Closed aarontabor closed 4 years ago
This turned out to be a calculation error when determining the number of readings to store in the rolling window, which caused the rolling window to actually be much longer than expected - in turn taking a long time to settle in.
Fixed in 729829677a53988353f6084e4b5a26ce3b8c19ab
This is related, but independent from #4. Since #4 is still an issue, leaving it open for now.
Overview
I'm currently computing normalized readings based on the raw data buffered over a 15-second rolling window. Therefore, I would expect the normalized readings to hone-in and basically oscillate around 0 after a 15-second start-up period, but this doesn't seem to be the case.
What I'm observing now is almost another full 15-seconds elapses while the normalized signal stays above 0, before an abrupt DC shift when it finally settles in as expected.
Tasks