STATISTICS OF TELEVISION SIGNALS 



755 



Fig. 3 — Close-up view of slide holding assembly and shifting mechanism of 

 picture autocorrelator. 



amount. The averaging process is inherent in such a measurement. 



The apparatus used to measure autocorrelation is sho^^^l in Figs. 1 

 and 2. The chamber at the bottom contains a Ught source of very 

 constant intensity and a convex lens to collimate the light. The middle 

 part, made of accurately machined aluminum, holds the two identical 

 slides of the picture under test, and an aperture exposing a large circular 

 area of the slides. The top chamber contains a collector lens and a photo- 

 multiplier tube which (on a microammeter not shown) gives a sensitive 

 indication of the total light transmitted through the slides. Fig. 3 

 shows a close-up \'iew of the slide-holding assembly. Two close-fitting 

 graduated aluminum rings permit accurately determined rotation of 

 both slides or one slide, and the micrometer drive permits translational 

 displacements measurable to within one mil (moving the two slides by 

 equal and opposite amounts); the separation between picture elements 

 is approxiinately 7.5 mils horizontally and 5 mils vertically (for the 

 2Y by 3i" shde size used). 



The light transmission is always a maximum when the two slides are 

 in precise register (As = 0). For large shifts the transmission fluctuates 

 about a nonzero asymptote. The nonzero asymptote results from the 

 fact that the average transmission is always positive, and the fluctuation 

 from the fact that large displacements introduce substantial amounts of 

 new picture material into the aperture. Since these components tend to 



