FIR vs Butterworth

[affective.n...@gmail.com]

Hi Mike,

Thanks for your response to the previous question. Just wondering - are there any known issues with using a Butterworth IIR filter compared to say an FIR zero-phase shift filter constructed using the MATLAB firls function and the filtfilt command? From fig 14.10 it does not seem to me that the Butterworth introduces any visible phase distortions when compared to the FIR...is this usually a safe assumption?

Thanks!

[Mike X Cohen]

Dear affective.neurodynamcs,

IIR filters can introduce phase instabilities, and FIR filters are generally more well-behaved (although some parameter combinations can also produce poor FIR filters, see figure 14.7). It's true that in the example in figure 14.10 there appears to be no appreciable difference between the FIR and the Butterworth, but I wouldn't make many strong general claims based on a single trial. 

Technically, IIR filters are faster to implement, but on modern computers, that really isn't a major issue unless you are doing on-line filtering (e.g., for BCI or neurofeedback). If you are focusing on ERPs or TF power, it probably doesn't matter whether you use IIR or FIR. If you are focusing on phase-based analyses, I'd stick with FIR. 

Mike

--
You received this message because you are subscribed to the Google Groups "AnalyzingNeuralTimeSeriesData" group.
To unsubscribe from this group and stop receiving emails from it, send an email to analyzingneuraltimes...@googlegroups.com.
Visit this group at http://groups.google.com/group/analyzingneuraltimeseriesdata.
For more options, visit https://groups.google.com/groups/opt_out.

--

Mike X Cohen, PhD
mikexcohen.com