Link to main foregrounds telecons page
Link to logbook
Attendance: Ragnhild, Clem, Mathieu, Shaul
Regrets:
Notes: Shaul
“r non-zero values”:
Raphael suggests r=0.003, and r=5e-4. Why? good for doing cross-checks. Action: SH will check whether there are other prevailing opinions.
Clem reminds us that Hans-Kristian requested some changes to simulations: no aliasing of power in the maps (simulation pixelization was not sufficient for the smallest PICO beams), and that noise be white in the map domain. Clem agrees with the former, not the latter. Action: Need to get Hans-Kristian's input. Clem to write to Hans-Kristian?
Ragnhild reports on her work with Commander1 (
slides here). She observes excess power and is investigating the source.
Clem suggests running the simulations on the 00 model, which has Gaussian distributed foregrounds, to distinguish between the aliasing issues described above and between foreground removal issue. Ragnhild says she believes she had done it some months ago, and all was well with the 00 case. Action: Ragnhild will check.
Brief discussion of Mathieu's priorities: Action: he will check his code and suggest options for improvement, if any.
Mathieu and SH discuss the r likelihood plots (see slides from two weeks ago). Since r=0 in the input, the peak of the likelihood indicates the level of residuals; the width indicates instrument noise, and remaining lensing variance. MR says he fixes A_lens to be 15%, as if we know the value with no uncertainty. MR and SH agree that (1) either A_lens should be fit together with r, or (2) some uncertainty in the value of A_lens needs to be folded in when estimate the likelihood for r.