Generate chi2 vs period graph about presto

[scottransom]

If you have a .pfd file, you can load it in using the prepfold module (into the pfd class). Then you can loop over a range of frequencies using the adjust_fold_frequency() method. That returns the adjusted profiles as well as the reduced chi^2. So then you could plot period (or freq) vs chi^2. Hope that helps.

[jmohta]

Thank you for your response. I tried doing the method you mentioned but I am not able to get exact P vs chi^2 graph as in the PFD file. Can I do something like use adjust_period() function to get the profile at desired period value and then calculate chi^2 using calc_redchi2 function? Would this work?

[scottransom]

That is basically what the adjust_fold_frequency() method does. How different do the plots look? They should be very similar....

But your method should basically work as well. However, if you want or need the exact curve that prepfold puts out in the .pfd plots, I'd recommend just inserting a printf() line in a for loop in the prepfold_plot.c file so that it prints out the values. Then running show_pfd() on the .pfd file.

[jmohta]

In the adjust_fold_frequency() method we need to pass phase bins as one of the parameter to the function. I have only the range of periods. So, in that case what should I pass to the function? I tried once passing directly the period values but I didn't get the desired graph. Then I tried to pass 1/period values to the function then too it doesn't give me the graph. So, how can I convert period values to number of phase bin shift? Thank you for your help.

[scottransom]

Since frequency is (by definition) the derivative of phase wrt time, the delta f (i.e. the change in freq) is equal to delta Phase / T, where Phase is the fraction of a rotation (i.e. the number of phase bins to shift / # of phase bins in the profile) and T is the total observation time. The range of frequencies that prepfold searches is usually +/-2*#of phase bins in the profile.