2.
Session 2 Sept« 20th
Mr, II. D. Rosenau, Perkin Elmer, TJSA. gave a paper on the application
of the 0.T. F. concept to the prediction and analysis of the performance of a
complete photographic system. The prediction process wr.s discussed in detail,
stressing the importance of using the O.T.F. of the lens appropriate to the
spectral distribution of the illuminent, and he illustrated the processes
involved in computing' this. Some of the factors involved were variable or
uncertain e. g. image motion, camera focussing, so that the estimate of the
total system performance must involve a probability element. The complete
estimate requires the use of a large computor. The analysis of air photographs
to compare with estimates, and diagnose camera faults is complicated by factors
which vary between successive frames. The use of tests of significance, for
example Students t-test, in comparing actual and predicted performance.
In discussion, Prof. Hallert drew attention to the need for proving
that the errors in the data to which Students t-test applied were normally
distributed. The paper was very interesting but he wondered whether situations
would arise very frequently in which the prediction of performance needed to be
done so thoroughly.
Mr. \7. L, Attoya, Itek Corporation, USA, described an exercise in the
prediction of the contrast transfer function of a lens-film combination from
the c. t.f.*s of the two elements. He emphasised the complicated nature of the
film response, even excluding considerations of granularity. The H and D curve
of small areas of film is dependent upon the spectral frequency of the impressed
light pattern for example. The c.t. f. also depends on the mean density level of
the film, contrast transfer being better at higher densities. The importance
of all aspects of processing in affecting to c.t, f. of the final image was
stressed. The c.t.f. of film is a very sensitive quantity which requires extreme
care in measurement to obtain consistent and meaningful results.
In discussion, it became apparent that caution is necessary in inter
preting the results of this paper, since granularity effects were ignored. In
practice the higher densities at which c t. f. is enhanced give higher granularity
and in ordinary use this would be apparent. The application of c. t. f. methods to
the prediction of system performance is by no means simple and the interpretation
of the results will still generally involve the computing of the resolution from
the c. t. f., whose direct significance to the user is not obvious.
Mr. U. A. Uoodley, Royal Radar Establishment, U.K. described the use of
O.T.F. *s in connection with the cathode ray tube recording of radar signals. The
CRT phosphor has, like photographic film, a granular structure and some light
spread. In addition the election beam striking the phosphor has a finite diameter,
so that the overall behaviour of a CRT has similarities to that of a camera.
RKE had evolved techniques to measure the response of CRT’s and had extended these
to measurement of 0.T.F.’s of lenses, using the CRT to generate a sinusoidally
modulated test object. The degree of precision aimed at was not as high as
in other methods described in the literature but the method was simple and gave
useful results.
3. Session 3 Sept. 22nd
Dr. H. K. Meier, Zeiss Oberkochen, Federal G-ermany discussed the work
of Heynacher on the formulation of a criterion for picture quality in terms of
O.T.F. and its extension to photogrammetry. (Heynacher 1 s original work had been
/done