisecs) 
light filt 
long exp 
incorrect 
chosen, ii 
ference o