Hi All, The function zp2sos in Signals always returns 1.0 for the gain of the second order sections. This isn't very useful. I believe it should return the gain factor to make the cascade of second order sections equal to the input filter which is in the form [zeros poles gain]. Indeed, this is the behavior described for the Matlab version at http://www.mathworks.com/help/toolbox/signal/ref/zp2sos.html.
An example that shows the problem is simply to convert a second order filter in z-p-g form to a single second order section: octave-3.4.0:1> [b,a] = butter(2, 0.2) b = 0.067455 0.134911 0.067455 a = 1.00000 -1.14298 0.41280 octave-3.4.0:2> [z, p, g] = butter(2, 0.2) z = -1 -1 p = 0.57149 + 0.29360i 0.57149 - 0.29360i g = 0.067455 octave-3.4.0:3> [sos, sos_g] = zp2sos(z', p') sos = e 1.00000 2.00000 1.00000 1.00000 -1.14298 0.41280 sos_g = 1 >>>>> Should be 0.067455 <<<<<< The filter from the first input line should be the same as from the last input line but it differs in the gain factor. This is easy to fix but as it is the user has to do it. On two related matters, the Octave documentation for zp2sos, http://octave.sourceforge.net/signal/function/zp2sos.html, states that the arguments to zp2sos should be column vectors but using either column or row vectors results in the same answer. Also, on the same doc page, a quantity called Bscale is mentioned, but only once; it does not reference anything on the page. This should also be fixed. It probably refers to g but g is incorrect as noted above. I suppose the functions mentioned in the "See also" should be checked for similar problems. Jerry ------------------------------------------------------------------------------ All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-novd2d _______________________________________________ Octave-dev mailing list Octave-dev@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/octave-dev
