When should noisepower divide by 2.0?

[zuoshifan]

I notice that noisepower divides by 2.0 in line 819 of timestream.py, but doesn't in elsewhere, e.g. line 388, 828, 1401, 1496 of beamtransfer.py and line 286 of kltransform.py. Is there any inconsistency in these cases?

I also notice the noisepower in unpolarized telescope is half of that in polarized telescope, see line 900 of telescope.py.