COSMOMC: all_l_exact data format

[kumar pandey]

Hi can you please have a look : http://cosmocoffee.info/viewtopic.php?t=1276

thanks,
R.

[Paniez Paykari]

Hi All,

Does the [tex]C_\ell[/tex]s in the planck.dat file have to be necessarily smooth? Or can we have a simulated [tex]C_\ell[/tex] from a CMB map? I appreciate the help.

Thanks
Paniez

[Antony Lewis]

You can use a simulation, just don't expect the posteriors to peak at exactly the simulation input values.

[Paniez Paykari]

Thank you. So, that means I should expect some bias? I can trust the errors though?
Paniez

[Alex Conley]

I'm a bit confused -- why, in the example code, does
one write C_TT+N_l and C_EE+noiseFac*N_l, but
does not include any noise term when writing out
C_TE?

[Antony Lewis]

Just assuming the simple case where there is no noise correlation between Q/U and T.

[Alessio Notari]

Hello,
I am trying to use the all_l_exact format on CosmoMC with the above procedure with some C_l table produced by the Healpix package anafast, and it works well :) .


One simple question: does it work only for full-sky data or should I do something more complicated (such as using planck_like.f90) when I put a mask with anafast on my sky maps?

And by the way, what is the planck_like.f90 routine doing exactly and how do I use it? (I cannot find any instructions on the web) ...

thanks a lot for your help!
Alessio

[Antony Lewis]

Planck_like implements the approximation of 0801.0554 as well as simple schemes like full-sky and Gaussian. You probably should use something that accounts for the covariance when using cut-sky data.

For sample data files see the bottom of
http://cosmologist.info/weightmixer/
or the python exact simulation script supplied with most recent cosmomc.

[Alessio Notari]

Thanks for the answer.

Though, I am not sure if I understand the way to preceed. How do I tell to cosmomc that there is nontrivial covariance due to cut-sky?

Moreover, in the all_l_exact format why is there an f_sky column if this is valid only for full sky?

cheers

[Alessio Notari]

I guess I should probably use the .newdat format for the data, which is related to this:

http://cosmocoffee.info/viewtopic.php?p=4068
and to this:
http://www.astro.caltech.edu/~tjp/CBI/d ... index.html

Could anyone clarify if this is the right way to proceed and how this format can be used for mock data? (there are no answers to the questions in the above link)

thanks again

[Jason Dossett]

Planck_like implements the approximation of 0801.0554 as well as simple schemes like full-sky and Gaussian. You probably should use something that accounts for the covariance when using cut-sky data.

For sample data files see the bottom of
http://cosmologist.info/weightmixer/
or the python exact simulation script supplied with most recent cosmomc.

Hi,
I am trying to use the makePerfectForcast.py python script to create a new dataset for some future simulated data. I understand how to produce the input files:
lensedTotClFileRoot and lensPotential.
I am however unsure how exactly to construct the reconNoise input file. Any help would be greatly appreciated.

Thanks,
Jason

[Antony Lewis]

You only need reconNoise if you want to include CMB lensing reconstruction (in which case it is the reconstruction noise variance, which you can calculate for isotropic full sky case using standard result involving 3j sum).

[Jason Dossett]

Thanks Antony.