Since the frequency noise is fairly high, this activation of the frequency noise subtraction has a significant impact on the sensitivity as it could be seen in the first figure comparing the h(t) spectrum before and after the change. Due to the current sampling rate of B2_8MHz, this noise subtraction is active only up to 3.5 kHz. This explains the jump of about 7 Mpc in the BNS range (see the second figure, since the update of the transfer function is done every 500 seconds, the improvement is visible only at GPS=1211979000 -> 12:50 UTC).
As a consequence, the coupling between h(t) and PRCL is now much reduced (see third figure). There is still some coupling with MICH (fourth figure), but we have to wait for the new calibration parameters to see if this coupling remains or not.
