The README states the following:
"c_footprint_file: (optional) a footprint file containing the mask power spectrum, which is read-in in C_survey_window in cosmolike_core/theory/covariances_3D.c; note that the normalization of the power spectrum is automatically adjusted,"
However, changing the c_footprint_file input normalization causes a roughly linear change in the covariances produced. I attach an example of two different DES-like mask files, one of which is normalized, and another which is not. Running the same covariance generation with the two causes different results in each case, i.e. the normalization of the power spectrum is not automatically adjusted and must be done before input.
Example mask files (with incorrect normalization, and with correct normalization):
mask_Cl_DES.txt
mask_Cl_DES_norm.txt
Using the incorrect mask causes covariances to be underestimated by a factor of ~8. I am also unable to find an instance of any automatic adjustment in the code, but open to clarifications on the documentation as I have not dug into it in great detail.
The README states the following:
"
c_footprint_file: (optional) a footprint file containing the mask power spectrum, which is read-in inC_survey_windowincosmolike_core/theory/covariances_3D.c; note that the normalization of the power spectrum is automatically adjusted,"However, changing the
c_footprint_fileinput normalization causes a roughly linear change in the covariances produced. I attach an example of two different DES-like mask files, one of which is normalized, and another which is not. Running the same covariance generation with the two causes different results in each case, i.e. the normalization of the power spectrum is not automatically adjusted and must be done before input.Example mask files (with incorrect normalization, and with correct normalization):
mask_Cl_DES.txt
mask_Cl_DES_norm.txt
Using the incorrect mask causes covariances to be underestimated by a factor of ~8. I am also unable to find an instance of any automatic adjustment in the code, but open to clarifications on the documentation as I have not dug into it in great detail.