RESEARCH ON VISIBILITY: INTENSITY-MODULATED SCOPES 



123 



sometimes accompanying higher PRF's is 

 due to screen brightening which is, in visual 

 effect, equivalent to reducing bias. The 

 writer made a simple test of this by varying 

 PRF from 25 to 1000 pps, in 11 steps, ad- 

 justing the bias at each PRF so as to keep 

 the sweep line just barely visible. The com- 

 pensation required over this range was 1.3 

 volts of bias; that is, the screen had to be 

 brighter, by an amount produced by 1.3 

 volts of bias reduction, with a PRF of 25 

 than with a PRF of 1000. However, when 

 bias was thus adjusted for each of the 11 

 PRF's, the threshold intensity of a medium- 

 sized pip was absolutely constant, within 

 the error of measurement. In short, visi- 

 bility was independent of PRF per se; it was 

 dependent on the brightness consequences 

 of PRF. This is precisely what we would 

 expect when using dim scopes. Likewise, if 

 the scope is already too bright, additional 

 brightness caused by increasing PRF may 

 be unfavorable rather than favorable. (See 

 Fig. 9.) 



Video Noise. The whole problem of noise 

 (PPI "snow" and A-scope "grass") deserves 

 better treatment than can be given here. 

 The mathematics of noise reception and fil- 

 tering has not been completely worked out, 

 especially in relationship to video indicators, 

 though some treatises are available on the 

 subject (46, 50). From a perceptual stand- 

 point, noise tends to reduce homogeneity 

 of the screen. It also adds brightness. 

 High frequency noise may merely make the 

 PPI screen look "grainy"; low frequencies 

 may make it appear "blobby" and patchy. 

 Large beamwidth targets, appearing as arc 

 line pips, may readily be distinguished from 

 most noise blobs; but extremely small pips 

 may resemble the shape of a noise pulse so 

 much as to be distinguishable only on the 

 basis of brightness or recurrence. A really 

 satisfactory experimental video noise, vari- 

 able in specturm and intensity, is not known 

 to have been employed in any experiments 

 to date. The data which come closest to 

 satisfying the requirements of spectrum 



analysis are those referred to above on i-f 

 bandwidth. Evidence on the brightening 

 effect of noise can be gleaned from Fig. 2. 

 The optimum of the visibility-bias curve 

 appears to be shifted about 1.5 volts to 2 

 volts; this particular noise (random; wide 

 band pass) was lighting up the screen in an 

 amount equivalent to 1.5 volts bias; being 

 fine-grain noise, it did not reduce the ulti- 

 mate visibility of the pip. In contradiction 

 to our earher conclusion (55), based on tests 

 of non-random noise, the presence of ran- 

 dom noise definitely lowers the optimal bias, 

 probably in proportion to its rms value. 



Optically simulated noise was used in one 

 experiment by Hopkinson (24). He took 

 photographs of operational radar scopes and 

 projected a simulated pip from a light source 

 on to the photographic representation. 

 Only prehminary data are reported, but 

 these show that the contrast threshold is the 

 same function of background brightness, 

 with or without noise. 



The clutter that appears on operating 

 radar scopes is not all noise; much of it 

 consists of echoes from the sea nearby, from 

 clouds, birds, and objects of no operational 

 significance. This clutter is a serious haz- 

 ard to detection; but being difficult to re- 

 produce in the laboratory, it has been largely 

 ignored by the researchers. 



Visual Scanning and Search Methods 



Search Area 



As a radar operator searches a PPI scope 

 for new targets, he must continually keep 

 his eyes roving over the scope, usually follow- 

 ing the sweep line, hunting for new pips. 

 Is a whole scope less efficiently scanned than 

 a part of it? Unpublished data of the 

 writer's (57) showed surprisingly little ef- 

 fect on visibility of variations in fractional 

 area of the scope to be searched. In a com- 

 parison of I with the entire 6-inch scope, 

 appearance thresholds were only about two 

 db better for the smaller area. This per- 

 haps ought not to be surprising in view of 



