Thank you so much for performing the silent run!!
Unfortunately, from the viewpoint of coherence with IMMT1 trans, it seems DARM has higher coherence with IMMT1 trans (see Fig. 1) than that of Mar 30 morning run. (for comparison, see 24634)
In fact I saw the REFL camera showed TEM01-like vertically devided pattern (Fig. 2 taken at 10:18 JST; see some general questions down to words in 24600), which should not happen when the alignment to PRC would be better.
Anyway, using the measured open-loop transfer function (24635), we would say more definite thing about the sensitivity. Hope more simple way to apply each measured OLTF to reproduce the sensitivity.
Anyway, having a look at summary page for distribution of each frequency range to the total BNS range, I'm repeating this, it would be more important to focus less on un-reduce-able peaks but rather more on reducable noise floor.