[Eeglablist] filters

[Stanley Klein]

The main thing I suggest about filters is that in addition to plotting the
frequency and phase response, to also be sure to plot the temporal response
to a pulse input (including negative time). One always wants to see how
much ringing there is. I presume the sharper the filter is in the frequency
domain the worse it is in the time domain.
Stan
On Sun, Jun 10, 2012 at 7:16 AM, Arnaud Delorme <arno at ucsd.edu> wrote:

> Dear Makoto, Clemens and Tim,
>
> in EEGLAB 11, you can plot the FIR and IIR filter frequency and phase
> response.  Looking at the filter response, you can figure out everything
> you need to know
> about the filter. Hope this helps,
>
> Arno
>
> On Jun 6, 2012, at 11:29 PM, Makoto Miyakoshi wrote:
>
> > Thanks Clemens!
> >
> > Makoto
> >
> > 2012/6/6 Clemens Brunner <clbrunner at ucsd.edu>:
> >> Makoto, it's the other way round for me -- I'm not an expert in analog
> >> filter design, but I do know a little bit about signal processing :-).
> Yes,
> >> your intuition is correct. You can still apply this to digital filters
> --
> >> IIR filters are typically designed using analog design technique and
> then
> >> transferring that to the digital domain. FIR filters are usually
> >> digital-only, so I'm not sure if these properties apply (of course, if
> you
> >> increase the filter order of an FIR filter, the cutoff will be steeper,
> but
> >> I cannot tell you if there is a way to determine the slope).
> >>
> >> There is a good introductory paper by the Mathworks, which talks about
> >> filters and has MATLAB examples:
> >> http://www.mathworks.com/tagteam/55876_digfilt.pdf
> >>
> >> HTH
> >> Clemens
> >>
> >>
> >>
> >> On 06/06/2012 11:28 AM, Makoto Miyakoshi wrote:
> >>>
> >>> Dear Tim and Clemens,
> >>>
> >>> Thanks for feedback.
> >>> Since I don't know much about signal processing, my understanding of
> >>> filter is the analog one used in an electric circuit with capacitors
> >>> and inductors (as used in speakers and amplifiers). Sor it is
> >>> intuitive for me, as Tim said, that slope and filter order are such
> >>> that 6dB for the first order (1 capacitor/inductor), 12dB for the
> >>> second order (capacitor + inductor), 18dB for the third order (3
> >>> elements), etc. But it seems filter can be 'designed' pretty freely in
> >>> signal processing.
> >>>
> >>> Clemens, I would appreciate if you could show us an introductory
> >>> website or paper to learn what is digital filter.
> >>>
> >>> Makoto
> >>>
> >>>
> >>>
> >>> 2012/6/6 Tim Schoof<t.schoof at ucl.ac.uk>:
> >>>>
> >>>> Hi Makoto,
> >>>>
> >>>> I got the following response from Clemens Brunner:
> >>>>
> >>>> Hi Tim,
> >>>>
> >>>>> I was wondering whether anyone knew the exact filter specifications
> of
> >>>>
> >>>> the
> >>>>>
> >>>>> Basic FIR and Short IIR filters implemented in eeglab.
> >>>>
> >>>>
> >>>> eegfilt() uses firls() or fir1() to compute the filter coefficients
> for a
> >>>> given order.
> >>>>
> >>>> iirfilt() designs an elliptic filter and automatically selects the
> order
> >>>> to meet the
> >>>> filter specs (high cutoff, low cutoff, passband ripple and stopband
> >>>> attenuation).
> >>>>
> >>>>> Also - and this is probably a question about filters in general - I
> was
> >>>>
> >>>> hoping someone could tell me how to specify a 24 dB/oct roll-off (or I
> >>>> suppose 12 dB/oct given that the script uses filtfilt). What order
> would
> >>>> that correspond to in an FIR filter?
> >>>>
> >>>> Depends on which filter design technique you use. If you want to go
> the
> >>>> manual way,
> >>>> first pick a design technique (windowing, Kaiser window, optimum
> filters
> >>>> via
> >>>> Parks-McClellan, ...), provide the desired specifications, and the
> >>>> corresponding
> >>>> parameters (including filter order) will pop out.
> >>>>
> >>>> You can play around with MATLAB's filter design tools if you have
> MATLAB
> >>>> and the
> >>>> Signal Processing Toolbox. You can use fdesign or filterbuilder, or
> >>>> fdatool. This
> >>>> will let you fully customize your filter.
> >>>>
> >>>> If you want to use such a custom FIR filter with EEGLAB, I guess you
> >>>> would
> >>>> have to
> >>>> write a short plugin or filter outside of EEGLAB (i.e. working with
> >>>> EEG.data
> >>>> directly).
> >>>>
> >>>> HTH
> >>>> Clemens
> >>>>
> >>>> Tim
> >>>>
> >>>>> Dear Tim,
> >>>>>
> >>>>> I'm also interested in knowing this. Arno?
> >>>>>
> >>>>> Makoto
> >>>>>
> >>>>> 2012/6/4 Tim Schoof<t.schoof at ucl.ac.uk>:
> >>>>>>
> >>>>>> Dear eeglab list,
> >>>>>> I was wondering whether anyone knew the exact filter specifications
> of
> >>>>>> the
> >>>>>> Basic FIR and Short IIR filters implemented in eeglab.
> >>>>>> Also - and this is probably a question about filters in general - I
> was
> >>>>
> >>>> hoping someone could tell me how to specify a 24 dB/oct roll-off (or I
> >>>> suppose 12 dB/oct given that the script uses filtfilt). What order
> >>>> would
> >>>>>>
> >>>>>> that correspond to in an FIR filter?
> >>>>>> Many thanks,
> >>>>>> Tim
> >>>>>> _______________________________________________
> >>>>>> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> >>>>>> To unsubscribe, send an empty email to
> >>>>>> eeglablist-unsubscribe at sccn.ucsd.edu
> >>>>>> For digest mode, send an email with the subject "set digest mime" to
> >>>>
> >>>> eeglablist-request at sccn.ucsd.edu
> >>>>>
> >>>>>
> >>>>>
> >>>>>
> >>>>> --
> >>>>> Makoto Miyakoshi
> >>>>> JSPS Postdoctral Fellow for Research Abroad
> >>>>> Swartz Center for Computational Neuroscience
> >>>>> Institute for Neural Computation, University of California San Diego
> >>>>>
> >>>>
> >>>>
> >>>>
> >>>
> >>>
> >>>
> >>
> >
> >
> >
> > --
> > Makoto Miyakoshi
> > JSPS Postdoctral Fellow for Research Abroad
> > Swartz Center for Computational Neuroscience
> > Institute for Neural Computation, University of California San Diego
> > _______________________________________________
> > Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> > To unsubscribe, send an empty email to
> eeglablist-unsubscribe at sccn.ucsd.edu
> > For digest mode, send an email with the subject "set digest mime" to
> eeglablist-request at sccn.ucsd.edu
>
>
> _______________________________________________
> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> To unsubscribe, send an empty email to
> eeglablist-unsubscribe at sccn.ucsd.edu
> For digest mode, send an email with the subject "set digest mime" to
> eeglablist-request at sccn.ucsd.edu
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20120610/fa01384c/attachment.html>