426 FLICKER WITH MOVING PICTURES [Cn. XI 



THEORY AND EXPERIMENTS ON FLICKER 



592. Experiments have been made to determine the speed at 

 which flicker disappears, that is, the speed at which the eye is un- 

 able to distinguish between a continuous and an intermittent 

 light. 



These experiments show that at a certain speed the appearance 

 of flicker disappears. This speed is practically the same for 

 different people. 



As the speed is increased the flicker disappears for the center of 

 the field of vision before it does for the edge. Thus, the light on a 

 screen may not appear to flicker when looked at directly but it may 

 appear to flicker when looked at out of the "tail of the eye." 



As the brightness of illumination is increased the appearance of 

 flicker is increased and a higher speed is required for flicker to 

 disappear. Thus, when showing a very dark film, the light may 

 not appear to flicker at all, while with a very transparent film or 

 no film at all the light may appear to flicker violently although the 

 speed is the same. 



If, instead of having the dark period and the light period equal, 

 either the dark period or the light period is made less in proportion 

 the flicker appears less violent, and it disappears entirely at a lower 

 speed. This effect is, however, not very great. 



Thus, the flicker with a shutter in which l /6 is light and % 

 is dark, is the same as one in which % is light and Yt> is dark 

 ( SQ2a). 



592a. A formula to express these factors numerically was worked out by 

 T. C. Porter of Eton College and published in the Proceedings of the Royal 

 Society,, Vol. 63, p. 347; Vol. 70, p. 313-329 (1902). 



The constants have been recalculated. 



Let f = number of light flashes per second at which flicker disappears when 

 light and dark flashes are equal. 



Let n = number of flashes per second; light and dark flashes are unequal. 



w = angle of white space in disc. 



(360 w) = angle of dark space in disc. 



I = intensity of illumination in meter candles. 



b = constant depending on illumination. 



From experimental data the formula comes out 



f = 26 -f- 12.2 log I 



b = 12.04 + 2 -3?8 log I 



n = f + b [logw log (360 w) 4.5106]. 



