approaching the theoretical cutoff. At zero frequency, however, the 
measured value for the standard deviation of the data is more than 
twice that of the calculated normalization error. He explains this 
by pointing to the possibility of other sources of experimental 
uncertainty in addition to shot noise. 
His basic premise that the noise can be held to an acceptable 
level is upheld, however, and for the specific lens and test parameters 
used is 0.0035 MTF units at zero spatial frequency and approximately 
0.018 MIF units at 1200 cycles/mm. 
5.9.6.2 Dutton/O theorizes that OTF values, derived from data 
obtained by radiometric scanning of test images are subject to 
random fluctuations due to several stochastic processes, mainly shot 0 
noise effects in the photodetector. Therefore, in his investigation 
of equipment parameters to reduce the noise to a minimum, he uses 
for a model the methodology of obtaining emperical noise power 
spectra from sampled data. Formulae were then developed relating 
the random component of the OTF result to experimental parameters 
such as photo-current, detector gain, scan length sample space, 
scanning velocity, etc. When the formulae were compared with 
experimental results, there was good agreement. 
In commenting on similar studies made in Japan by Takeda 
and 0se/1,72, Dutton says that the Japanese study i developed a 
a general matrix formulation for noise contribution and applied it to 
slit and edge scanning; magnitude and frequency dependence of the 
noise in the answers is obtained in terms of observable system 
parameters. Although the American model and the Japanese model 
appear rather different, final formulae in the two papers can be 
shown to be equivalent. Slit scanning is reported to have some 
theoretical advantage over edge scanning balanced, however, by 
practical disadvantages in many cases. Both Dutton/Ü and Takeda and | 9 
0se/? show that excessive scan length is deleterious when using an 
edge. 
5.9.6.3 Image Evaluation’? 
From the vantage position of years of teaching and designing 
of optics, testing of lenses for design evaluation and for quality 
control, and optical shop management, Hopkin reevaluates the 
status of OTF, offering some practice suggestions: 
OTF testing (says Hopkins) is not simple if data necessary 
to evaluate the lens MIF characteristics for focus, field angle, 
and spectrum are required. Data can become voluminous and a figure 
of merit (FOM) summing this data is needed such that evaluations 
and comparisons of quality can be made. (Such data is requisite 
for high performance aerial and space optics if optimum information 
gathering capabilities are assured.)