uency limit of the 
as sine waves. In 
e lines limits the 
» frequencies much 
ilts on a 400 lines 
je" at its limit by 
(The harmonies 
tal amplitude and 
ns? Can they help 
st methods, better 
  
and series of 
pear to be: “Yes”, 
ions in object and 
now preferred to 
"his gives a simple 
smits. 
a contrast-transfer 
or large areas, and 
rical expression to 
THE PHOTOGRAPHIC IMAGE, BROCK 11 
the contrast-transfer for large areas and to the spatial frequency at which the contrast- 
transfer becomes zero, but we cannot show contrast-transfer as a continuous function of 
frequency. This would not matter if the normalised function could be shown to have the 
same shape for all combinations of emulsion and opties; in practice the shape is not 
always the same, and we stand to learn more about the behaviour of the system by having 
full information about its contrast-transfer over the whole working range of frequency. 
There is an important difference between the CT behaviour of lens and emulsion. 
The CT funetion for the lens is quite independent of the absolute levels of luminance 
in the object; whatever these may be, the transmission of contrast at a given frequency 
will always be the same. In the lens only one contrast-reducing factor is at work (Except, 
of course, for veiling glare). In the emulsion, however, we can recognise two processes; 
an optical reduction of contrast due to scatter, which is frequency-dependent but linear 
with intensity, plus a recording process (making the optical image visible) which is not 
linear with intensity and carries with it the random structure known as granularity. For 
perfect reproduction the transparencies of the negative should be inversely proportional 
to the corresponding object luminances, i.e. 
B, T, 
B, T 
Strict proportional only occurs, however, on the straight part of the characteristic 
curve, and if the slope of this is y, 
B, n 
so, unless y is unity, the relation is not linear, even when B, and B, are not widely dif- 
ferent. For large differences of image illuminance, as with a black and white object, the 
curvature of the lower part of the characteristic also comes into play, and further dis- 
tortion takes place. In the electronie analogy, a non-linear characteristic generates har- 
monies, e.g. a quadratie curve generates even harmonies. A square wave of say 3 lines/ 
mm has odd harmonies at 9, 15, 21, 27, etc. lines per mm. On a fine grained emulsion 
capable of resolving say 60 lines/mm low contrast these would be reproduced with small 
loss up to the 9th harmonie or so, but non-linearity should by analogy generate other 
harmonics, e.g. 6, 12, 18 etc. lines per mm which were not in the original wave and must 
distort its shape. How far the analogy holds is not known, but clearly these effects need 
to be worked out and their implications studied. There can be no doubt that non-linear 
photographie reproduction of a single sine wave will distort its shape, which is merely 
another way of saying that harmonies are introduced. Since the overall gamma in air 
photography is normally greater than unity, the non-linearities of the negative stage 
will not be compensated in the printing stage and the inference is that some distortion of 
edges is always present. There would appear, in fact, to be inevitable penalties in in- 
creasing contrast as a device to increase apparent sharpness. 
Also there are the neighbourhood and tanning phenomena acting to spoil the faithful 
rendering of small contrasting details by emphacis on particular frequencies. 
Fortunately, in air photography the image contrasts are small enabling some of 
these effects to be reduced by keeping to the straight part of the characteristic curve, or 
by working over a small part of a non-linear characteristic. Nevertheless, it appears that 
there is a good deal of work to be done before emulsion frequency response can be defined 
in unambiguous terms. In the meantime it. is probably better to make determinations 
under the conditions of practical use accepting any non-linearity due to working at high 
gammas, and keeping target contrast small. A few CT curves have been published for 
emulsions of interest in air photography, but the precise meaning of such curves is not 
clear. They usually show the CT function as unity at low frequencies, yet the quoted