In trying to set XCP-D parameters, I’m not sure what motion filter order to select, and I’m hoping somone can explain the rationale.
https://xcp-d.readthedocs.io/en/latest/usage.html
For selecting the motion filter order, the ReadTheDocs say “Number of filter coefficients for the motion parameter filter. If the motion filter type is ‘lp’, the order is divided by 2 as filtfilt applies the filter twice. If the motion filter type is ‘notch’, the order is divided by 4 as iirnotch is a second-order filter and filtfilt applies the filter twice. Make sure to set this parameter to a multiple of 2 if you choose the ‘lp’ filter and a multiple of 4 if you choose the ‘notch’ filter.” The default motion filter order is 4.
I’m using a low pass filter (–motion-filter-tyle lp), and I’m not sure if I want to set my motion filter order as 2, 4, or something else. I know using a Butterworth filter, the higher the order, the steeper the transition band. I’m unclear what the ordering options in XCP-D will result in, in terms of the Butterworth filter. If we choose order 4, will two second-order low pass filters cascade together to produce a fourth-order low pass filter (since the ReadTheDocs say the filter is applied twice)? Should we opt for just the second-order filter instead (and is this what we would get if we set the parameter to 2 in XCP-D)? We’re hesitant to use filter order 2 since the rationale says the order is divided by 2, and we want to avoid using a 1st-order filter. We’re leaning toward order 4, but want to understand what’s actually being applied.
If anyone can explain the different motion filter order settings in XCP-D, it would be a great help!