noise questions

[lbaumo]

1. The noise for simulations seems large.
2. are we following the correct procedure by adding noise to the simulation then dividing by the noise? should we divide by a noise map of gaussian noise with a given standard deviation, like we do for the data, or should we just divide by the standard deviation, as appears in virginia's paper
3. There is a difference of 1/sqrt(2) in the noise in Emma's paper and virginia's paper. (see Eq 3 vs Eq 8) Why?

[lbaumo]

for the third point- the factor of 2 in emma's paper is consistent with Eq 8 in the SLICS paper. Changing it to virginia's convention does not correct our sqrt(2) problem, it makes things worse (confirmed by @AndreasTersenov)

[martinkilbinger]

2. Sounds like the correct procedure.
   I guess dividing by the std is the simplest option. Probably more accurate is to divide by a noise map, to capture local variations in galaxy density. If I understand this point correctly.
3. The dispersion <sigma_eps^2> is also different, in our case it is for the complex shear (two components), sigma_eps = 0.44. Since kappa is a scalar (one component) we divide by 2. Virginia's <sigma_eps> is for one component, 0.3.
   So 0.44 ** 2 / 2 is very close to 0.3 ** 2.