D. Benesch

In my opinion, it would make more sense to set a default sampling frequency for the interbeat interval interpolation independent of the sampling frequency of the raw heartbeat signal, for...

> We could add detrending too though That's true! Also, in case helpful here is a notebook showing the power for different interpolation rates: https://github.com/danibene/hrv_freq_interpl/blob/main/neurokit_hrv_freq_interpl.ipynb

Honestly, I'm not sure how to evaluate what is better here - I've seen a lot of variability in the literature ([including not interpolating/resampling at all and using the original...

Hi @Tam-Pham , thanks for the feedback! I opened a PR with your suggested changes: https://github.com/neuropsychology/NeuroKit/pull/680 I defined `x_new` a bit differently than you suggested: > `x_new = np.arrange(1, desired_length,...

@Tam-Pham > (please excuse my midnight brain) Relatable :D #### Re detrending: > I think Kubios implemented a more complicated detrend method than that. The Kubios [user guide](https://www.kubios.com/downloads/Kubios_HRV_Users_Guide.pdf) (pg. 12)...

PS: > > Should we also add an option to not interpolate for the the lomb-scargle periodogram? (And if so, should this be a separate PR?) > Off the top...

@DominiqueMakowski I still would like to make it possible to not interpolate the RRI when using the lomb-scargle periodogram method, but will address after https://github.com/neuropsychology/NeuroKit/issues/684

> guess the stale bot _can_ be useful who knew ^^ @DominiqueMakowski I'm happy to submit a PR for this, checking whether the `peaks` argument in `hrv()` is a dict...

> If the sampling rate is too low / high to achieve a given bandpass filtering we should just apply a highcut / lowcut (depending on what's possible) no? also...

Actually, I don't understand how a bandstop filter is supposed to be indicated with the current implementation of `signal_filter`, since `lowcut` shouldn't be higher than `highcut` for the underlying default...