Correlation of reconstructed lensing template to ideal II
February 18 2019, Clem Pryke
This a one plot update to Correlation_of_reconstructed_lensing_template_to_ideal. Julien Carron provided two new sets of reconstructed lensing templates: i) Recon from 95GHz rather than previous 145GHz. The Q/U white noise noise in the 95/145GHz sim maps is 1.40/2.19 uK-arcmin respectively. ii) Recon from 95GHz signal only maps with Weiner filtering which "knows there is no noise" - to get a measure of the algorithmic limitation of the present Quadratic Estimator method.
The plot below shows the results averaging over the first 10 realizations.
Here are some comments from Julien as to why the 95GHz is not much better than the 145GHz, and why there is the step up at ell=200.
Unfortunately the beam at 95GHz is too large at 6 armin. Were it 1 arcmin the QE tracer quality would be twice larger.
The step comes about because I apply a sharp cut on E and B multipoles ell > 200 before building the lensing tracer. The cross-power to B at ell > 200 can be a bit hard to interpret, since there is contribution from the noise part of the tracer. (Gaussian BB EE contractions in (B x ( phi_noise(E, B) x E_template))