50 
SYMPOSIUM PHOTO INTERPRETATION, DELFT 1962 
looking at the stereoscope from a distance), measured with respect to the last 
glass surface. This property is of great importance, because the pupil of the 
human eye should coincide with the exit pupil of the instrument in order to 
use the total field of view. An eye-clearance of 15 mm (when using spectacles, 
25 mm) is sufficient if the rim of the eye-piece is shaped properly. Most stereo 
scopes do not deserve the qualification “good” for this property. 
16. Rotation of one image with respect to the other, caused by misalignment of 
mirrors and prisms, results in poor stereoscopy. Correction of the effect by 
counter-rotation of the print is not allowed when parallaxes have to be meas 
ured. The easiest way to measure the relative rotation of one image with 
respect to the other is by stereoscopic observation of two lines. Tolerances are 
specified for a normal field of view; an accuracy of i 5 mrad can be obtained 
and systematic errors of 10 mrad can be checked with the help of a pocket 
stereoscope. Most stereoscopes receive the qualification “tolerable” or “good”. 
21. Although “Scanning” belongs more to the application of the stereoscope, 
it might well be an important aspect of the performance of the instrument 
for prolonged use. It is not the scanning itself which is tiresome, but the control 
by eyes and brain of the tolerable y-parallaxes. With scanning stereoscopes 
and the use of rectified prints, these parallaxes will be small and stereoscopy 
is maintained over the complete model. Maybe this is, in the long run, more 
important than a perfect image quality. 
34. Flare is expressed as a percentage of the average illuminance of the 
print. It is caused by all optical surfaces and depends on such factors as the 
polishing and coating of the surfaces of lenses, and scratches and dust on these 
surfaces. Flare is measured with, for example, a densitometer; the result 
depends on the type of densitometer, the magnification and the area which 
is illuminated. The observation of details in the dark areas of the print is 
hampered by flare. Our measurements indicate that flare is one of the most 
important single properties of the stereoscope. Qualifications run from “good” 
to “poor”. 
41. Image Quality is expressed nowadays in Modulation Transfer Functions 
(see paper by Welander), so we determined the M.T.F. for the central rays 
(“on axis”) for a number of stereoscopes. As expected, the Transfer Function 
of a stereoscope can be approximated by a Gauss Transfer Function. From 
this we derive the linespread function, also a Gauss function, which is specified 
when 2u is given. This 2cr (in fj.m) has the physical meaning of the minimum 
width of a thin line, when seen through the stereoscope. The T.F.’s are com 
parable properties, because magnification is not incorporated. The qualifica 
tion corresponds roughly to the performance of the optical design. The proposed 
specifications are based on the results of the measurements, because existing 
literature is of no help. 
51. The Position of the eyepiece is of utmost importance to the operator who 
has to use the instrument for his daily work. The large arteries that nourish 
his eye and brain should not be constricted by a bent neck. Looking down 
wards at an angle of 45° seems to be the best; looking vertically downwards