lt than measuring 
microscopic scale. 
h illuminated tar- 
or positioning the 
Jontrast transmis- 
'anning the aerial 
quivalent appara- 
aréchal points out 
  
e of lenses. 
em, but they have 
. The descriptions 
presentative tech- 
ies sinusoidally in 
view inset below. 
ne slit S forming 
ut is amplified at 
The photocell can 
tions. 
For photographie 
10u. For accurate 
traversing of the 
uch fine working, 
n S and the focus 
jjective is so good 
nally be neglected. 
e, and for televi- 
and the work can 
f targets, forming 
at constant speed 
form is shown in 
"he D.C. amplifier 
oscilloscope as a 
THE PHOTOGRAPHIC IMAGE, BROCK 19 
trace of amplitude against frequency. The square wave response can be corrected to sine 
wave if necessary. 
A different group of methods dispenses with periodic gratings of any kind and 
analyses the image of a fine slit or series of slits. In a method used for testing television 
lenses in England the test object in Fig. 8 is replaced by a single illuminated slit, and 
the intensity distribution in the image of this is measured by traversing the photomulti- 
plier slit across it. A perfect image would be a line of zero width, but the actual image 
is broadened as suggested in Fig. 15. The Fourier transform of this curve gives the 
contrast transmission. Graphical and electro-mechanical methods for performing the 
transform have been devised. In another method using slits, several are mounted around 
a rotating drum and imaged via a collimator and the test lens on to a fixed search slit. 
As the moving slit images sweep past the search slit they generate a series of pulsed 
signals. Such pulses contain the fundamental frequency, corresponding to the number 
of slits and the rate of rotation, plus all the harmonics at equal amplitude up to very 
high orders. The resulting complex wave-form from the photomultiplier can be analysed 
electronically (by a “multiplication” process) for 
measurement of the relative amplitude of all the 
frequencies in the image cast by the lens. The 
first of these slit methods is simpler and probably 
more accurate, but requires mathematical anal- 
ysis of each energy distribution curve before the 
INTENSITY 
  
  
  
  
  
  
  
  
  
  
DISTANCE 
Fig, 14. Progressive-frequency Fig.15. Intensity distribution in 
line test-object. the image of a fine slit. 
contrast transmission can be derived. The second method will give results quickly, but at 
the cost of complex electronics. 
As yet, no accurate method for determining contrast transmission competes in sim- 
plicity with the resolution test. With the improvement of techniques, however, and in 
particular the automation of the process as far as possible, this situation will change, 
and the difficulty will become that of assimilating the information rather than ob- 
taining it. 
Some of the practical difficulties are concerned with the reproducibility and relia- 
bility of electronic amplifiers. These would be less serious in the application of frequency 
response technique to the focusing of cameras. The object of the focusing operation is 
then to obtain maximum response at some definite frequency chosen to suit the film in 
use and the kind of scene photographed. Suppose this to be 20 lines per mm, a target 
giving this frequency is chosen and the focus varied for maximum response; questions 
of repeatability do not arise and the operation could be much quicker than a photographic 
determination. 
Some difficulties in the way of measuring film contrast transmission have already 
been mentioned. Subject to these, the methods can be essentially the same as for lenses. 
The method described for measuring threshold contrast as a function of frequency, using 
sinusoidal test objects, could be adapted to contrast transmission measurement by micro- 
densitometry on the image. Alternately, micro-densitometry on a slit image recorded by 
the emulsion, followed by the transform operations on the transparency distribution curve, 
could be used.