620 BELL SYSTEM TECHNICAL JOURNAL 



In this position, a complete cycle of black and white is obtained with 

 each line instead of with each 50 lines. The frequency components 

 in this case are multiples of 800 cycles and bear the same relations to 

 800 cycles as the components spoken of above bear to 16 cycles. A 

 transient which blurs the picture outline n picture elements (hori- 

 zontally, this time) corresponds to a time interval of n forty 

 thousandths of a second. Evidently the delay requirements are 50 

 times more lenient in the former case than in the latter so that the 

 delay requirement at the highest frequencies, which determine the 

 fine detail in the direction of scanning, is 50 times as severe as at low 

 frequencies, which determine the fine detail in a direction perpendicular 

 to the direction of scanning. 



In the still pictures referred to, the transients extended in the 

 direction of travel of the light spot and there were no transients 

 analogous to those discussed here in connection with Fig. lA. For 

 this reason the delay limits determined from still picture transmission 

 are the ones which apply to the higher frequencies. For the lower 

 frequencies the requirements are obtained by multiplying the high- 

 frequency requirements by 50. For these reasons, together with the 

 result of a Fourier analysis of the picture current, the limits were 

 set at ± 10 or ± 20 microseconds from 400 to 20,000 cycles. Below 

 400 cycles, the departures from the constant delay were permitted to 

 be ± 500 or dz 1000 microseconds. 



Noise. Another important requirement is that relating to the ratio 

 of the picture currents to the extraneous interfering currents which 

 may arise in the line from power induction and other sources. Early 

 experience with the television apparatus showed that considerably 

 more noise was permissible in the case of television than in the case of 

 still picture transmission so that in this case comparison with the still 

 picture transmission would result in an unduly severe requirement. 

 This is thought to be explained by the fact that in the case of television 

 the pictures are flashed before the eye 16 times per second and the 

 effects of the extraneous currents occur on successive flashes in different 

 positions, so that defects of one flash are corrected on the next. 



A set of experiments was performed from which it was determined 

 that if the ratio of average picture currents to average noise currents 

 exceeded about 10 the results were satisfactory. In order to assure 

 considerable margin above this figure, it was decided to make the 

 average television current to be transmitted into the line 4 milli- 

 amperes. 



Echoes. If two paths exist by which the currents may travel from 

 the sending point to the receiving point, the length of the two paths 



