understanding filter properties

Hello,

I can’t figure out how to interpret the representation of the filters in the time domain.
When I plot a default FIR vs IIR band-pass filters for my data, the impulse response is centered over very different times.
My question: what does the x position of the impulse’s peak mean and what are the consequences of this difference?
Same, although less worrisome, for the discrepancy in amplitude: does it point to anything?

Thank you so so much,

Giulia

tagging @larsoner on this one. Question about FIR vs IIR filtering.

Depends on how you actually apply the filter. If you use phase='zero', the time axis in the FIR plot is not correct – it should go from about -0.25 to about +0.25.

For the amplitude, the frequency response plots should tell you this.

I would recommend reading the backround info on filtering tutorial, it tries to help with this stuff a bit:

https://mne.tools/dev/auto_tutorials/preprocessing/25_background_filtering.html

1 Like