PSRCHIVE user documentation: pat

1.0 Purpose

pat is used to determine the time of arrival (TOA) of each integrated Profile in an Archive. It can use a variety of methods to determine the phase shift between each Profile and a standard template. This shift is added to the reference epoch of the Integration in which the Profile resides to give the representative site arrival time. These TOAs can then be analysed with TEMPO or TEMPO2.

2.0 Usage

pat is very easy to use, all commands are of the form
pat options filenames
where the options allow the user to specify a standard template profile, the fitting algorithm to use, the output format and several other small configuration parameters. For a full list of options, use
pat -h
. Multiple filenames can be specified on the command line either individually or using wildcard characters.

3.0 Algorithms

At the present time, pat can use any one of five algorithms to determine the phase shift between the standard template and the observed Profile.
  • Fourier Phase Gradient (PGS) This algorithm was described By J. Taylor in the Philosophical Transactions of the Royal Society of London A, 1992, 341, 117. It takes advantage of a property of the Fourier transform known as the "shift theorem", which states that the Fourier transform of a function and a shifted copy of the function differ only by a linear phase gradient. Fitting for this gradient in the Fourier domain can determine the shift between two similar Profiles. This algorithm is very precise when the S/N of the Profile is high.
  • Gaussian Interpolation Shift (GIS) This algorithm was described by Hotan, Bailes and Ord in the Monthly Notices of the Royal Astronomical Society, 362, 1267. pat calculates the discrete cross correlation function of the Profile with the template and a Gaussian curve is then fitted to the resulting points to allow interpolation between each phase bin. In this manner, TOAs can be determined to within approximately 1/10 of the width of an individual phase bin. This algorithm is less susceptible to noise contamination than the PGS method, but is less precise when the S/N is high.
  • Parabolic Interpolation Shift (PIS) This is the oldest method available. It is very similar to the GIS method, but uses only the peak bin of the cross correlation function and one bin on either side to define a parabola that is used for interpolation.
  • Zero Pad Fitting (ZPF)This method interpolates the cross correlation function by Fourier transforming, padding the result with zeroes and transforming back to the time domain. It is somewhat experimental and the error estimate it returns is not reliable.
  • Sinc Interpolation Shift (SIS) This algorithm is similar in effect to the ZPF method.
All of the above methods fit only the Stokes I profile. The user can also choose to fit the full polarimetric profile in the Fourier domain using an algorithm described by van Straten in the Astrophysical Journal (in press). Normally, the standard template profile is loaded from a Pulsar::Archive, but it is also possible to use an analytic standard template constructed from Gaussian components. pat is compatible with multiple TOA output formats, including parkes, itoa, princeton and the more modern tempo2 format. Additional flags can be added to the default output when using the tempo2 format, allowing the user to carry extra information (like the name of the instrument used to record the data) along with the TOAs.

4.0 Testing and examples

Typically, pat will be invoked in the following fashion:
pat -A GIS -s std.rf *.rf 
The above command prints a formatted list containing the TOAs and associated information like frequency and uncertainty. If no algorithm is specified (with the -A option), the PGS method is used.

5.0 Known bugs and features that require implementation

  • None at this time.