Hi,
I've tried to use high-pass filter in MLC in MEMS Studio, but the results are incorrect for first windows in each data file. I found that result is different from the one from Unico, but probably because Unico just drops first 2 windows from each file (which is an error itself).
Signal is a quite constant noise (sample data and project is attached).
These are F2_MEAN_ACC_Z_FILTER_1 & F3_MEAN_ACC_V_FILTER_2 from MEMS Studio and Unico.
Decision tree generated based on these filters makes no sense.
Values of these features read from FIFO looks ok. Looks like a bug in the app.
Best regards
Tomasz
