32 THE PHOTOGRAPHIC IMAGE, AUTHOR’S PRESENTATION 
also to measure in principle; I think in practice 
it is not so easy but let us give all credit to this, 
that it is in principle a real physical property 
which you can define and measure. Of course, if 
all our measuring methods are good, it is in- 
dependent of the method of measuring, which 
is a very great advantage. 
It also should be sufficiently obvious that 
the CT curve is so much better than the re- 
solution figure, because it covers a range of 
frequencies; it is better in the same way that a 
transmission curve or a minus blue filter is 
better than a nominal specification of a cut-out 
line, which tells you nothing whatsoever about 
the transmission of the filter at other wave- 
lengths. 
Another very useful feature is that in prin- 
ciple you get out of this curve the behaviour 
of the lens for a whole range of image contrasts, 
which is very valuable. In general, the image 
contrasts in our photography are very low, they 
can be very low but they are-not always ex- 
tremely low; there are cases in low altitude pho- 
tography of towns, shall we say, where you can 
get very high contrast. So neither low contrast 
nor high contrast resolution testing in them- 
selves give you all the information you need 
and, of course, with the CT tests you can derive 
everything you want. I know of nothing else 
that gives you that comprehensive kind of per- 
formance assessment. 
As to methods of measurement, this still, it 
seems to me, is very much in a state of flux. AII 
kinds of methods are published, a lot of people 
use them and a lot more people talk about them. 
I should make our own position clear. At the 
Royal Aircraft Establishment we have not made 
very many measurements, for various reasons 
with which I will not bother you. However, after 
some fairly lengthy adventures in electronics we 
have for the time being gone back to measuring 
the spread function by photo-electric means, 
and doing the necessary mathematical analysis 
of that on the Mercury computor at RAE, which 
is slightly laborious but seems to be giving quite 
good results; unfortunately, it is a little early to 
be worth quoting any of them. 
For the other results we have seen, I ac- 
knowledge again my indebtedness to Professor 
Ingelstam for the only result on 6 inch wide 
angle lenses, and even these are on axis only, 
they do not give us off axis results. 
Passing on again to some more aspects of 
the frequency response approach I do agree that 
the analogy of frequency response in optical 
systems with frequency response in other phy- 
sical systems has been very fruitful. It helps us 
to a better understanding of limitation in the 
photographic process. However, I think we have 
to be on our guard all the time against assuming 
the analogy is perfect. I feel the photographic 
case is very different from many of the electronic 
cases, of which we very often think. It is so 
often the case in electronics that the amplifier 
or filter for whatever system is substantially 
perfect over the bandwidth in which you are 
interested. The gain can be constant and the 
phase shift negligible over the region you want 
to use. You can regard it, therefore, as easily 
specified by something like the half-power 
points, you can specify the band width in a very 
legitimate way. But in photography we are al- 
ways pushing up to the limits of the band 
widths. We are really interested in the region 
where the image is beginning to deteriorate, and 
this is the region where, in the analogy, the gain 
is falling rapidly and the phase shift is begin- 
ning to be important. To my simple mind, this 
makes the interpretation much more difficult 
and complicated. 
In conclusion, I say again that while CT tests 
are obviously the thing for the future, I do not 
think we have to burn up all our resolution tests 
and say that they are no use. They are still quite 
valid if they are properly carried out. We would 
say, especially if they are low contrast tests they 
valid. In that connection I am very interested 
to see a paper by Mr Bousky in the technical 
literature to this Congress, which is, in effect, an 
impassioned plea for low contrast testing. Mr 
Bousky is with Chicago Aerial Industries. 
The CT tests help us to a better understand- 
ing of the value and the limitations of our reso- 
lution tests and we should push on with them, 
but in the meantime we can carry on with our 
resolution tests and anything we can do to im- 
prove their precision is well worth doing. 
I do not want to take up the whole time I 
have been allocated, I think it is much more 
important that other people should have a 
chance to say something, so with the President's 
approval I will sit down and invite comments 
and questions on this subject. 
Mr E. WEL. 
excellent paper 
I am in full agi 
contrast functi 
estimating the q 
In Sweden 
functions for sc 
camera lenses. - 
lens function b 
the vibration, t 
In this manner 
for the entire a 
predictive funct 
function obtain: 
tographing a li; 
an altitude of 1, 
this on a slide. 
here]. 
On the horiz 
line and the vert 
the entire aerial 
fer. The line of 
termined curve 
result obtained 
agreement is p: 
limitations, we 
this could be ste 
functions in prac 
450 Sec and vel 
I think the imag 
this altitude. 
Mr.P. D. C. 
this subject so fi 
if not here verb 
say. I think the 
consider what w 
plication in the 
cameras. 
It does appee 
is not ready foi 
us with resoluti 
mendation, whic 
four different ty 
and this numbe 
the comments w 
tion with revisi 
have been some 
there have also 
that we needed 
targets. In gene 
opinion expresse 
felt it safe to m: 
from Mr Brock 
discussions and