Why does the FIRPM function in Filter Design Toolbox 3.2 (R14SP2) return incorrect end-point coefficients when I use it to design an equiripple filter?

I am using the FIRPM function in the Filter Design Toolbox. However, when I use this function the early and late coefficients seem to be incorrect. For example, I use the following code to design my filter:

% All frequency values are in Hz. 

Fs = 768000;  % Sampling Frequency 

N     = 256;    % Order 

Fpass = 21000;  % Passband Frequency 

Fstop = 27000;  % Stopband Frequency 

Wpass = 1;      % Passband Weight 

Wstop = 2;      % Stopband Weight 

dens  = 16;     % Density Factor 

% Calculate the coefficients using the FIRPM function. 

b  = firpm(N, [0 Fpass Fstop Fs/2]/(Fs/2), [1 1 0 0], [Wpass Wstop],{dens}); 

stem(b)

However, the end-point coefficients are very large.

Answer 1

[MathWorks Support Team]

This is expected behavior when using the FIRPM function in the Filter Design Toolbox 3.2 (R14SP2). Large end-point coefficients are common with FIRPM designs due to the equiripple nature of the response.