BayesLine issue with flow (and possibly srate)
Looking into the recent PSD concerns, I noticed some weird behavior with changing srate depending on the flow.
Here is flow 8
and here flow 16
The flow 16 ones have the expected behavior as the low f parts of the PSD are consistent regardless of srate. I have confirmed that the (low frequency part of the) data are the same in all runs above.
The following is the AD for flow 8 (crosses) and flow 16 (dots)
showing that the bad PSDs are not due to bad data, just really bad fitting.
Possibly related to that, I see non-integer spline frequency points in the chains
8 8 -85.535 16 -95.4118 54.4516 -99.1214 69.9355 -101.174 104 -99.5692 108 -99.5047 108.645 -100.165 127.875 -96.1224