Even more track correction stuff...

At a recent local meeting, it was suggested that using correlations of 1/pt may be better than correlations of pt. This page shows my attempts to correct the FF spectra using the modified RFF 1/pt Vs unmodified RFF 1/pt correlation matrix.

Figure 1: This firgure shows the modified RFF vs unmodified RFF 1/pt correlation. The top plot shows tracks which originally had a positive charge sign. The events at negative Y are positive tracks that have had their sign changed. The bottom plot shows tracks which originally had a negative charge sign.

The 1/pt correlations are much tighter than the pt correlations. Before we can use the correlations shown in figure 1 to correct the FF spectra, we need to determine where the true negative distribution ends and the negative tracks which were originally positive tracks begins.

Figure 2: This plot shows the modified RFF q/pt spectra for tracks which originally had a positive charge sign (Red) and tracks which originally had a negative charge sign (Blue). The left plot shows the full range and the right plot is a blow up of the +/- overlap region.

The above figure shows the overlap between the modified negative RFF spectra and the modified positive RFF spectra. Looking at the figure, I have decided to treat all negative charge sign tracks wich have 1/pt < 0.049 (pt > 20.408) as positive tracks which have changed sign.

Figure 3: This figure compares the unmodified FF (Black), unmodified RFF (Blue), and modified FF (Red) spectra. The FF spectra is modified by using gausian fits to the correlation matrices shown in Figure 1. The top plot shows positive charge sign spectra and the bottom plot shows the negatvie charge sign spectra.