Error message for an invalid angle of incidence (alpha) for a beam splitter
didn't match the valid (and checked) range. Fixes #27

Removing sqrt(2) factor that wasn't needed in qnoised detector. Caused additional noise when using qnoised with only vacuum contributions which had non-unique frequency bins.

Changing conjugates back, needs more testing. Changed qshot demod phase too so it doesn't change its output when varying it in tandem with a modulators phase.

The value of `alpha_2` was not being updated when sweeping a
beamsplitter's `alpha` parameter with HOMs disabled, so now a manual
check is inserted.

Adding quiet option to lock output, adding in --no-log option to not create a log file. Cluster runs seem to be getting disk IO limited