We here provide quick tips to get started. Two example scripts are provided: IDL/poker_example.pro and IDL/check_ipoker.pro. The former proposes a basic example of power spectrum estimation, the latter proposes an example of cross-power spectrum between different maps, together with the Monte-Carlo simulation scheme to derive error bars.

In a nutshell:

• ipoker.pro can be called as simply as:
  ipoker, map, pixel_resolution_in_arcmin, k, pk.
  In this case, there is no mask or boundary effect correction, ipoker just bins the squared modulus of the Fourier transform of the map.
• To correct for boundary effects, give a value to nx_large and ny_large larger than the map x and y sizes, e. g. if the map is 100x100 pixels:
  ipoker, map, pixel_resolution_in_arcmin, k, pk, nx_large=128, ny_large=128
• To account for a mask:
  ipoker, map, pixel_resolution_in_arcmin, k, pk, nx_large=128, ny_large=128, mask=mask
• The binning width can be adjusted in the command line:
  ipoker, map, pixel_resolution_in_arcmin, k, pk, nx_large=128, ny_large=128, mask=mask, delta_l_over_l=0.2, dk_min=100
• Quantities such as the mode mixing matrix and the binning parameters may be common to different maps and do not need to be recomputed. They are output by ipoker.pro when it is called the first time, in the structure out_arrays, together with the relevant parameters in another structure out_params. These can then be passed as input to the next calls of ipoker:
  - First call: ipoker, map, pixel_resolution_in_arcmin, k, pk, nx_large=128, ny_large=128, mask=mask, out_arrays=arrays, out_params=params
  - Next calls: ipoker, new_map, k, pk_new_map, in_arrays=arrays, in_params=params
• The calling sequences are the same for the estimation of the cross-power spectrum of two maps. Just pass the second map as an additional keyword:
  ipoker, map, k, pk_new_map, map1=the_other_map
• map and map1 may have been observed with different point spread functions (e.g. in Planck-HFI when the maps were those of two different bands). These beams can be passed to ipoker via the keywords a_beam and b_beam. See IDL/check_ipoker.pro for more details and a worked example.

IMPORTANT NOTE: POKER computes binned quantities pk for a bin refered to as k. The user must be careful when it comes to comparison with theoretical models and parameter estimation. Indeed, Eq. (14) of the paper is an equality only in the case case of a known power law. Depending on how the "continuous" spectrum varies accross the bin, the binned value will be more or less close to the average value in the bin or to the center value of the bin. There is no way to avoid this. In the (rather common) case of a power law, the continuous value at the center of the bin equals pk/k^beta, which is why ipoker.pro outputs this for convenience. But the user should keep in mind (and therefore check on simulations) that this is only an approximation and that the only exact quantity is the binned pb_out output by ipoker.pro.