18 THE PHOTOGRAPHIC IMAGE, BROCK 
3.1. Frequency response measurements on lenses and emulsions. 
Measuring the frequency response of lenses is much more difficult than measuring 
their resolving power, mainly because it involves photometry on a microscopic scale. 
Photographic resolving power measurements require a collimator with illuminated tar- 
gets plus some mechanical arrangements of reasonable precision for positioning the 
camera or lens. Given that, and very little else, work can commence. Contrast transmis- 
sion requires the same «essentials, plus a precision arrangement for scanning the aerial 
image with a photo-multiplier tube behind an extremely fine slit (or equivalent appara- 
tus), with associated amplifiers and recording gear. Though Prof. Maréchal points out 
S L C T 
TARGET SEEN AS 
S Wl 
Fig.13. Elements of a system for measuring frequency-response of lenses. 
  
  
  
  
  
  
  
| 
| 
| 
     
    
  
  
  
  
  
  
| 
| 
1 
| 
  
that there are possibilities. These are the essential features of the system, but they have 
been applied in diverse ways, and there is no sign of standardisation. The descriptions 
which follow are not meant to be complete, but merely to indicate representative tech- 
niques. 
The most direct approach would be as shown in Fig. 9. 
Light from an illuminated opal traverses the target 7, which varies sinusoidally in 
transparency as suggested by the shading of the section and the facing view inset below. 
The lens L under test images the target via the collimator C on a fine slit S forming 
the front of a housing for the photo multiplier cell P. The cell output is amplified at 
the D.C. amplifier A and the signal is registered on the recorder R. The photocell can 
be traversed across the image to record the sinusoidal amplitude variations. 
In practice this would be a difficult apparatus to make and use. For photographic 
lenses it is desirable to measure responses down to 100 lines/mm, i.e. 10 nu. For accurate 
tracing at this size the slit should not be much wider than 1u and the traversing of the 
cell should be done to the same accuracy. To avoid the necessity for such fine working, 
the aerial image is normally enlarged by a microscope objective between S and the focus 
of L. The performance of a large aperture well corrected microscope objective is so good 
that its effect on the image formed by the photographic lens can normally be neglected. 
Variable density targets of the kind shown are very difficult to make, and for televi- 
sion work are often replaced by black and white lines. (Square waves). 
Measurement step by step on a series of targets is very laborious, and the work can 
be speeded up by various expedients. One method is to mount a series of targets, forming 
a continuous progression, around the axis of a drum, which is rotated at constant speed 
with the illuminant inside. An opened-out view of such a target in line form is shown in 
Fig. 14. This rotating target is substituted for the targets in Fig. 13. The D.C. amplifier 
is replaced by an A.C. amplifier and the signal is displayed on an oscilloscope as a 
trace of 
wave if 
A € 
analyses 
lenses ii 
the inte 
plier sli 
is broac 
contrast 
transfor 
a rotati 
As the 
signals. 
of slits 
high orc 
electron 
measure 
frequen 
first of 
more a 
ysis of 
contrast 
the cost 
As 
plicity v 
particul: 
and the 
taining 
Son 
bility of 
response 
then to 
use and 
giving t 
of repeal 
determir 
Som 
been me 
The met 
sinusoid: 
densitom 
the emul 
could be