One problem, typical for the data translation framegrabber board used, is that every now and then only the odd or even field of a frame is passed to the program.
This will show up in the raw curve as a double
image of the curve at half intensity. For some events, like the 28
Sgr/Titan event of 1989, this poses a problem (figure 7). In such
cases you can use the background measurement to distinguish
between good and bad points. Simply take the difference between the
two embracing windows. Figure 8 shows this difference. As you can
see in this figure some 
of the original signal is still present
in the background.
Thus, this background can also be used to judge if window sizes are chosen well. It is relatively easy to define a filter to reject the bad points. The final result, using an extra moving average over 2 seconds, is shown in figure 9.
This result concerns a video recording made by Hans Bode in Germany using an image intensifier.The tracking of the portable Celestron he used was quite bad. Titan was visible throughout the event, though, making it easy to use it as the object to be tracked, using the method described in the previous chapter.
