Reproduce PySM 2 templates for Dust

[zonca]

I am trying to reproduce the processing of the Commander Dust map into the PySM templates for Dust, as described in the PySM 2 paper: https://arxiv.org/pdf/1608.02841.pdf

Something looks wrong, in particular the BB spectrum I get is too flat.
Any hints?
You can comment directly line-by-line on the notebook here: https://app.reviewnb.com/healpy/pysm/pull/66/files/

Or just view a read-only version of the notebook here: https://nbviewer.jupyter.org/github/healpy/pysm/blob/37eb60355979b49a90405dab8624c830fbb64bd7/docs/template_preprocessing/reproduce_pysm2_dust.ipynb

@NicolettaK @bthorne93 @brandonshensley @jodunkley

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[zonca]

I tried to compute the spectrum with namaster instead of PolSpice, using all default options, just passing the mask with no apodization:

f_2 = nmt.NmtField(total_mask, m_planck[1:])
binning = nmt.NmtBin.from_nside_linear(nside, 1)
cl_22 = nmt.compute_full_master(f_2, f_2, binning)

but I pretty much get the same results:

the BB spectrum is very flat, the gamma fit is -0.0018

[zonca]

thanks @bthorne93 for suggesting to denoise the spectra,
I made a rough estimate of the white noise level in EE and BB and now the agreement with the PySM 2 paper is reasonable.
I'm merging this into the documentation of PySM 3 for future reference.

Plot from the PySM 2 paper

Plot from my implementation

[zonca]

https://pysm3.readthedocs.io/en/latest/template_preprocessing/reproduce_pysm2_dust_pol.html