Re: Error using fir1()

[Mike Miller]

On Thu, Jun 6, 2013 at 17:03:25 +0800, Matthew Tay Han Yang wrote:
> Hi,
> I am trying to create a filter, N_order = 8000, and cutoff frequency 0.01.
> Thus I type: b = fir1(8000, 0.01).
>
> However, the error message is " error: fir2 : grid size must be greater than
> half the filter order".

Hi Matt, I can confirm this, can you please report a bug against fir1 at:

https://savannah.gnu.org/bugs/?func=additem&group=octave

> In this case, I read from
> http://octave-matcompat.sourcearchive.com/documentation/20010225-7/fir1_8m-source.html
> that it uses fir2 instead of fir1 by default, since the window is not
> specified.
>
> However, what does "grid size" mean, and how should I overcome this?

See "help fir2" for an explanation of the grid size argument.

Until this is fixed, you could adapt your filter requirements to call
fir2 directly. For your example above, that would be

  b = fir2 (8000, [0, 0.1, 0.1, 1], [1, 1, 0, 0]);

HTH,

-- 
mike