The effect of decimating on TDEV data

 In this section, results will be presented verifying that reducing the SPW output data file from 30Hz sampling frequency to 3Hz has a minimal effect on the TDEV data subsequently calculated. This allowed the TDEV processing time as well as the hard drive space utilised to be reduced significantly without inordinately affecting the quality of the results. Shown in figure 8.25 are plots of TDEV produced from 30Hz data (the solid line) as well as 30Hz data decimated to 3Hz (the circles). It can be seen that the TDEV data produced from the decimated data is essentially the same except at short observation intervals. This is fortuitous in that it is the calculation of data over long observation intervals that makes the largest demands on processing time and file storage. Thus it was determined that it was possible to decimate the output file to produce TDEV data for observation intervals $\tau\geq 10^0s$. It was appropriate to use un-decimated data for observation intervals smaller than 1 second but this placed minimal demands on processing time and file storage and so was satisfactory.

 

Figure 8.25:   Plot verifying that the use of the SPW decimate has a minimal effect on TDEV results