CLASS 40 GHz 2016 - 2022 Observations (DR1)
Mapping Transfer Function
Time-stream filtering of CLASS data in the mapmaking pipeline removes large scale power from the maps and biases power spectra estimation. A harmonic domain mapping transfer function is used to correct for this spectrum bias. The function takes the form of a matrix Fℓℓ' such that
Cℓ,out = Fℓℓ' Cℓ',in,where Cℓ',in refers to the VV, EE, BB angular power spectra of the unfiltered sky. The transfer function Fℓℓ' relates these to the angular power spectrum as observed by CLASS, Cℓ, out, accounting for correlations between multipole moments ℓ and between spectra. When used with a pseudo-Cℓ estimator, the recovered estimate of the input spectra have been shown through simulations to be unbiased for ℓ > 4. The transfer function defined here includes both the low-pass filtering in the demodulation process (at higher-ℓ) and the filtering from the map-making process (low-ℓ), i.e. the pink and green curves shown in Figure 14 of Li et al. (2023).
Two forms of the transfer function are provided. One form assumes a uniform weight within the CLASS survey region, and is useful when simulating the impact of the CLASS filtering strategy on the power spectrum of other experiments. The other form assumes the non-uniform CLASS weight is used, a modification appropriate when correcting for the transfer function bias on CLASS cross spectra.
The Fℓℓ' matrix multipole range is 0 - 383 inclusive. There are 9 384 x 384 blocks in a 3x3 configuration, for a total matrix size of 1152 x 1152. Ordering is VV, EE, BB on the diagonals. The matrix is written in hdf5 format, which may be read in python using the h5py package.
CLASS Data Pipeline and Maps for 40 GHz Observations through 2022, Li et al. (2023)
CLASS Angular Power Spectra and Map-Component Analysis for 40 GHz Observations through 2022, Eimer et al. (2023)