Re: A question regarding unwrap

[Doug Stewart]

Joe Koski wrote:

> on 9/24/05 7:44 PM, Doug Stewart at address@hidden wrote:
>
>  
>
> > Ron Crummett wrote:
> >
> >    
> >
> > > Hi -
> > >
> > > I have a recent problem with unwrap; I tried plotting the phase of a
> > > transfer function and it jumped halfway through the plot, from -180 to
> > > 180.  I tried to unwrap the phase using unwrap and plotted the phase
> > > again, but still had the same jump.  Reading the help on unwrap, it
> > > sounds like it will only really work if the phase is positive (it adds
> > > 2*pi).  For me, I have a negative phase, so I need to subtract 2*pi.
> > > Any idea of what the solution could be?  I tried looking at the code
> > > but couldn't make too much sense out of it from my few glances.
> > >
> > > Thanks,
> > >
> > > Ron Crummett
> > >
> > >      
> >
> > # fix the phase wrap arownd
> > ll=length(ph);
> > [qm iqm]=max(abs(diff(ph)));
> > if(qm>300)
> > ph(iqm+1:ll)=ph(iqm+1:ll)-360;
> > endif
> >
> >    
> Ron, Doug,
>
> My experience with an alternative (Stearns and David, Signal Processing
> Routines in Matlab, which is compatible with Octave) is that the phase
> unwrap routines are good for 2*pi+, but then get confused. Maybe a more
> robust version of unwrap is in order with an "if" (or some such) to check
> the curve continuation.
>
> Joe
>
>
>  
This works for High and Low pass filters

My first file was only for Low pass filters.
Doug Stewart
There is probably a better way but this works for me. :-)




# fix the phase wrap around
do
        ll=length(ph);
        [qm iqm]=max((diff(ph)));
        if(qm>300)
         ph(iqm+1:ll)=ph(iqm+1:ll)-360;
        endif
        [qm iqm]=min((diff(ph)));
        if(qm<-300)
        ph(iqm+1:ll)=ph(iqm+1:ll)+360;
        endif
qm
until (abs(qm)<300)





-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------