High-pass Butterworth filter and filtfilt

Hello everyone, I'm trying to high-filter a signal beacause when I perform a fft of the original signal I find a contribution at very low frequency even if I take the fft of detrend(signal). I want to cut out that contribution. I wrote this simple script:

Fcp=3;  %cutoff frequency
[z,p,k]=butter(8,Fcp/(Fsp/2),'high');
sos=zp2sos(z,p,k);
%fvtool(sos,'Analysis','freq')

x101filtfilt=filtfilt(z,p,x101);
figure;
plot(t101,x101,t101,x101filtfilt);
grid on
title('PE101 (pressure signal)')
xlabel('t (s)')
ylabel('Pressure (kPa)')
legend({'Original PE101 Signal','Actual PE101 Signal (filtered and realigned signal)'});

The sampling frequency is 600 Hz and the signal x101 is about 180 s long. I receive the error: "Warning: Imaginary parts of complex X and/or Y arguments ignored". As a matter of fact the new signal x101filtfilt contains imaginary numbers. Where am I wrong? Thank you very much.

[Jan]

And does x101, z or p contain complex values? Which command produces this error?

Answer 1

[Star Strider]

You have two errors that I can see.

1. You need both outputs from zp2sos:

[sos,g] = zp2sos(z,p,k);

2. Then use those in your filtfilt call:

x101filtfilt = filtfilt(sos,g,x101);

That should solve your imaginary number problem (unless ‘x101’ is complex).

I also analysed your filter with freqz:

freqz(sos, 2^16, Fsp)

It appears stable and that it will do what you want it to do.

[Guglielmo Giambartolomei]

Star Strider, you're a fallen angel; it works now! Your syntax is different from the one recommended in matlab guide. Thank you very much!

[Star Strider]

As always, my pleasure!