Another problem encountered was that there were several kinds of video noise, that caused error in estimating position of the ball. First, vertical line noise occurred all the time in the rightmost of the screen. Second, pixel (15,4) or (15,0) had a white value when read from memory, but didn't show up on the screen. Third, there could be noise caused by external force applied to the pinball machine, for example, using the plunger to put the ball into play.
For the first and second kinds of noise, it was hard to figure out what caused the problem. It seems to be something wrong in the hardware. Those two cases were simply solved by ignoring the affected pixels. To avoid the first kind of vertical noise on the screen, just pixel columns 0 through 62 were read. That is, the procedure which gets the position of the ball does not read the rightmost pixel on each line (63). For the second kind of noise, it also ignores the pixel in the two positions that were observed to be noisy, i.e. (15,4) and (15,0). No compensation was attempted for noise due to external forces. A filtering method might be useed, but there was no time to try this.