R devel - Sep 2008 - package "signal", function "filter" (PR#12752)

Full_Name: Jeff Schenck
Version: 2.6.2 (signal v. 0.5)
OS: Windows XP
Submission from: (NULL) (128.107.248.220)


The "filter" function in the package "signal" does not
properly handle recursive
filters with only two feedback taps.  I looked at the source code in filter.R. 
There is a line that says "if (length(a) > 2)" that I think should
have ">=".  I
tried it that way and it seems to work fine.

FYI...  My test filter was a DC notch:
  b <- c(1, -1)
  a <- c(1, -0.95)
  y <- filter(b, a, x)

Thanks,
Jeff Schenck

Package 'signal' is not the responsibility of the R developers, so you 
need to follow the FAQ and report this to the maintainer, rather than 
clog up R-bgs with an inappopriate report.

You might find that R's own function filter() is better written.

The FAQ also asks you not to report on long obsolete versions of R.

On Tue, 9 Sep 2008, jschenck at cisco.com wrote:
> Full_Name: Jeff Schenck
> Version: 2.6.2 (signal v. 0.5)
> OS: Windows XP
> Submission from: (NULL) (128.107.248.220)
>
>
> The "filter" function in the package "signal" does not
properly handle recursive
> filters with only two feedback taps.  I looked at the source code in
filter.R.
> There is a line that says "if (length(a) > 2)" that I think
should have ">=".  I
> tried it that way and it seems to work fine.
>
> FYI...  My test filter was a DC notch:
>  b <- c(1, -1)
>  a <- c(1, -0.95)
>  y <- filter(b, a, x)
>
> Thanks,
> Jeff Schenck
>
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>
-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595