You could alternatively enter the sampling frequency and view results in Hz
    b=[1 1 1 1];
    a=1;
    freqz(b,a,256,1000);

OR, if you really want normalized

    freqz(b,a,256,1);
    axs=(findobj(gcf,'type','axes'));
    for idx=1:2;
         xlabel(axs(idx),'Normalized Freq');
    end

another usefull tool is fvtool

fvtool(b,a,256,1);

now right hand click on xlabel and change to what you want



<aine_canby@yahoo.com> wrote in message
news:1120066800.034628.194360@g44g2000cwa.googlegroups.com...
> Lets say I have a filter with a cutoff at 100Hz and my sampling rate is
> 1kHz. This means my cutoff is at 0.1 in normalized frequency terms. But
> it seems that the matlab freqz function plots my cutoff as being at
> 0.2, because it defines normalized as being relative to the nyquist. Is
> this not really confusing?
>
> If I'm referencing the freqz graph, I guess I'll need to say that the
> cutoff has a normalised frequency of 0.2 (with respect to the nyquist).
> Would that be clear?
>

aine_canby@yahoo.com wrote:

> Lets say I have a filter with a cutoff at 100Hz and my sampling rate is
> 1kHz. This means my cutoff is at 0.1 in normalized frequency terms. But
> it seems that the matlab freqz function plots my cutoff as being at
> 0.2, because it defines normalized as being relative to the nyquist. Is
> this not really confusing?

Well, obviously the folks at The Mathworks (tm) liked it.  I would 
prefer something relative to sampling rate.
> 
> If I'm referencing the freqz graph, I guess I'll need to say that the
> cutoff has a normalised frequency of 0.2 (with respect to the nyquist).
> Would that be clear?
> 
As clear as you can get.  You could also see if you can't plot a graph 
against actual frequency.

-- 
-------------------------------------------
Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Lets say I have a filter with a cutoff at 100Hz and my sampling rate is
1kHz. This means my cutoff is at 0.1 in normalized frequency terms. But
it seems that the matlab freqz function plots my cutoff as being at
0.2, because it defines normalized as being relative to the nyquist. Is
this not really confusing?

If I'm referencing the freqz graph, I guess I'll need to say that the
cutoff has a normalised frequency of 0.2 (with respect to the nyquist).
Would that be clear?